NHS R&D HTA Programme - LSHTM Research...

Health Technology Assessm

ent 2006;Vol. 10: No. 23

The treatm

ent of attention deficit hyperactivity disorder in children and adolescents

A systematic review and economicmodel of the effectiveness and cost-effectiveness of methylphenidate,dexamfetamine and atomoxetine forthe treatment of attention deficithyperactivity disorder in children andadolescents

S King, S Griffin, Z Hodges, H Weatherly, C Asseburg, G Richardson, S Golder, E Taylor, M Drummond and R Riemsma

Health Technology Assessment 2006; Vol. 10: No. 23

HTAHealth Technology AssessmentNHS R&D HTA Programme

The National Coordinating Centre for Health Technology Assessment,Mailpoint 728, Boldrewood,University of Southampton,Southampton, SO16 7PX, UK.Fax: +44 (0) 23 8059 5639 Email: [email protected]://www.hta.ac.uk ISSN 1366-5278

FeedbackThe HTA Programme and the authors would like to know

your views about this report.

The Correspondence Page on the HTA website(http://www.hta.ac.uk) is a convenient way to publish

your comments. If you prefer, you can send your comments to the address below, telling us whether you would like

us to transfer them to the website.

We look forward to hearing from you.

July 2006

Copyright notice

© Queen's Printer and Controller of HMSO 2006 HTA reports may be freely reproduced for the purposes of private research and study and may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising Violations should be reported to [email protected] Applications for commercial reproduction should be addressed to HMSO, The Copyright Unit, St Clements House, 2–16 Colegate, Norwich NR3 1BQ

How to obtain copies of this and other HTA Programme reports.An electronic version of this publication, in Adobe Acrobat format, is available for downloading free ofcharge for personal use from the HTA website (http://www.hta.ac.uk). A fully searchable CD-ROM isalso available (see below).

Printed copies of HTA monographs cost £20 each (post and packing free in the UK) to both public andprivate sector purchasers from our Despatch Agents.

Non-UK purchasers will have to pay a small fee for post and packing. For European countries the cost is£2 per monograph and for the rest of the world £3 per monograph.

You can order HTA monographs from our Despatch Agents:

– fax (with credit card or official purchase order) – post (with credit card or official purchase order or cheque)– phone during office hours (credit card only).

Additionally the HTA website allows you either to pay securely by credit card or to print out yourorder and then post or fax it.

Contact details are as follows:HTA Despatch Email: [email protected]/o Direct Mail Works Ltd Tel: 02392 492 0004 Oakwood Business Centre Fax: 02392 478 555Downley, HAVANT PO9 2NP, UK Fax from outside the UK: +44 2392 478 555

NHS libraries can subscribe free of charge. Public libraries can subscribe at a very reduced cost of £100 for each volume (normally comprising 30–40 titles). The commercial subscription rate is £300 per volume. Please see our website for details. Subscriptions can only be purchased for the current orforthcoming volume.

Payment methods

Paying by chequeIf you pay by cheque, the cheque must be in pounds sterling, made payable to Direct Mail Works Ltdand drawn on a bank with a UK address.

Paying by credit cardThe following cards are accepted by phone, fax, post or via the website ordering pages: Delta, Eurocard,Mastercard, Solo, Switch and Visa. We advise against sending credit card details in a plain email.

Paying by official purchase orderYou can post or fax these, but they must be from public bodies (i.e. NHS or universities) within the UK.We cannot at present accept purchase orders from commercial companies or from outside the UK.

How do I get a copy of HTA on CD?

Please use the form on the HTA website (www.hta.ac.uk/htacd.htm). Or contact Direct Mail Works (seecontact details above) by email, post, fax or phone. HTA on CD is currently free of charge worldwide.

The website also provides information about the HTA Programme and lists the membership of the variouscommittees.

HTA


S King,1* S Griffin,2 Z Hodges,1 H Weatherly,2

C Asseburg,2 G Richardson,2 S Golder,1 E Taylor,3

M Drummond2 and R Riemsma1

1 Centre for Reviews and Dissemination, University of York, UK2 Centre for Health Economics, University of York, UK3 Institute of Psychiatry, King’s College London, UK

* Corresponding author

Declared competing interests of authors: E Taylor acted as a representative of theRCP at the NICE Appraisal Committee meeting. R Riemsma is a member of the editorialboard for Health Technology Assessment, but he was not involved in the editorial processfor this report.

Published July 2006

This report should be referenced as follows:

King S, Griffin S, Hodges Z, Weatherly H, Asseburg C, Richardson G, et al. A systematicreview and economic model of the effectiveness and cost-effectiveness of methylphenidate,dexamfetamine and atomoxetine for the treatment of attention deficit hyperactivitydisorder in children and adolescents. Health Technol Assess 2006;10(23).

Health Technology Assessment is indexed and abstracted in Index Medicus/MEDLINE,Excerpta Medica/EMBASE and Science Citation Index Expanded (SciSearch®) and Current Contents®/Clinical Medicine.

NHS R&D HTA Programme

The research findings from the NHS R&D Health Technology Assessment (HTA) Programme directlyinfluence key decision-making bodies such as the National Institute for Health and Clinical

Excellence (NICE) and the National Screening Committee (NSC) who rely on HTA outputs to help raisestandards of care. HTA findings also help to improve the quality of the service in the NHS indirectly inthat they form a key component of the ‘National Knowledge Service’ that is being developed to improvethe evidence of clinical practice throughout the NHS.

The HTA Programme was set up in 1993. Its role is to ensure that high-quality research information onthe costs, effectiveness and broader impact of health technologies is produced in the most efficient wayfor those who use, manage and provide care in the NHS. ‘Health technologies’ are broadly defined toinclude all interventions used to promote health, prevent and treat disease, and improve rehabilitationand long-term care, rather than settings of care.

The HTA Programme commissions research only on topics where it has identified key gaps in theevidence needed by the NHS. Suggestions for topics are actively sought from people working in theNHS, the public, service-users groups and professional bodies such as Royal Colleges and NHS Trusts.

Research suggestions are carefully considered by panels of independent experts (including service users)whose advice results in a ranked list of recommended research priorities. The HTA Programme thencommissions the research team best suited to undertake the work, in the manner most appropriate to findthe relevant answers. Some projects may take only months, others need several years to answer theresearch questions adequately. They may involve synthesising existing evidence or conducting a trial toproduce new evidence where none currently exists.

Additionally, through its Technology Assessment Report (TAR) call-off contract, the HTA Programme isable to commission bespoke reports, principally for NICE, but also for other policy customers, such as aNational Clinical Director. TARs bring together evidence on key aspects of the use of specifictechnologies and usually have to be completed within a short time period.

Criteria for inclusion in the HTA monograph seriesReports are published in the HTA monograph series if (1) they have resulted from work commissionedfor the HTA Programme, and (2) they are of a sufficiently high scientific quality as assessed by the refereesand editors.

Reviews in Health Technology Assessment are termed ‘systematic’ when the account of the search,appraisal and synthesis methods (to minimise biases and random errors) would, in theory, permit thereplication of the review by others.

The research reported in this monograph was commissioned and funded by the HTA Programme onbehalf of NICE as project number 03/33/01. The protocol was agreed in July 2004. The assessment reportbegan editorial review in May 2005 and was accepted for publication in July 2005. The authors have beenwholly responsible for all data collection, analysis and interpretation, and for writing up their work. TheHTA editors and publisher have tried to ensure the accuracy of the authors’ report and would like tothank the referees for their constructive comments on the draft document. However, they do not acceptliability for damages or losses arising from material published in this report.

The views expressed in this publication are those of the authors and not necessarily those of the HTA Programme, NICE or the Department of Health.

Editor-in-Chief: Professor Tom WalleySeries Editors: Dr Aileen Clarke, Dr Peter Davidson, Dr Chris Hyde,

Dr John Powell, Dr Rob Riemsma and Dr Ken SteinManaging Editors: Sally Bailey and Sarah Llewellyn Lloyd

ISSN 1366-5278

© Queen’s Printer and Controller of HMSO 2006

This monograph may be freely reproduced for the purposes of private research and study and may be included in professional journals providedthat suitable acknowledgement is made and the reproduction is not associated with any form of advertising.

Applications for commercial reproduction should be addressed to NCCHTA, Mailpoint 728, Boldrewood, University of Southampton, Southampton, SO16 7PX, UK.

Published by Gray Publishing, Tunbridge Wells, Kent, on behalf of NCCHTA.Printed on acid-free paper in the UK by St Edmundsbury Press Ltd, Bury St Edmunds, Suffolk. T

Objectives: To assess the clinical and cost-effectiveness of oral methylphenidate hydrochloride(MPH), dexamfetamine sulphate (DEX) andatomoxetine (ATX) in children and adolescents (<18 years of age) diagnosed with attention deficithyperactivity disorder (ADHD) (including hyperkineticdisorder). Data sources: Electronic databases covering 1999–July 2004 for MPH, 1997–July 2004 for DEX and1981–July 2004 for ATX.Review methods: Selected studies were assessedusing modified criteria based on CRD Report No. 4.Clinical effectiveness data were reported separately foreach drug and by the type of comparison. Data forMPH were also analysed separately based on whetherit was administered as an immediate release (IR) orextended release (ER) formulation. For all drugs, thedata were examined by dose. Data for the coreoutcomes of hyperactivity (using any scale), ClinicalGlobal Impression [as a proxy of quality of life (QoL)]and adverse events were reported. For crossoverstudies, the mean and standard deviation (SD) for eachoutcome were data extracted for end of trial data (i.e.baseline data were not considered). For parallelstudies, change scores were reported where given,otherwise means and SDs were presented for end oftrial data. In addition, mean differences with 95%confidence intervals were calculated for each study. Foradverse events, self-ratings were reported when used,otherwise, parent reports were utilised. Percentages ofparticipants reporting adverse events were used tocalculate numbers of events in each treatment arm. Allthe clinical effectiveness data and economic evaluations(including accompanying models) included in the

company submissions were assessed. A new model wasdeveloped to assess the cost-effectiveness of thealternative treatments in terms of cost per quality-adjusted life-year. To achieve this, a mixed treatmentcomparison model was used to estimate the differentialmean response rates. Monte Carlo simulation was usedto reflect uncertainty in the cost-effectiveness results.Results: In total, 65 papers met the inclusion criteria.The results suggest that MPH and DEX are effective atreducing hyperactivity and improving QoL (asdetermined by Clinical Global Impression) in children,although the reliability of the MPH study results is notknown and there were only a small number of DEXstudies. There was consistent evidence that ATX wassuperior to placebo for hyperactivity and Clinical GlobalImpression. Studies on ATX more often reported thestudy methodology well, and the results were likely tobe reliable. Very few studies made direct head-to-headcomparisons between the drugs or examined a non-drug intervention in combination with MPH, DEX orATX. Adequate and informative data regarding thepotential adverse effects of the drugs were also lacking.The results of the economic evaluation clearlyidentified an optimal treatment strategy of DEX first-line, followed by IR-MPH for treatment failures,followed by ATX for repeat treatment failures. WhereDEX is unsuitable as a first-line therapy, the optimalstrategy is IR-MPH first-line, followed by DEX and then ATX. For patients contraindicated to stimulants,ATX is preferred to no treatment. For patients inwhom a midday dose of medication is unworkable, ER-MPH is preferred to ATX, and ER-MPH12 appearsmore cost-effective than ER-MPH8. As identified in theclinical effectiveness review, the reporting of studies


iii

© Queen’s Printer and Controller of HMSO 2006. All rights reserved.

Abstract

A systematic review and economic model of the effectivenessand cost-effectiveness of methylphenidate, dexamfetamine andatomoxetine for the treatment of attention deficit hyperactivitydisorder in children and adolescents

S King,1* S Griffin,2 Z Hodges,1 H Weatherly,2 C Asseburg,2 G Richardson,2

S Golder,1 E Taylor,3 M Drummond2 and R Riemsma1

1 Centre for Reviews and Dissemination, University of York, UK2 Centre for Health Economics, University of York, UK3 Institute of Psychiatry, King’s College London, UK* Corresponding author

was poor, therefore this should be borne in mind wheninterpreting the model results.Conclusions: Drug therapy seems to be superior to no drug therapy, no significant differences between thevarious drugs in terms of efficacy or side effects werefound, mainly owing to lack of evidence, and theadditional benefits from behavioural therapy (incombination with drug therapy) are uncertain. Giventhe lack of evidence for any differences in effectivenessbetween the drugs, the economic model tended to be

driven by drug costs, which differed considerably.Future trials examining MPH, DEX and ATX shouldinclude the assessment of tolerability and safety as apriority. Longer term follow-up of individualsparticipating in trials could further inform policy makersand health professionals. Such data could potentiallydistinguish between these drugs in a clinically usefulway. In addition, research examining whether somaticcomplaints are actually related to drug treatment or tothe disorder itself would be informative.

Abstract

iv


v


Contents

Glossary and list of abbreviations ............. vii

Executive summary .................................... xiii

1 Aim of the review ...................................... 1

2 Background ................................................ 3Description of underlying health problem ....................................................... 3Current service provision ........................... 4Description of intervention ........................ 5

3 Methods for reviewing effectiveness and cost-effectiveness ....................................... 7Search strategy ........................................... 7Inclusion and exclusion criteria: clinicaleffectiveness ................................................ 8Inclusion and exclusion criteria for systematic reviews: clinical effectiveness ................................................ 9Data extraction strategy: clinical effectiveness ................................................ 10Quality assessment strategy: clinicaleffectiveness ................................................ 10Analysis strategy: clinical effectiveness ...... 10Cost-effectiveness ....................................... 11

4 Clinical effectiveness .................................. 13Quantity and quality of research available ...................................................... 13Assessment of effectiveness ........................ 13Summary of clinical effectiveness data ...... 75Summary of adverse events data ............... 76

5 Review of economic evaluations of ADHD drug interventions in children and adolescents ................................................. 79Aim ............................................................. 79Literature review ........................................ 79Review of quality of life and cost-effectiveness literature ........................ 79Review of the Janssen-Cilag submission .................................................. 86Review of the Celltech submission ............. 93Review of the Eli Lilly submission ............. 96

6 Economic model ......................................... 101Methods ...................................................... 102Results ........................................................ 109Discussion ................................................... 122

7 Discussion ................................................... 125Effectiveness evidence ................................ 125Adverse events ............................................ 125Economic evidence ..................................... 126Limitations of the clinical effectiveness studies and need for further research ........ 127

8 Conclusion .................................................. 129Research recommendations ....................... 129

Acknowledgements .................................... 131

References .................................................. 133

Appendix 1 Clinical effectiveness research ...................................................... 147

Appendix 2 Economic evaluations and health-related quality of life research ........ 165

Appendix 3 Excluded studies from theupdated search (after the first screening) (n = 115) .................................................... 171

Appendix 4 Excluded studies from NICE, CCOHTA and AHRQ reviews (n = 28) ...................................................... 175

Appendix 5 Quality assessment questions used for clinical effectiveness studies (as modified from CRD Report No. 4) ...... 177

Appendix 6 Economic evaluation qualityassessment checklist ................................... 179

Appendix 7 Economic evaluation dataextraction forms ......................................... 181

Appendix 8 All possible treatment strategies ..................................................... 187

Appendix 9 WinBUGS code ...................... 189

Appendix 10 Health state descriptions used to elicit standard gamble utility estimates from parents of children with ADHD ......................................................... 191

Appendix 11 WinBUGS code for extendedMTC model ................................................ 199

Appendix 12 Data extraction tables of clinical effectiveness studies ....................... 201

Appendix 13 Data extraction table of thesystematic review of adverse events ........... 365

Health Technology Assessment reportspublished to date ....................................... 367

Health Technology Assessment Programme ................................................ 379

Contents

vi

GlossaryAdverse effect An abnormal or harmfuleffect caused by and attributable to exposure toa chemical (e.g. a drug), which is indicated bysome result such as death, a physical symptomor visible illness. An effect may be classed asadverse if it causes functional or anatomicaldamage, causes irreversible change in thehomeostasis of the organism or increases thesusceptibility of the organism to other chemicalor biological stress.

Attention deficit hyperactivity disorder Amental disorder characterised by a persistentpattern of inattention and/orhyperactivity–impulsivity that is morefrequently displayed and more severe than istypically observed in individuals at acomparable level of development.

Bias Deviation of results or inferences from the truth, or processes leading to such deviation. Any trend in the collection,analysis, interpretation, publication or review of data that can lead to conclusions that are systematically different from the truth.

Blinding (synonym: masking) Keepingsecret group assignment (e.g. to treatment or control) from the study participants orinvestigators. Blinding is used to protectagainst the possibility that knowledge ofassignment may affect patient response totreatment, provider behaviours (performancebias) or outcome assessment (detection bias).Blinding is not always practical (e.g. whencomparing surgery with drug treatment). The importance of blinding depends on how objective the outcome measure is; blinding is more important for less objectiveoutcome measures such as pain or quality of life.

Chi-squared (�2) test Any statistical testbased on comparison of a test statistic with achi-squared distribution.

Concealment of allocation The process usedto prevent foreknowledge of group assignmentin a randomised controlled trial, which shouldbe seen as distinct from blinding. Theallocation process should be impervious to anyinfluence by the individual making theallocation by having the randomisation processadministered by someone who is notresponsible for recruiting participants, forexample, a hospital pharmacy or a centraloffice. Methods of assignment such as date ofbirth and case record numbers (see quasi-random allocation) are open to manipulation.Adequate methods of allocation concealmentinclude: centralised randomisation schemes;randomisation schemes controlled by apharmacy; numbered or coded containers inwhich capsules from identical-looking,numbered bottles are administeredsequentially; on-site computer systems, whereallocations are in a locked unreadable file; andsequentially numbered opaque, sealedenvelopes.

Conduct disorder A mental disordercharacterised by a repetitive and persistentpattern of behaviour in which the basic rightsof others or major age-appropriate societalnorms or rules are violated, manifested byaggressive, defiant or antisocial behaviours.

Confidence interval A measure of precisionof statistical estimate; quantifies the uncertaintyin measurement. Usually reported as 95% CI,i.e. the range of values within which one can be95% sure that the true values for the wholepopulation lie.

continued


vii


Glossary and list of abbreviations

Technical terms and abbreviations are used throughout this report. The meaning is usually clear fromthe context, but a glossary is provided for the non-specialist reader. In some cases, usage differs in the

literature, but the term has a constant meaning throughout this review.


viii

Glossary continued

Confounding (1) The masking of an actualassociation or (2) false demonstration of anapparent association between the studyvariables when no real association betweenthem exists.

Construct validity An instrument exhibitsthis if it is demonstrated that it correlates withother trusted measures of the same effect beingmeasured and that it can discriminate betweengroups with known differences.

Cost–benefit analysis An attempt to give theconsequences of the alternative interventions amonetary value. In this way, the consequencescan be more easily compared with the costs ofthe intervention. This involves measuringindividuals’ ‘willingness to pay’ for givenoutcomes, and can be difficult.

Cost–consequence analysis Costs arereported separately from health effects.

Cost-effectiveness analysis Theconsequences of the alternatives are measuredin natural units, such as years of life gained.The consequences are not given a monetaryvalue.

Cost-effectiveness acceptability curve Agraphical representation of the probability ofan intervention being cost-effective over arange of monetary values for society’swillingness to pay for an additional unit ofhealth gain.

Cost minimisation When two alternatives arefound to have equal efficacy or outcomes(consequences). Therefore, the only differencebetween the two is cost. This is sometimesconsidered to be a subtype of cost-effectivenessanalysis.

Cost–utility analysis The consequences ofalternatives are measured in ‘health statepreferences’, which are given a weighting score.In this type of analysis, different consequencesare valued in comparison with each other, andthe outcomes (e.g. life-years gained) areadjusted by the weighting assigned. In this way,an attempt is made to value the quality of lifeassociated with the outcome so that life-yearsgained become quality-adjusted life-yearsgained.

Crossover trial A type of clinical trialcomparing two or more interventions in whichthe participants, upon completion of thecourse of one treatment, are switched toanother. For example, for a comparison oftreatments A and B, half the participants arerandomly allocated to receive them in theorder A, B and half to receive them in theorder B, A. A problem with this design is thatthe effects of the first treatment may carry overinto the period when the second is given.

Discounting The process of converting futurepounds sterling and future health effects totheir present value.

Discriminant validity An instrument exhibitsthis for the extent to which it does notcorrelate with variables and measures thoughtto be unrelated to the construct beingmeasured.

Dominance The state when an interventionunder study is both less costly and moreeffective than the comparator(s).

Economic evaluation Comparative analysisof alternative course of action in terms of boththeir costs and effects.

Effectiveness The extent to which a specificintervention, when used under ordinarycircumstances, does what it is intended to do.

Extended dominance The state when astrategy is both more costly and less effectivethan a linear combination of two otherstrategies with which it is mutually exclusive.

External validity The ability to generalisethe results from a particular experiment to alarger population.

First-line treatment The first regimen givento patients

Incidence The number of new cases of adisease or event in a population during aspecific period of time.

Incremental cost-effectiveness ratio Anexpression of the additional cost of health gainassociated with an intervention relative to anappropriate comparator. Expressed as thedifference in mean costs (relative to thecomparator) divided by the difference in mean

continued

Glossary continued

effects. Sometimes expressed with confidenceintervals.

Intention-to-treat An ITT analysis is one inwhich all the participants in a trial are analysedaccording to the intervention to which theywere allocated, whether they received it or not.ITT analyses are favoured in assessments ofeffectiveness as they mirror the non-compliance and treatment changes that arelikely to occur when the intervention is used inpractice, and because of the risk of attritionbias when participants are excluded from theanalysis.

Internal validity The degree to which astudy is logically sound and free ofconfounding variables.

Markov Chain Monte CarloA mathematical model containing a finitenumber of mutually exclusive and exhaustivehealth states, with uniform time periods and inwhich the probability of movement from onestate to another depends on the current stateand remains constant over time.

Methodological quality (synonyms: validity,internal validity) The extent to which thedesign and conduct of a study are likely to haveprevented systematic errors (bias). Variation inquality can explain variation in the results ofstudies included in a systematic review. Morerigorously designed (better ‘quality’) trials aremore likely to yield results that are closer to the‘truth’.

Meta-analysis The statistical pooling of theresults of a collection of related individualstudies, to increase statistical power andsynthesise their findings.

Mixed treatment comparison Mixedtreatment comparison is a form of meta-analysis used to strengthen inferenceconcerning the relative efficacy of twotreatments. It uses data based on directcomparisons (A versus B and B versus C trials)and indirect comparisons (A versus C trials);also, it facilitates simultaneous inferenceregarding all treatments in order to select thebest treatments.

p-Value In the context of significant tests, thep-value represents the probability that a given

difference is observed in a study sample, whensuch a difference does not exist in the relevantpopulation. Small p-values indicate strongerevidence to reject the null hypothesis of nodifference.

Parallel group trial (synonym: independentgroup design) A trial that compares twogroups of people, one of which receives theintervention of interest and one of which is acontrol group. Some parallel trials have morethan two comparison groups and somecompare different interventions withoutincluding a non-intervention control group.

Prevalence The measure of the proportion ofpeople in a population who have someattribute or disease at a given point in time orduring some time period.

Oppositional defiant disorderA mental disorder characterised by a pattern ofnegativistic, defiant, disobedient and hostilebehaviour towards authority figures as evidentin such behaviour as temper tantrums,argumentativeness, refusing to comply withrequests and deliberately annoying others.

Placebo An inactive substance or procedureadministered to a patient, usually to compareits effects with those of a real drug or otherintervention, but sometimes for thepsychological benefit to the patient through abelief that they are receiving treatment.Placebos are used in clinical trials to blindpeople to their treatment allocation. Placebosshould be indistinguishable from the activeintervention to ensure adequate blinding.

Quality of life A concept incorporating allthe factors that might impact on anindividual’s life, including factors such as theabsence of disease or infirmity and also otherfactors which might affect their physical,mental and social well-being.

Quality-adjusted life-year An index ofhealth gain where survival duration is weightedor adjusted by the patient’s quality of lifeduring the survival period. QALYs have theadvantage of incorporating changes in bothquantity (mortality) and quality (morbidity) oflife.

continued


ix



x

Glossary continued

Randomised controlled trial (synonym:randomised clinical trial) An experiment inwhich investigators randomly allocate eligiblepeople into intervention groups to receive ornot to receive one or more interventions thatare being compared. The results are assessedby comparing outcomes in the treatment andcontrol groups.

Relative risk (synonym: risk ratio)The ratio of risk in the intervention group tothe risk in the control group. The risk(proportion, probability or rate) is the ratio ofpeople with an event in a group to the total inthe group. RR = 1 indicates no differencebetween comparison groups. For undesirableoutcomes an RR that is <1 indicates that theintervention was effective in reducing the riskof that outcome.

Second-line treatment The second regimenadministered either as a result of relapse afterfirst-line therapy or immediately following onfrom first-line therapy.

Sensitivity analysis A mathematical methodthat examines uncertainty associated withparameters estimated into the analysis to testthe robustness of the analysis findings. In one-way sensitivity analysis each parameter is variedindividually, for multi-way analysis two or moreparameters are varied at the same time,threshold analysis identifies the critical valuesabove or below which the results of a study varyand analysis of extremes is used to examine themost pessimistic and the most optimisticscenarios. Finally, probabilistic sensitivityanalysis attributes distributions of probabilitiesto uncertain variables that are incorporatedwithin a model.

Standard gamble Measuring a health stateutility by comparing life in a particular givenhealth state with a gamble with a probabilitythat perfect health is the outcome and thatimmediate death is the outcome. Theprobability is varied until a point ofindifference between the two choices (i.e. untilthe preference for the given health state isequal to the preference for the gamble).

Statistical significance An estimate of theprobability of an association (effect) as large orlarger than what is observed in a studyoccurring by chance, usually expressed as a

p-value. For example, a p-value of 0.049 for arisk difference of 10% means that there is lessthan a one in 20 (0.05) chance of anassociation that is as large or larger havingoccurred by chance and it could be said thatthe results are ‘statistically significant’ at p = 0.05. The cut-off for statistical significanceis usually taken at 0.05, but sometimes at 0.01or 0.10. These cut-offs are arbitrary and haveno specific importance. Although it is oftendone, it is inappropriate to interpret the resultsof a study differently according to whether thep-value is, say, 0.055 or 0.045 (which aresimilar values, not diametrically opposed ones).

Systematic review (synonym: systematicoverview) A review of a clearly formulatedquestion that uses systematic and explicitmethods to identify, select and criticallyappraise relevant research, and to collect andanalyse data from the studies that are includedin the review. Statistical methods (meta-analysis) may or may not be used to analyseand summarise the results of the includedstudies.

Time trade-off Measuring a health state bytrading off life-years in a state of less thanperfect health for a shorter life span in a stateof perfect health.

Utility A measure of the strength of anindividual’s preference for a given health stateor outcome. Utilities assign numerical valueson a scale from 0 (death) to 1 (optimal or‘perfect’ health), and provide a single numberthat summarises health-related quality of life.Hence utility has been described as a globalmeasure of health-related quality of life.Sometimes ‘utility’ is only used to refer topreferences (on the 0–1 scale) that are elicitedusing methods which introduce risky scenariosto the respondent (standard gamble), with theterm ‘values’ used to refer to other types ofpreferences.

Values An alternative measure of the strengthof an individual’s preference for a given healthstate or outcome. In contrast to utilities, valuesreflect preferences elicited in a riskless context.

Visual analogue scale Direct rating whererates are asked to place a mark at a

continued

Glossary continued

point between two anchor states appearing ateither end of the line. It is used as a method ofvaluing health states.

Washout period The stage in a crossover trialwhen treatment is withdrawn before the secondtreatment is given. Washout periods are usuallynecessary because of the possibility that the

intervention administered first can affect theoutcome variable for some time after treatmentceases. A run-in period before a trial starts issometimes called a washout period iftreatments that participants were using beforeentering the trial are discontinued.


xi


List of abbreviationsACTeRS ADD-H Comprehensive Teachers’

Rating Scale

ADD attention deficit disorder

ADD-H attention deficit disorder withhyperactivity

AE adverse event

ADHD attention deficit hyperactivitydisorder

AHRQ Agency for Healthcare Research andQuality

ARS ADHD Rating Scale

ASQ Abbreviated SymptomsQuestionnaire

ATX atomoxetine hydrochloride

BM behaviour modification

BNF British National Formulary

BSEQ Barkley Side Effects Questionnaire

BT behavioural therapy

CBCL Child Behaviour Checklist

CBT cognitive behavioural therapy

CCOHTA Canadian Coordinating Office forHealth Technology Assessment

CCT Children’s Checking Test

CCT clinically controlled trial

CD conduct disorder

CEAC cost-effectiveness acceptability curve

C-GAS Children’s Global Assessment Scale

CGI Clinical Global Impression

CGI-I Clinical Global Impressionimprovement subscale

CGI-S Clinical Global Impression severitysubscale

CHQ Child Health Questionnaire

CI confidence interval

CIC commercial-in-confidence

CNS central nervous system

CON concerta

CPRS-R Conners’ Parent Rating Scale –Revised

CPRS Conners’ Parent Rating Scale

CPRS-H Conners’ Parent HyperactivitySubscale

CTRS-R Conners’ Teacher Rating Scale –Revised

CTRS Conners’ Teacher Rating Scale

CTRS-H Conners’ Teacher HyperactivitySubscale

DEC Development and EvaluationCommittee

DEX dexamfetamine sulphate

DEX-SR sustained-release dexamfetaminesulphate

DEX-TR time-release dexamfetaminesulphate

DSM-III Diagnostic and Statistical Manual ofMental Disorders (3rd edition)

continued

Glossary and list of abbreviation

xii

List of abbreviations continued

DSM-IV Diagnostic and Statistical Manual ofMental Disorders (4th edition)

ECG electrocardiogram

EEG electroencephalogram

EQ-5D EuroQol instrument

ER-MPH extended-release methylphenidatehydrochloride

FFD Freedom from Distractability Factor

GDS Gordon Diagnostic System

HKD hyperkinetic disorder

HRQoL health-related quality of life

ICD-10 International StatisticalClassification of Diseases andRelated Health Problems (10threvision)

ICER incremental cost-effectiveness ratio

IHRQoL Index of Health-related Quality ofLife

IOWA Inattention/Overactivity withAggression

IOWA-C Inattention/Overactivity withAggression Conners’ Rating Scale

IQ intelligence quotient

IR-MPH immediate-release methylphenidatehydrochloride

ITT intention-to-treat

MCD Metadate CD

MCMC Markov Chain Monte Carlo

MD mean difference

MFFT Matching Familiar Features Test

MPH methylphenidate hydrochloride

MPH-SR sustained-release methylphenidatehydrochloride

MTA Multimodal Treatment Study ofADHD

MTC mixed treatment comparison

NICE National Institute for Health andClinical Excellence

NS not significant

ODD oppositional defiant disorder

PACS Parental Account of ChildhoodSymptoms

PEM pemoline

PIAT Peabody Individual AchievementTest

PPA Prescription Pricing Authority

PSS Personal Social Services

QALY quality-adjusted life-year

QoL quality of life

RCT randomised controlled trial

RR relative risk

SD standard deviation

SEM standard error of the mean

SERS Side Effects Rating Scale

SG standard gamble

SHS Schenectady Hyperkinetic Scale

SKAMP Swanson, Kotkin, Agler, M-Flynnand Pelham (scale)

SNAP Swanson, Nolan and Pelham (scale)

SR systematic review

STESS Subject’s Treatment EmergentSymptom Scale

STP Summer Treatment Programme

TIP Telephone Interview Probe

TOTS Loney’s Time on Task Scale

TTO time-trade-off

VAS visual analogue scale

WISC-R Weschlar Intelligence Scale forChildren – Revised

WRAT Wide Range Achievement Test

All abbreviations that have been used in this report are listed here unless the abbreviation is well known (e.g. NHS), or it has been used only once, or it is a non-standard abbreviation used only in figures/tables/appendices in which case the abbreviation is defined in the figure legend or at the end of the table.


xiii


BackgroundAttention deficit hyperactivity disorder (ADHD)(including hyperkinetic disorder) is defined by the‘core’ signs of inattention, hyperactivity andimpulsivity, and is characterised by an early onset.The estimated prevalence for ADHD in school-aged children varies widely (e.g. 3–7%), beingdependent on a number of variables, includingthe methods of ascertainment, the informants, thepopulation sampled, the diagnostic criteriaapplied and the sex of the affected individual.Data on prevalence in adolescence and adulthoodare limited. The disorder is frequently observed ingreater numbers of males than females, with ratiosranging from 2:1 to 9:1 depending on subtypeand setting.

There are two generally used diagnostic criteria:the International Statistical Classification ofDiseases and Related Health Problems (ICD-10)and Diagnostic and Statistical Manual of MentalDisorders (DSM-IV) criteria. The ICD-10 presentsdetails on the diagnosis of hyperkinetic disorders(HKD) and the DSM-IV criteria define ADHDmore broadly to include three subtypes: acombined subtype in which all three core signs arepresent, a predominantly inattentive subtype inwhich inattention is present but nothyperactivity/impulsivity and a predominantlyhyperactive–impulsive subtype in whichhyperactivity/impulsivity are present but notinattention. As the ICD-10 criteria are similar tothe severe combined type ADHD defined by theDSM-IV criteria, prevalence rates may be higherusing the DSM-IV criteria than when diagnosedusing the ICD-10 criteria.

Current treatments for ADHD include social,psychological and behavioural interventions inaddition to medical management. Medicationscurrently licensed for the treatment of ADHD inthe UK include methylphenidate hydrochloride(MPH), dexamfetamine sulphate (DEX) andatomoxetine (ATX), although clinicians sometimesprescribe tricyclic and other antidepressants. MPHis available in immediate-release (Ritalin® andEquasym®) and extended-release forms [Concerta®

XL and Equasym XL® (a licence application forEquasym XL had been submitted; it has been

specifically developed to provide efficacy across theschool day and replace the need for twice dailydosing for children who do not consistently require evening medication)]. They are allindicated in children over 6 years of age, and inadolescents. DEX can be given to children asyoung as 3 years, whereas ATX (licensed in the UK in May 2004) is indicated in children aged 6 years and above.

ObjectiveThe objective was to assess the clinical and cost-effectiveness of oral MPH, DEX and ATX inchildren and adolescents (under 18 years of age)diagnosed with ADHD (including hyperkineticdisorder).

MethodsThis systematic review incorporated studies from,and built upon, three previous systematic reviews:

� A review conducted by the American Agency forHealthcare Research and Quality (AHRQ)published in 1999 (Jadad and colleagues, 1999).

� A report for the Canadian Coordinating Officefor Health Technology Assessment (CCOHTA)(Miller and colleagues, 1999).

� A previous National Institute for Health and Clinical Excellence (NICE) review, which was also based primarily on evidencefrom the AHRQ report (Lord and Paisley,2000).

Search strategyThe searches, conducted in July 2004, aimed toretrieve both published and unpublished paperswith no language restrictions. A date restriction of1999 onwards was placed on the methylphenidatesearches to update the report produced by Lordand Paisley published in 2000. A date restrictionof 1997 onwards was placed on the searches fordexamfetamine to update the AHRQ report(which included a review of this drug). Researchon atomoxetine was searched for from 1981onwards. The search strategy was based on thatused in the AHRQ report.

Executive summary

xiv

Inclusion/exclusion criteriaTo be eligible for inclusion, studies had to berandomised controlled trials (RCTs) of at least 3 weeks’ duration (3 weeks per treatment arm inparallel studies and 3 weeks in overall trial lengthfor crossover studies). In addition, systematicreviews were included to examine adverse eventsdata. For the assessment of cost-effectiveness, abroader range of studies was considered.

The studies had to examine MPH, DEX or ATXused alone or in combination with non-druginterventions and be compared with placebo, withone another in head-to-head comparisons or withnon-drug interventions. Non-drug interventionsincluded any type of psychological andbehavioural strategies (e.g. cognitive behaviouraltherapy, child or parent training, bibliotherapy)and/or nutritional interventions. Studies thatcompared MPH, DEX or ATX with other drugs(e.g. Adderall) not licensed in the UK for ADHDwere included as long as there was a placebogroup. This was applied to both efficacy andadverse events data.

Participants included children and adolescentsunder 18 years of age diagnosed with ADHD(including hyperkinetic disorder). There was nolower age limitation (although there was no lowerage limitation for the report, it is noted that MPHis indicated for children older than 6 years, andATX is indicated for children aged 6 years andover). The diagnosis must have been made in anexplicit way, preferably using either the ICD-10criteria or the DSM-IV criteria. Studies includingparticipants with conditions other than ADHD(e.g. Tourette’s syndrome) were excluded unlessthey reported separate analyses for patients withADHD alone.

To be included in the review, trials had to reportresults on one or more of the following:

� core symptoms (including measures ofinattention, hyperactivity, impulsivity)

� quality of life (QoL) (Clinical Global Impressionor overall severity indices were used as a proxyof QoL)

� adverse effects (including loss of appetite,insomnia, headache, stomach ache and weightloss).

Studies that only examined tests of psychologicalfunction (e.g. the continuous performance test),measures of depression and/or anxiety, ormeasures of coexistent problems (including poorpeer relationships, and conduct/oppositional-

disorder-related outcomes) were not included inthe review. Studies that presented results in figureswithout presenting actual numbers, or onlysignificance values for comparisons, were excludedfrom the review.

Studies that met the inclusion criteria above, butwere only published as abstracts or as conferencepresentations were not included in the reviewunless a full paper could be obtained that relatedto the abstract.

Two reviewers independently screened all titlesand abstracts, including economic evaluations,identified in the updated literature search. Fullpaper manuscripts of any titles/abstracts that wereconsidered relevant by either reviewer wereobtained where possible. In addition, full papercopies of relevant studies presented in the NICE,AHRQ and CCOHTA reports were obtained. Thefull papers were then assessed against theinclusion criteria by one reviewer and checked byanother. Any discrepancies were resolved byconsensus and, if necessary, a third reviewer wasconsulted.

Data extraction and quality assessmentThe quality of the clinical effectiveness studies wasassessed using modified criteria based on CRDReport No. 4. Each study was assessed and datawere extracted by one reviewer and independentlychecked for agreement with a second reviewer.Disagreements were resolved by consensus and, ifnecessary, a third reviewer was consulted.

Methods of analysis/synthesisClinical effectiveness data were reportedseparately for each drug and by the type ofcomparison. Data for MPH were also analysedseparately based on whether it was administered asan immediate release or extended releaseformulation. For all drugs, the data wereexamined by dose. Data for the core outcomes ofhyperactivity (using any scale), Clinical GlobalImpression (as a proxy of QoL) and adverse eventswere reported. For crossover studies, the meanand standard deviation (SD) for each outcomewere data extracted for end of trial data (i.e.baseline data were not considered). Wherepossible, we aimed to calculate mean differenceand standard errors for crossover studies in orderto facilitate meta-analysis. However, owing to thelack of information needed to calculate meandifferences in many of the studies, this was notpossible. For parallel studies, change scores werereported where given, otherwise means and SDswere presented for end of trial data. In addition,

Executive summary

mean differences with 95% confidence intervalswere calculated for each study. For adverse events,self-ratings were reported when used, otherwise,parent reports were utilised. Percentages ofparticipants reporting adverse events were used to calculate numbers of events in eachtreatment arm.

For the cost-effectiveness section of the report,details of each identified published economicevaluation, together with a critical appraisal of itsquality, were presented in structured tables.

Handling company submissionsAll the clinical effectiveness data included in thecompany submissions were assessed. Where thesemet the inclusion criteria they were included inthe clinical effectiveness review. All economicevaluations (including accompanying models)included in the company submissions wereassessed and detailed assessments of theassumptions underlying the submitted analyseswere undertaken.

A new model was developed to assess the cost-effectiveness of the alternative treatments in terms of cost per quality-adjusted life-year. Toachieve this, a mixed treatment comparison modelwas used to estimate the differential meanresponse rates. Monte Carlo simulation was usedto reflect uncertainty in the cost-effectivenessresults.

ResultsClinical effectivenessIn the previous systematic reviews (NICE, AHRQand CCOHTA), 65 studies were identified aspotentially relevant to the current systematicreview, and full paper copies were ordered. Ofthese, 40 met the inclusion criteria. In theupdated search, a total of 2908 titles and abstracts relating to clinical effectiveness orsystematic reviews of adverse events wereidentified and screened for relevance. Of these,409 full paper copies were examined in detail and assessed for inclusion. Of these, 20 RCTs andone systematic review met the inclusion criteria. Inaddition, four commercial-in-confidence paperswere included. Overall, this gives a total of 65papers.

As reported in the previous NICE report, and inthe AHRQ and CCOHTA reviews, the plethora ofMPH studies suggest that MPH is effective atreducing hyperactivity and improving QoL (as

determined by Clinical Global Impression) inchildren. It was noted, however, that the majorityof studies that evaluated the effectiveness of MPHdid not adequately report their studymethodology. Hence, the reliability of the studyresults is not known. There appears to be littleevidence supporting a difference in theeffectiveness of immediate-release (IR) andextended-release (ER) MPH.

Similarly, DEX also appears to be effective atreducing hyperactivity and improving QoL,although this is based on a small number ofstudies. Only one study adequately reported thestudy methodology.

There was consistent evidence that ATX wassuperior to placebo for hyperactivity and ClinicalGlobal Impression. Studies on ATX more oftenreported the study methodology well, and theresults are likely to be reliable.

Very few studies made direct head-to-headcomparisons between the drugs. The previousNICE report stated that there appeared to be littleevidence of difference in the effectiveness of MPHand DEX. No recent studies were found in theupdated search. Although the studies reportedvariable results, the one study that reported nostatistically significant differences between the twodrugs was deemed to be of good quality, whereasthe quality of the others was uncertain given thepoor reporting of study methodologies.

One study that compared MPH and ATX reportedno differences between the drugs for hyperactivityor Clinical Global Impression. This study did notadequately report study methodology, and theresults should be interpreted with caution.[Confidential information removed].

Few studies were included in the review thatexamined a non-drug intervention in combinationwith MPH, DEX or ATX. Generally, the resultswere variable. The studies were, however,heterogeneous regarding the type of non-druginterventions examined and the scales used tomeasure outcomes.

Adequate and informative data regarding thepotential adverse effects of MPH, DEX and ATXare lacking. Overall, higher dosages of IR-MPHappear to be associated with the occurrence ofheadache, lost appetite, stomach ache andinsomnia compared with placebo. ER-MPHappears to be associated with decreased appetiteand increased insomnia. However, a previous


xv


Executive summary

xvi

systematic review highlighted the need for furtherresearch into somatic complaints, which may beassociated with the disorder itself rather thanmethylphenidate treatment. Similarly, high dosesof DEX appear to be associated with decreasedappetite and increased sleeping problems. ATX ofany dose may impair appetite.

Cost-effectivenessThe review highlighted a number of potentiallimitations in the existing literature. In particular,the review highlighted limitations in estimatingtreatment effectiveness and associated utilityvalues. These limitations may stem from a lack ofavailable data. A new economic model wasdeveloped for this report. Pooling was limited inthe clinical effectiveness review, owing toheterogeneity between trials. However, somedegree of pooling is necessary to proceed with aneconomic model. The issue of heterogeneity wasovercome by basing the base case on trials that aremore similar in terms of how they measure theoutcome of interest. In a series of sensitivityanalyses more trials were included by relaxing thecriterion of similarity in outcome measurement.Data on resource use associated with ADHD in theUK were lacking, and so the model relies onestimates from experts.

Given the lack of available evidence for statisticallysignificant differences in efficacy between thealternative drugs, the results of the economicmodel were largely driven by drug cost, in which there are marked differences. Theeconomic evaluation clearly suggests an optimaltreatment strategy, that is, DEX first-line, followedby IR-MPH for treatment failures, followed by ATX for repeat treatment failures. If DEX is considered not suitable as a first-line therapy,the optimal strategy is IR-MPH first-line, followed by DEX as second-line and ATX again as third-line. For patients contraindicated tostimulants, ATX is preferred to no treatment. For patients in whom a midday dose of medicationis unworkable, ER-MPH is preferred to ATX, andER-MPH12 appears more cost-effective than ER-MPH8.

The model is not without limitations. As identifiedin the clinical effectiveness review, the reporting

of studies was poor, there are few data todiscriminate between the drugs in efficacy oradverse events and there are few data on long-term efficacy and adverse events associated withmedical management of ADHD. The data do not allow discrimination between patients with ADHD in terms of ADHD subtype, age,gender or previous treatment. These caveats mustbe borne in mind when interpreting the modelresults.

ConclusionsThe main conclusions from this report are asfollows:

1. Drug therapy seems to be superior to no drugtherapy.

2. No significant differences between the variousdrugs in terms of efficacy or side effects werefound – mainly owing to lack of evidence.

3. The additional benefits from behaviouraltherapy (in combination with drug therapy) areuncertain.

The main additional feature of the economicmodel is the consideration of costs. Given the lackof evidence for any differences in effectivenessbetween the drugs, the model tends to be drivenby drug costs, which differ considerably.

Research recommendationsFuture trials examining MPH, DEX and ATXshould include the assessment of tolerability andsafety as a priority. Reporting should bestandardised and transparent. Researchers shouldrefer to the CONSORT approach to study design.

Longer term follow-up of individuals participatingin trials could further inform policy makers andhealth professionals. Such data could potentiallydistinguish between these drugs in a clinicallyuseful way.

In addition, research examining whether somaticcomplaints are actually related to drug treatmentor to the disorder itself would be informative.

This review examines the clinical and cost-effectiveness of oral methylphenidate

hydrochloride (MPH) (Ritalin®, Equasym®,Equasym XL®, Concerta® XL), dexamfetaminesulphate (DEX) (Dexedrine®) and atomoxetinehydrochloride (ATx) (Strattera®) in children and

adolescents (under 18 years of age) diagnosedwith attention deficit hyperactivity disorder(ADHD) [including hyperkinetic disorder (HKD)].It includes an update of the existing NationalInstitute for Health and Clinical Excellence(NICE) Report No. 13.1


1


Chapter 1

Aim of the review

Description of underlying healthproblemADHD (including HKD) is defined by the ‘core’signs of inattention, hyperactivity and impulsivity,and is characterised by an early onset. Theestimated prevalence for ADHD in school-agedchildren varies widely (e.g. 3–7%), beingdependent on a number of variables including themethods of ascertainment, the informants, thepopulation sampled, the diagnostic criteriaapplied and the sex of the affected individual. Forexample, prevalence rates may be lower whendiagnosed using the International StatisticalClassification of Diseases and Related HealthProblems (10th edition) (ICD-10) criteria (seebelow) for hyperkinetic disorder. Data onprevalence in adolescence and adulthood arelimited. The disorder is frequently observed ingreater numbers of males than females, with ratiosranging from 2:1 to 9:1 depending on subtypeand setting.2,3

In England and Wales, the prevalence of ADHD inchildren 6–18 years of age has been estimated tobe about 5% (based on mid-1997 estimates).4 TheBritish Child and Adolescent Mental HealthSurvey conducted in 1999 examined a sample of10,438 children and found 2.23% to have ADHD.Of the 5212 males, 3.62% had ADHD comparedwith 0.85% of the 5226 females in the sample(p < 0.001). Overall, 1.41% of the sample wasclassified as ADHD Combined Type, comparedwith 0.67% and 0.16% Inattentive and HyperactiveType, respectively.5

AetiologyResearch has indicated that the aetiology ofADHD is of a multifactorial and complex nature.Family, twin and adoption studies consistentlyimplicate the role of genetic factors. Moleculargenetic studies have focused on dopaminereceptors such as the D4 dopamine receptor gene(DRD4), the mRNA of which appears to play arole in cognitive and emotional functions.

A number of conditions have a very high risk forthe presence of ADHD symptoms, includingneurofibromatosis Type 1,6 epilepsy,7 very lowbirth weight,8,9 chronic tic disorders/Tourette’s

syndrome, developmental coordination disorders10

and high functioning autism.11 In addition,numerous biological environmental risk factorshave been studied, such as diet, prenatal alcoholand nicotine exposure, some of which appear toaccount for selected cases. Many studies providestrong evidence for the importance of psychosocialadversity such as severe marital discord; however,these factors tend to emerge as universalpredictors of a child’s emotional health andadaptive functioning rather than being specificpredictors of ADHD. Although there is no singlepathophysiological profile of ADHD,neurobiological research suggests that both geneticand environmental factors modify the front-subcortical pathways in the brain that controlattention and motor behaviour.12,13

DiagnosisICD-10, published in 1992,14 details the diagnosisof hyperkinetic disorders. Such disorders,characterised by onset during childhood oradolescence, feature both impaired attention andoveractivity evident in more than one situationover a sustained period of time.

According to these guidelines, impaired attentionmanifests itself in sufferers by premature breakingoff from tasks and leaving activities unfinished.Overactivity implies excessive restlessness and isjudged in terms of normal expectations for thesituation, age and IQ of the individual involved.

Published in 2000, the fourth edition of theDiagnostic and Statistical Manual of Mental Disorders(DSM-IV)2 cites five criteria to be considered inthe diagnosis of Attention deficit/hyperactivitydisorder:

A. At least six symptoms of inattention orhyperactivity/impulsivity should have persistedfor at least 6 months to an extent inconsistentwith normal development.

B. Some symptoms must have been presentbefore 7 years of age.

C. Resulting impairment should be evident intwo or more settings.

D. Impairment should be clinically significantwith regard to social, academic oroccupational functioning.


3


Chapter 2

Background

E. Symptoms should not occur exclusively duringthe course of a pervasive developmentaldisorder, schizophrenia or other psychoticdisorder. Neither should they be betterexplained by another mental disorder.

The DSM-IV criteria distinguish between threesubtypes of attention deficit/hyperactivity disorderaccording to the predominant symptom patternfor the previous six months: combined type,predominantly inattentive type and predominantlyhyperactive–impulsive type.

Choice of diagnostic criteria has importantimplications in terms of measuring diseaseprevalence and also making treatment decisions.The ICD-10 criteria are similar to the severecombined type ADHD defined by the DSM-IVcriteria. Hence, diagnoses based on the formercriteria may be fewer. The two contendingapproaches continue to be a matter ofcontroversial discussion.15,16

The internal validity of the ICD-10 and DSM-IVmodels is also a focus of research anddiscussion.17–24 Studies have been carried out, forexample, to compare the impact of applicationacross cultures and age groups. Issues ofprevalence, co-morbidity and aetiology, related tointernal validity, are highlighted below.

Co-morbidity and associated featuresConsiderable overlap exists between hyperkinesisand other patterns of disruptive behaviour such asconduct disorder. The ICD-10 diagnosticguidelines consequently distinguish betweenhyperkinetic conduct disorder and simpledisturbance of activity and inattention.14 The

DSM-IV manual estimates that approximately halfof clinic-referred children with ADHD also haveoppositional defiant disorder (ODD) or conductdisorder (CD).2

Common patterns of behaviour and development,insufficient and unnecessary for diagnosis inthemselves, have also been noted amongstsufferers:

� disinhibition in social relationships� recklessness in situations involving some danger� impulsive flouting of social rules� cognitive impairment� specific motor and language developmental

delays� dissocial behaviour such as temper outbursts.

These associated characteristics often result innegative interactions with peers, school authoritiesand family members and subsequently diminishedself-esteem.2,14

Current service provisionFor the first quarter of 2004, there were 68,000prescribing items for MPH, at a cost of over £2million [Figure 1; Prescription Pricing Authority(PPA) data cover March 1999 to March 2004 inquarterly periods].

For the first quarter of 2004, there were just under10,000 prescribing items for DEX, at a cost of justunder £80,000 (Figure 2; PPA data cover March1999 to March 2004 in quarterly periods). FromDecember 2001 the cost increased considerablywhereas prescribing remained fairly static. The

Background

4

0

10

20

30

40

50

60

70

Mar-99

Sep-99

Mar-00

Sep-00

Mar-01

Sep-01

Mar-02

Sep-02

Mar-03

Sep-03

Mar-04

Quarter to

Item

s (t

hous

ands

)

0

500

1000

1500

2000

2500

NIC

(£, t

hous

ands

)Items NIC (£)

FIGURE 1 Trends in prescribing of and spending on methylphenidate in general practice in England and Wales. Source: PPA. NIC, net ingredient cost.

reason for this is that the majority of theprescribing for dexamfetamine is for the 5-mgtablet (available as the brand Dexedrine) and inOctober 2001 the price of this product doubled,hence the cost increased.

Health, social and education services will alsoincur costs for assessment and follow-up of thesechildren. These costs have been estimated at about£23 million for initial specialist assessment(£21.8 million in England and £1.5 million inWales) and £14 million for follow-up care over1 year (£13.4 million in England and £0.9 millionin Wales). However, some of these costs relate toservices that are already in place, and 100%uptake of medication is unlikely. It is also possiblethat better treatment for children with ADHDcould avoid some health, education and socialcosts in the longer term.

Description of interventionMethylphenidate hydrochloride25

MPH is a mild CNS stimulant with moreprominent effects on mental than other motoractivities. The mode of therapeutic action is notcompletely understood, but one key action is theinhibition of the dopamine transporter, withconsequent magnification of dopamine-medicatedsignalling in areas such as frontal lobe andstriatum.

Ritalin® (Cephalon UK Ltd), Equasym® (CelltechPharmaceuticals Ltd), Equasym XL® (CelltechPharmaceuticals Ltd) and Concerta® XL (Janssen-

Cilag Ltd) are each indicated as part ofcomprehensive treatment programme for ADHDin children (>6 years of age) and adolescents.[Although Tranquilyn® holds a UK MarketingAuthorisation (PL 18153/0001-3), its manufactureappears to have been discontinued. The licenceholder is Laboratorios Rubio SA.]

Ritalin is available in 10-mg tablets, Equasym isavailable in 5-, 10- and 20-mg tablets andConcerta XL is available in 18- and 36-mgprolonged-release tablets. A new form of slow-release MPH that may gain license in the UK isEquasym XL. It is to be an addition to theimmediate-release 5-, 10- and 20-mg Equasymtablets already established.

Ritalin and EquasymAccording to current guidelines, the recommendeddose is 0.2–0.7 mg/kg body weight, given twice orthree times daily, starting with the lowest andincreasing by increments of 5 mg in each dose untila good response is achieved, adverse events appearor the ceiling dose is reached.

Equasym XLMetadate CD® (Celltech Pharmaceuticals Ltd) iscurrently licensed in the USA and, if licensed inthe UK, it is likely to be known as Equasym XL.Metadate CD is administered once daily in themorning, before breakfast. The recommendedstarting dose is 20 mg once daily. Dosage may beadjusted in weekly 20-mg increments to amaximum of 60 mg per day. The manufacturers ofthis drug claim that its effects last through8 hours.


5


0

2

4

6

8

10

12

14

Mar-99

Sep-99

Mar-00

Sep-00

Mar-01

Sep-01

Mar-02

Sep-02

Mar-03

Sep-03

Mar-04

Quarter to

Item

s (t

hous

ands

)

0

10

20

30

40

50

60

70

80

90

NIC

(£, t

hous

ands

)

Items NIC (£)

FIGURE 2 Trends in prescribing of and spending on dexamfetamine in general practice in England and Wales. Source: PPA.

Concerta XLConcerta XL is administered orally once daily inthe morning and must be swallowed whole withthe aid of liquids, and must not be chewed,divided or crushed. Concerta XL may beadministered with or without food. Dosage shouldbe individualised according to the needs andresponses of the patient. Dosage may be adjustedin 18-mg increments to a maximum of 54 mg perday taken once in the morning. In general, dosageadjustment may proceed at approximately weeklyintervals. The manufacturers of this drug claimthat its effects last through 12 hours.26

Dexamfetamine sulphate(dexamphetamine sulphate)DEX is a symphathomimetic amine with centralstimulant and anorectic activity. It is indicated innarcolepsy. It is also indicated for children withrefractory hyperkinetic states under thesupervision of a physician specialising in childpsychiatry.

Dexedrine® (Celltech Pharmaceuticals Ltd) isavailable in 5-mg tablets.

The usual starting dosage for children aged3–5 years is 2.5 mg per day, increased if necessaryby 2.5 mg per day at weekly intervals. For childrenaged ≥ 6 years, the usual starting dose is 5–10 mgper day, increasing if necessary by 5 mg at weeklyintervals. The usual upper limit is 20 mg a dayalthough some older children have needed 40 mgor more for optimal response.

Atomoxetine hydrochloride27

ATX is reported to be a selective noradrenalinereuptake inhibitor. The precise mechanism by

which ATX works on ADHD is not known. It iscurrently licensed in the USA and the UK underthe brand name Strattera® (Eli Lilly & Co. Ltd).The manufacturers claim that it is the first non-stimulant medication approved for the treatmentof ADHD in children ≥ 6 years of age. Inadolescents who have shown a clear benefit fromtreatment, and whose symptoms persist intoadulthood, treatment may be continued intoadulthood. However, starting treatment withStrattera in adults is not appropriate. It isindicated as an integral part of a total treatmentprogramme for patients with ADHD.

Strattera is available in 10-, 18-, 25-, 40- and 60-mg capsules (for oral administration with orwithout food).

In children and adolescents up to 70 kg bodyweight, Strattera should be initiated at a total dailydose of approximately 0.5 mg/kg and increasedafter a minimum of 3 days to a target daily dose ofapproximately 1.2 mg/kg administered either as asingle daily dose in the morning or as evenlydivided doses in the morning and lateafternoon/early evening. The total daily dose inchildren and adolescents should not exceed1.4 mg/kg or 100 mg, whichever is less.

In children and adolescents >70 kg body weight,Strattera should be initiated at a total daily dose of40 mg and increased after a minimum of 3 days toa target total daily dose of approximately 80 mgadministered as stated above. After 2–4 additionalweeks, the dose may be increased to a maximumof 100 mg in patients who have not achieved anoptimal response.

Background

6


7


Chapter 3

Methods for reviewing effectiveness andcost-effectiveness

Search strategyClinical effectivenessDetails of the search strategies are presented inAppendix 1. They aimed to retrieve papersrelating to MPH, DEX and ATX for children withADHD. The strategy was based on that used in theAgency for Healthcare Research and Quality(AHRQ) report28 and used a range of free textterms and subject headings for ADHD. Inaddition, terms were added for methylphenidate,dexamfetamine and atomoxetine, to provide amore focused strategy. No study design filters wereadded to the searches in order to retrieve a rangeof study designs, including systematic reviews(SRs), randomised controlled trials (RCTs) andclinically controlled trials (CCTs), economicevaluations and adverse events. A date restrictionof 1999 onwards was placed on themethylphenidate searches as this review updatesthe report produced by Lord and Paisleypublished in 2000.4 A date restriction of 1997onwards was placed on the searches for DEX inorder to update the AHRQ report (which includeda review of this drug). Research on ATX wassearched for from 1981 onwards. Most of thesearches were conducted in July 2004. Nolanguage restrictions were placed on any of thesearch strategies.

The following databases were searched for relevantpublished literature:

� CINAHL� CENTRAL � Database of Abstracts of Reviews of Effects

(DARE)� EMBASE � Health Technology Assessment (HTA) Database� Health Economic Evaluations Database (HEED)� MEDLINE � NHS Economic Evaluation Database (NHS

EED)� PreMEDLINE� PsycINFO� Science Citation Index (SCI).

Ongoing and recently completed research wasidentified from:

� Controlled Clinical Trials � clinicaltrials.gov� National Research Register (NRR)� ReFeR database.

Conference proceedings were identified by:

� ISI Proceedings: Social Sciences and Humanities � ISI Proceedings: Science and Technology � Inside Conferences.

Reports, dissertations and other grey literaturewere identified by:

� Dissertation Abstracts� System for Information on Grey Literature in

Europe (SIGLE).

A number of evidence based websites weresearched or browsed, including:

� Agency for Healthcare Research and Quality(AHRQ)

� Health Evidence Bulletins Wales� Health Services/Technology Assessment Text

(HSTAT)� National Coordinating Centre for Health

Technology Assessment� National Guideline Clearinghouse� National Institute for Health and Clinical

Excellence (NICE) (published appraisals)� National Horizon Scanning Centre (NHSC)� Scottish Intercollegiate Guidelines Network

(SIGN) Guidelines� Turning Research Into Practice (TRIP+).

Paper resources to be scanned include the latestClinical Evidence and the BNF.29

Additional specific searches for adverse eventsrelated to MPH, DEX and ATX were conductedusing TOXLINE. Searches for economicevaluations were conducted using NHS EED and

HEED. These databases were also searched forresearch relating to the health-related quality oflife (HRQoL) of people with ADHD, but who werenot necessarily taking any of the three drugs.

The bibliographies of any eligible publicationidentified from the above sources were checked foradditional references. Manufacturer and sponsorsubmissions made to NICE were reviewed toidentify any additional studies.

Cost-effectivenessIn addition to identifying relevant papersretrieved from the clinical effectiveness searches,economic evaluations were identified by searchingthe following resources (details of the searchstrategies are presented in Appendix 2):

� Health Economic Evaluations Database (HEED)(Issue: July 2004)Searched: 22 July 2004 on CD-ROM

� NHS Economic Evaluation Database (NHS EED)Searched: 22 July 2004 on CRD’s internaladministration database.

Additionally, HRQoL research was sought bysearching the following resources:

� CINAHL (1982–June week 2, 2004) Searched: 18 June 2004 on OvidWeb athttp://gateway1.uk.ovid.com/ovidweb.cgi

� Database of Abstracts of Reviews of Effects(DARE)Searched 22 June 2004 on CRD’s internaladministration database

� EMBASE (1980–2004 week 11) Searched: 18 June 2004 on OvidWeb athttp://gateway1.uk.ovid.com/ovidweb.cgi

� Health Economic Evaluations Database (HEED)(Issue: June 2004)Searched 22 June 2004 on CD-ROM

� Health Technology Assessment Database (HTA)Searched 22 June 2004 on CRD’s internaladministration database

� MEDLINE (1966–March week 2, 2004) Searched: 18 June 2004 on OvidWeb athttp://gateway1.uk.ovid.com/ovidweb.cgi

� MEDLINE In-process and other non-indexedcitations (18 June 2004)Searched: 22 June 2004 on OvidWeb athttp://gateway1.uk.ovid.com/ovidweb.cgi

� NHS Economic Evaluation Database (NHS EED)Searched 22 June 2004 on CRD’s internaladministration database

� PsycINFO (1967–2004/ June week 1) Searched: 23 June 2004 on WebSPIRS via BIDSat http://www.bids.ac.uk/

� Social Science Citation Index (SSCI) (1981–2004)Searched: 22 June 2004 on ISI Web ofKnowledge via MIMAS at http://wos.mimas.ac.uk/

� Science Citation Index (SCI) (1981–2004) Searched: 22 June 2004 on ISI Web ofKnowledge via MIMAS at http://wos.mimas.ac.uk/

Inclusion and exclusion criteria:clinical effectivenessTwo reviewers independently screened all titlesand abstracts identified in the updated literaturesearch. Full paper manuscripts of anytitles/abstracts that were considered relevant byeither reviewer were obtained where possible. Inaddition, full paper copies of relevant studiespresented in the NICE, AHRQ and CanadianCoordinating Office for Health Technology andAssessment (CCOHTA)30 reports were obtained.The decision to include studies was assessedaccording to criteria presented below. Anydiscrepancies were resolved by consensus and, ifnecessary, a third reviewer was consulted.

Study designOnly RCTs examining MPH, DEX or ATX usedalone, in combination with each other or incombination with non-drug interventions thatwere compared with placebo, with one another inhead-to-head comparisons or with non-druginterventions were included in the review. Studiesthat compared MPH, DEX or ATX with otherdrugs (e.g. Adderall) were included provided thatthey included a placebo group. This was appliedto both efficacy and adverse events data. Inaddition, SRs were included to examine adverseevents data [see the section ‘Inclusion andexclusion criteria for systematic reviews: clinicaleffectiveness’ (p. 9)].

To be included, studies had to be at least 3 weeksin duration (3 weeks per treatment arm inparallel studies and 3 weeks in overall trial lengthfor crossover studies). The reason why this cut-offwas used is because the literature suggests that3 weeks is the minimum length of treatmentchosen by investigators who are examiningclinical outcome. [The literature contains anumber of random-allocation comparisons ofMPH and placebo based either on single-doseadministration or on treatment over a few days.These have been carried out to clarify the modeof action, e.g. the effect of different doses onlaboratory tests, rather than as therapeutic trials,so should not be included in assessments ofclinical value. The effect of medication on

Methods for reviewing effectiveness and cost-effectiveness

8

behaviour is often (not always) apparentimmediately, but the impact on the socialadjustment of the child may well not be apparentin the first days of therapy. We recognise,however, that even 3 weeks is a short period inwhich to examine the effect of a drug intended tomodify a chronic condition.]

InterventionsMPH (Ritalin, Equasym, Equasym XL, ConcertaXL), DEX (Dexedrine) and ATX (Strattera), usedalone or as part of a multi-modal treatmentprogramme (involving other drugs and/or non-drug interventions), were included, provided thateffectiveness or adverse event data were presentedfor the three drugs of interest. Non-druginterventions included any type of psychologicaland behavioural strategies [e.g. cognitivebehavioural therapy (CBT), child or parenttraining, bibliotherapy] and/or nutritionalinterventions.

Studies that compared MPH, DEX or ATX withother drugs not licensed for ADHD in the UKwere excluded from the review, unless the studyalso included a placebo group. In these studies,data for comparisons between MPH, DEX andATX and placebo were extracted.

ParticipantsParticipants included children and adolescents<18 years of age diagnosed with ADHD(including HKD). There was no lower agelimitation (although there was no lower agelimitation for the report, it is noted that MPH isindicated for children >6 years old and ATX isindicated for children ≥ 6 years of age). Thediagnosis must have been made in an explicit way,preferably using either the ICD-10 criteria or theDSM-IV criteria.

Studies that included conditions other than ADHD(e.g. Tourette’s syndrome) or children with ADHDand mental retardation were only included here ifthey reported separate analyses for patients withADHD alone. RCTs excluded from this review thatexamined children with ADHD and other co-morbid conditions are presented in Appendices 3and 4.

OutcomesGiven the large number of outcomes presented inthe literature, the inclusion criteria were restrictedto four broad outcome categories. These efficacymeasures are widely used and, after discussionswith a clinical expert, were recognised as the mostrelevant clinical outcomes.

To be included in the review, trials had to reportresults on one or more of the following:

� core symptoms (including measures ofinattention, hyperactivity, impulsivity)

� quality of life (QoL) (clinical global impressionor overall severity indices were used as a proxyof QoL)

� adverse effects (including loss of appetite,insomnia, headache, stomach ache and weightloss).

Studies that also reported on educationalperformance (including various tests of reading,spelling, mathematics, or by accuracy andproductivity of seatwork tasks) were also includedin the first stage of the screening, but were laterexcluded if this was the only outcome assessed.Studies that only examined tests of psychologicalfunction (e.g. the continuous performance test),measures of depression and/or anxiety or measuresof coexistent problems (including poor peerrelationships, and conduct/oppositional disorder-related outcomes) were not included in the review.However, if the trial examined these outcomes inaddition to one or more of the four presentedabove, it was noted in data extraction tables.

Studies that presented all results in figures withoutpresenting actual numbers, or only significancevalues for comparisons, were excluded from thereview.

PublicationStudies that met the inclusion criteria above, butwere only published as abstracts or as conferencepresentations, were not included in the reviewunless a full paper could be obtained that relatedto the abstract (a list of these excludedpublications is presented in Appendix 3).

As there was no language restriction in the searchstrategy, trials reported in any language wereconsidered for inclusion in the review. However,time limitations resulted in the exclusion of twopapers written in Mandarin (see Appendix 3).

Inclusion and exclusion criteria for systematic reviews: clinicaleffectivenessTo be included, systematic reviews with adverseevents data had to:

� provide evidence of a search for primaryliterature


9


� analyse safety and/or tolerability as a primaryobjective

� examine children and/or adolescents withADHD

� present data by individual drug type: MPH,DEX or ATX.

Of the studies that met our inclusion criteria, wesubsequently chose only to assess those withprimary studies not already included in our review.

The BNF29 was used to provide information onthe side-effect profiles of MPH, DEX and ATX.

Data extraction strategy: clinicaleffectivenessData relating to both study design and qualitywere extracted by one reviewer into an Accessdatabase and independently checked for accuracyby a second reviewer. Any discrepancies wereresolved by consensus and, if necessary, a thirdreviewer was consulted. Data from studies withmultiple publications were extracted and reportedas a single study.

Quality assessment strategy:clinical effectivenessThe quality of the clinical effectiveness studies wasassessed using modified criteria based on CRDReport No. 431 (see Appendix 5). Each study wasassessed by one reviewer and independentlychecked for agreement with a second reviewer.Disagreements were resolved by consensus and, ifnecessary, a third reviewer was consulted.

Analysis strategy: clinicaleffectivenessClinical effectiveness data were reportedseparately for each drug and by the type ofcomparison. Data for MPH were also analysedseparately based on immediate-release orextended-release formulation. For all drugs, thedata were examined by dose. The cut-offs used fordefining low, medium and high for each type ofdrug were those presented by Swanson andcolleagues.32 Data for the outcomes ofhyperactivity (using any scale that measuredhyperactivity specifically), Clinical GlobalImpression (CGI) (as a proxy of QoL) and adverseevents were analysed. Owing to time constraints,

information on other core outcomes and academicperformance were not analysed; however, thisinformation was extracted and is presented in thedata extraction tables (Appendix 12).

Any scale that appeared to have assessed purehyperactivity was included in the analysis. Resultsfrom scales that may incorporate hyperactivity, butalso include other symptoms (for example, theConners’ Abbreviated Rating Scale) were notanalysed but are included in the data extractiontables (Appendix 12). The type/version of the scaleused to assess hyperactivity has been reported aspresented in the original papers. Many differentscales and versions of the same scales (e.g. theConners’ scales) were used in the trials. Owing totheir complexity, no attempt was made to combineresults from data using different scales, differentversions of a scale or scales which may be the samebut have different names.

For crossover studies, the mean and standarddeviation (SD) for each outcome were dataextracted for end of trial data (i.e. baseline datawere not considered). Where possible, we aimed tocalculate mean difference and standard errors forcrossover studies in order to facilitate meta-analysis where possible.33 However, owing to thelack of data information needed to calculate meandifferences in many of the studies, this was notpossible.

For parallel studies, change scores were reportedwhere given, otherwise means and SDs werepresented for end of trial data. In addition, meandifferences with 95% confidence intervals (CIs)were calculated for each study.

For adverse events, self-ratings were reportedwhen used, otherwise parent reports were utilised.Percentages of participants reporting adverseevents were used to calculate numbers of events ineach treatment arm. Where actual numbers ofparticipants included in safety analyses wereunclear, denominators were based on numbers ofparticipants originally randomised to eachtreatment arm. Relative risks (RRs) were examinedwithin predefined subgroups (based on drug,dosage and inclusion of a behaviouralintervention). Data on weight were also analysedin view of recent concerns regarding the effects ofstimulants on growth in children. Meandifferences (with 95% CIs) were calculated for eachstudy. Where results were highly variable, possiblecauses of this were explored in terms ofparticipant age, duration of intervention andmethod of outcome measurement.

Methods for reviewing effectiveness and cost-effectiveness

10


11


Cost-effectivenessHRQoL studies were selected if they containedhealth outcomes data for use in economicevaluations of ADHD. The review of the economicevaluation literature included studies thatcompared two or more ADHD interventions interms of their costs and outcomes and where atleast one drug intervention was assessed.Economic evaluations could includecost–consequence, cost–utility, cost–effectiveness,cost–minimisation and cost–benefit analyses.Economic evaluations reported as conferenceproceedings or abstracts were excluded since thedata they contain may not be complete. A dataextraction form was used to abstract data on alleconomic evaluations selected in the literaturereview. The data extraction form used has beenused in previous Technology Assessment Reviews.

Each economic evaluation selected for theliterature review was quality assessedindependently by two health economists based onan economic evaluation checklist.34 Anydiscrepancies in quality assessment were resolvedthrough reaching a consensus between the twohealth economists involved.

In Chapter 5, a systematic review of the HRQoLand cost-effectiveness literature is conducted,followed by a review of the company submissionsto NICE. Following this, in Chapter 6, a newmodel is constructed, drawing on data obtainedfrom the literature review and the companysubmissions. The methods used to review theliterature, construct the new model and analysethe cost-effectiveness data are described in detailin the relevant sections.

Quantity and quality of researchavailableNumber of studies identified, includedand excludedIn the previous systematic reviews (NICE, AHRQand CCOHTA), 70 papers were identified thatappeared relevant to the current systematic review,and full paper copies were ordered. Of these, 42papers (describing 40 studies) met the inclusioncriteria.

In the updated search, 2515 references wereidentified (following de-duplication) and screenedfor relevance. A total of 423 full papers wereordered for more detailed examination, some ofwhich were obtained by checking references ofrelevant studies (see Figure 3). Of these, 20 RCTsand one SR met the inclusion criteria. Inaddition, four commercial-in-confidence (CIC)papers were included and the MultimodalTreatment Study of ADHD (MTA) trial wasassessed. Overall, this gives a total of 66 studies(40 studies from previous SRs, 21 papers from theupdated search, four CIC papers and the MTAtrial). An additional 52 papers/abstracts related tothe trials described in the 66 papers. These werenot considered to be fully included, but arereferenced in the data extraction tables with thestudies to which they relate. Reasons forexclusions are presented for each of the 115papers from the updated search in Appendix 3and for each of the 28 papers from the previousreviews in Appendix 4.

Quality of studiesThe quality of reporting for each of the includedstudies is presented in Table 1. A key to how poorand good were defined for each question ispresented in Appendix 5. No summary scorecould be tabulated; however, the first two columnsof the table can be used to identify immediatelystudies that reported their study methodologywell. It is acknowledged that the qualityassessments presented here reflect only the qualityof reporting of trials.

The majority of included studies poorly reportedon study methodology. Only five out of the 60non-confidential RCTs adequately reported the

method used to assign participants to patientgroups. Nine out of the 60 non-confidentialstudies reported that the sequence of allocationwas truly random. Most studies were blinded, butnone of the included studies reported whetherblinding was successful. Eleven out of the 60 non-confidential studies reported to have used anintention-to-treat (ITT) analysis, and 18 out of the60 non-confidential studies provided a completedescription of withdrawals. Of 64 RCTs, 35 werecrossover trials and 29 were parallel trials. Formost of the crossover trials, the statistical analysiswas either not clearly reported or not appropriate;for parallel trials, the quality of the analysis wasbetter reported. [Confidential informationremoved from this paragraph].

Trials included in the reviewAn overview of comparisons made in each of thetrials is presented in Table 2 (information fromfour trials reported as CIC is omitted).

Assessment of effectivenessAs noted above, the clinical effectiveness data arereported separately for each drug (by dose) and bythe type of comparison. This was done in order toreproduce accurately the data/results as reportedin the original papers. The cut-offs used fordefining low, medium and high for each type ofdrug were those presented by Swanson andcolleagues.32 Data for MPH were also analysedseparately based on immediate-release orextended-release formulation.

Where possible, data on hyperactivity specificallyare presented in separate tables. The scales usedto measure this core symptom are as reported inthe trials, hence the level of detail regarding whatscale was used may vary.

MPH versus placeboMPH low dose (≤ 15 mg/day) versus placeboTwelve studies examined low-dose (≤15 mg/day)immediate-release MPH compared with placebo(Table 3, with additional information presented inAppendix 12). Of these, three examined MPHadministered once daily and nine examined MPHadministered twice or three times daily.


13


Chapter 4

Clinical effectiveness

MPH administered once dailyOf the studies that examined MPH administeredonce daily, only one examined hyperactivity(discussed below).96 The two other studies bothused the Abbreviated Conners’ Teacher RatingScale (CTRS) total score to evaluate low-doseMPH.49,85 In the study by Rapport andcolleagues85 three groups of children wererandomised: a low-weight group (22–26 kg), a mid-

weight group (27–31 kg) and a high-weight group(32–36 kg). All three MPH low doses (5, 10 and15 mg/day) resulted in statistically significantimprovements compared to placebo for all weightgroups (p < 0.01). Similarly, in the study by DuPauland Rapport,49 all low-dose MPH treatment groups(5, 10 and 15 mg/day) were significantly betterthan placebo for Abbreviated CTRS total score butthe level of significance was not clear.


14

References identified:n = 5718

Unavailable/not received:n = 16

Previously included papers identified:n = 70

RCTs of clinical effectiveness meetinginclusion criteria from previous

reviews: n = 40[42 papers]

Studies meeting inclusion criteria:n = 20 RCTs plus n = 1 SR

In addition, the MTA trial was discussed in the review(with 20 papers/abstracts referring to this trial) and

33 papers were identified that related to all the trials includedin the review

Excluded papers:n = 28

Industry submission:n = 4

Excluded (first screen):Background or commentary: n = 60

Excluded: n = 158

Total n = 218

Full papers ordered:n = 423

References identified (following de-duplication) and screened:n = 2515

RCTs of clinical effectiveness: n = 65 (20 plus MTA plus 40 plus 4 industry)SRs of adverse events: n = 1

Excluded (second screen):(RCTs)

Alternative condition: n = 1Co-morbid condition: n = 20Unlicensed comparator: n = 3Irrelevant outcomes: n = 23

Inadequate data presentation: n = 8Abstract only: n = 33

Inadequate trial duration: n = 8Translation required: n = 2

Delayed receipt of paper: n = 4

(Systematic reviews)

Primary studies assessed for inclusion: n = 7Inappropriate outcomes/comparators: n = 1Inappropriate outcomes/population: n = 1

Inadequate data presentation: n = 4Total: n = 115

FIGURE 3 Process of study selection for clinical effectiveness


15


TABLE 1 The quality of included studies

Study

Ahmann, 199335 Poor Poor Yes P/#2 Unclear No No Yes YesArnold, 197636 Good Good Yes NR Unclear Unclear Unclear Yes YesArnold, 197837 Good Good Yes CL/#2 Unclear Unclear Unclear Yes YesArnold, 198938 Good Poor Yes NR Unclear Unclear Unclear Unclear YesBarkley, 199039 Poor Poor Yes C/P/T/I Unclear Unclear Yes Yes? NoBarkley, 200040 Poor Poor Yes C/P/T/I Unclear No Yes No NoBrown, 198541 Poor Poor NR NA Unclear NA NA Yes NoBrown, 198642 Poor Poor Yes P/T/I Unclear Unclear Partially Yes NoBrown, 198843 Poor Poor Yes NR Unclear Unclear Unclear Yes NoBuitelaar, 199644 Poor Poor Yes NR Unclear Yes NA Unclear NoConners, 197245 Poor Poor NR NA Unclear Unclear No Yes NoConners, 198046 Poor Poor Yes C/P/I Unclear Unclear Unclear Yes NoConrad, 197147 Poor Poor Yes P/T/CL/I Unclear Unclear No Yes NoDopfner, 200348 Poor Poor Yes NR Unclear No Yes Yes NoDuPaul, 199349 Poor Poor Yes T/I Unclear No No Poor NoEfron, 199750 Poor Poor Yes C/P/T/I Unclear Unclear Unclear Unclear NoElia, 199151 Poor Poor Yes NR Unclear Unclear Unclear No NoFine, 199352 Poor Poor Yes C/P/T/I Unclear NA NA No YesFirestone, 198653 Poor Poor Yes P/T/CL Unclear Unclear Partially Yes? NoFischer, 199154 Poor Poor Yes C/P/T/I Unclear Unclear Yes Unclear NoFitzpatrick, 199255 Good Poor Yes #1 Unclear Unclear Unclear Partially NoGillberg, 199756 Poor Poor Yes C/P Unclear Yes Partially Yes NoGittelman-Klein, 197657 Poor Poor Yes NR Unclear Unclear Yes No YesGreenberg, 197258 Poor Poor Yes NR Unclear Unclear Partially Yes? YesGreenhill, 200259 Poor Poor Yes NR Unclear Yes/LOCF Partially Yes YesHanden, 199960 Poor Poor Yes #6 Unclear Unclear Yes No NoHoeppner, 199761 Poor Poor Yes NR Unclear Unclear Unclear No NoJames, 200162 Poor Good Yes NR Unclear Unclear NA No YesKelsey, 200463 Poor Poor Yes NR Unclear Yes Yes Yes YesKemner, 200464 [Confidential information removed]Klein, 199765 Poor Poor Yes #4 Unclear Unclear Partially Yes NoKlorman, 198766 Poor Poor Yes NR Unclear Unclear Unclear Poor NoKlorman, 199067 Poor Poor Yes NR Unclear Unclear Unclear No NoKlorman, 199468 Poor Poor Yes I/#2 Unclear Unclear Yes Unclear NoKolko, 199969 Poor Poor Yes #5 Unclear No Yes Unclear NoKratochvil, 200270 Poor Poor No NA Yes/LOCF Yes Yes YesKupietz, 198871 Poor Poor Yes P/T/I Unclear Unclear Partially Yes NoManos, 199972 Poor Poor Yes #3 Unclear Unclear NA No NoMichelson, 200173 Good Good Yes NR Unclear Unclear Yes Yes YesMichelson, 200274 Poor Poor Yes C/CL/I Unclear Yes/LOCF Partially Yes YesMichelson, 200475 Poor Good Yes C/I Unclear Yes? Partially Yes YesMTA, 199976 Poor Good Yes #7 Unclear Yes Partially Unclear NoPelham, 198777 Poor Poor Yes I/#2 Unclear Unclear Unclear Yes NoPelham, 199078 Poor Poor Yes NR Unclear Unclear Unclear Unclear NoPelham, 199379 Poor Poor Yes NR Unclear Unclear Unclear Unclear NoPelham, 199980 Poor Poor Yes P/T/I/CL Unclear Unclear Yes Unclear Yes

continued

Was

the

met

hod

used

to

assi

gnpa

rtic

ipan

ts t

o th

e tr

eatm

ent

grou

ps r

eally

ran

dom

?

Was

the

seq

uenc

e of

rand

omis

atio

n co

ncea

led?

Was

blin

ding

car

ried

out

?

Who

was

blin

ded?

Was

blin

ding

suc

cess

ful?

Was

an

ITT

ana

lysi

s pe

rfor

med

?

Was

a c

ompl

ete

desc

ript

ion

of a

nyw

ithd

raw

als

give

n?

Was

the

sta

tist

ical

ana

lysi

sap

prop

riat

e?

Was

an

asso

ciat

ion

wit

h in

dust

ryre

port

ed b

y th

e au

thor

s?

MPH administered two or more times dailyOf the studies that examined MPH administeredmore than once daily, five examined hyperactivityas a core outcome measure. Results from thesestudies are presented in Table 4 by age group andare described below. The remaining four studiespresented in Table 3 did not examine hyperactivity,but did measure other core outcomes (with theexception of Fine and Johnston52 who reportedresults for adverse events only). These studiesinclude that of Manos and colleagues72 whereoutcome measures included the AbbreviatedSymptoms Questionnaire (ASQ) as measuredseparately by parents and teachers, and the ADHD

rating scale (parents). However, Manos andcolleagues did not report data separately for low-and medium-dose MPH (see Appendix 12). Henceany results comparing low-dose MPH versusplacebo cannot be extracted from this study.

In the study by Barkley and colleagues,40 fivetreatment arms were examined including low-doseMPH and a placebo group. The main coreoutcome examined was ADHD total ratings asevaluated by parents, and teachers (mathematicsand English teachers). Overall statistical analysesresulted in no significant differences between thetreatment arms.


16

TABLE 1 The quality of included studies (cont’d)

Study

Pelham, 199981 Poor Poor Yes C/P/T/CL Unclear Unclear Unclear Yes? YesPelham, 200182 Poor Poor Yes NR Unclear Unclear Yes Unclear YesPliszka, 200083 Poor Poor Yes P/T/CL Unclear Yes Yes Unclear YesQuinn, 200384 [Confidential information removed]Rapport, 198985 Poor Good Yes C/P/T/I Unclear No No Poor NoSchachar, 199786 and Good Good Yes C/P/T/I Unclear Unclear Yes Yes No

Diamond, 199987

Smith, 199888 Poor Poor Yes NR Unclear Unclear Yes Unclear NoSpencer, 200289 Good Good Yes C/P/I Unclear Unclear Partially Yes YesSteele, 200490 [Confidential information removed]Stein, 199691 Poor Good Yes C/P/T/I Unclear Unclear Yes Yes NoStein, 200392 Poor Poor Unclear NR Unclear Unclear Unclear Unclear NoSwanson, 200432 Poor Poor Yes NR Unclear Yes No Good? YesTervo, 200293 Poor Poor Yes C/P/I Unclear No Unclear No NoWeiss, 200494 [Confidential information removed]Wernicke, 200495 Poor Poor Yes NR Unclear Unclear No Unclear YesWerry, 198096 Poor Poor Yes NR Unclear Unclear Unclear Poor NoWolraich, 200197 Poor Poor Yes NR Unclear Yes/LOCF Yes Yes YesZeiner, 199998 Poor Poor Yes All raters Unclear NA NA Yes No

C, children; CL, clinicians/psychiatrists/therapists/counsellors; I, investigators/staff/outcome assessors/clinical assistant; LOCF,last observation carried forward; NA, not applicable; NR, not reported; P, parents/families; T, teachers; #1, some? raters,parents and ?; #2, others unclear; #3, clinicians and parents were not blind to allocated medication, but together withteachers were blind to dosage level; #4, psychologists were blind to all treatment conditions. Psychiatrists and parents wereblinded to medication type for those subjects allocated to behavioural therapy arms. Teachers were blind to type ofmedication, and study design. All classroom observers were blind to study design, and six out of seven were also blind tothe type of children under study and the purposes of observation; #5, nurse, programme staff, staff, students, and parentswere blind to dosages and schedules; #6, states that ‘Preschool staff, patients, laboratory staff were unaware that lowerMPH dose preceded higher MPH dose’; 7, classroom observers, peers, laboratory raters, clinicians selecting ‘best dose’.

Was

the

met

hod

used

to

assi

gnpa

rtic

ipan

ts t

o th

e tr

eatm

ent

grou

ps r

eally

ran

dom

?

Was

the

seq

uenc

e of

rand

omis

atio

n co

ncea

led?

Was

blin

ding

car

ried

out

?

Who

was

blin

ded?

Was

blin

ding

suc

cess

ful?

Was

an

ITT

ana

lysi

s pe

rfor

med

?

Was

a c

ompl

ete

desc

ript

ion

of a

nyw

ithd

raw

als

give

n?

Was

the

sta

tist

ical

ana

lysi

sap

prop

riat

e?

Was

an

asso

ciat

ion

wit

h in

dust

ryre

port

ed b

y th

e au

thor

s?


17


TA

BLE

2An

ove

rvie

w o

f tria

ls in

clud

ed in

the

revie

w –

and

com

para

tors

Stud

y

Age

(ye

ars)

IQ

Ahm

ann,

199

3355.

0–15

NR

××

×A

rnol

d, 1

97636

4.6–

12N

R×

××

Arn

old,

197

8375.

0–12

NR

××

×A

rnol

d, 1

98938

6.0–

12N

R×

××

Bark

ley,

199

0395.

0–13

105

××

×Ba

rkle

y, 2

00040

12.0

–17

104

××

×2

Brow

n, 1

98541

6.3–

11.8

102

××

×Br

own,

198

6425.

7–13

.197

××

××

Brow

n, 1

98843

13–1

593

××

××

Buite

laar

, 199

6446.

0–13

93×

××

Con

ners

, 197

2456.

0–12

NR

××

×C

onne

rs, 1

98046

6.0–

11N

R×

××

Con

rad,

197

1474.

0–6

NR

××

××

Dop

fner

, 200

3486.

0–16

104

××

DuP

aul,

1993

496.

0–11

NR

×3

×Ef

ron,

199

7505.

0–14

.999

××

Elia

, 199

1516.

0–12

106

××

×Fi

ne, 1

99352

6.0–

10N

R×

××

Fire

ston

e, 1

98653

5.0–

9N

R×

××

Fisc

her,

1991

542.

4–17

.2N

R×

××

Fitz

patr

ick,

199

2556.

9–11

.511

4×

××

×G

illbe

rg, 1

99756

6.0–

11N

R×

×G

ittel

man

-Kle

in, 1

97657

6.0–

12N

R×

××

×G

reen

berg

, 197

2586.

5–11

85×

××

×G

reen

hill,

200

2595.

0–15

NR

××

cont

inue

d

Placebo/st. comm. care/attent. contr.Non-drug (+ placebo)MPH low doseMPH medium doseMPH high doseMPH low dose + non-drugMPH medium dose + non-drugMPH high dose + non-drugMPH-SR low doseMPH-SR medium doseMPH-SR high doseMPH-SR low dose + non-drugMPH-SR medium dose + non-drugMPH-SR high dose + non-drugMPH low dose + MPH-SR 20 mgMPH + thioridazine hydrochlorideDEX low doseDEX medium doseDEX high doseDEX medium dose + non-drugDEX high dose + non-drugDEX-TR/ERDEX-SR + non-drugATX L+MATX HCaffeinePemolinePemoline + non-drugImipramineAdderallAdderall + non-drugMAOI + dietGLA 500 mg + vitamin EChlorpromazineHydroxyzineLevoamphetamineThioridazine hydrochloridePindolol


18 TA

BLE

2An

ove

rvie

w o

f tria

ls in

clud

ed in

the

revie

w –

and

com

para

tors

(co

nt’d

)

Stud

y

Age

(ye

ars)

IQ

Han

den,

199

9604.

0–5.

160

××

×H

oepp

ner,

1997

616.

1–18

.2N

R×

××

Jam

es, 2

00162

6.9–

12.2

100

××

××

Kelse

y, 2

00463

6.0–

12N

R×

×Ke

mne

r, 20

0499

[Con

fiden

tial

info

rmat

ion

rem

oved

]K

lein

, 199

7656.

0–12

NR

××

×K

lorm

an, 1

98766

12.0

–19

101

××

Klo

rman

, 199

06712

.0–1

810

9×

×K

lorm

an, 1

99468

5.6–

11.9

109

××

Kolk

o, 1

99969

7.0–

13N

R×

××

××

×K

rato

chvi

l, 20

0270

7.0–

15N

R×

×Ku

piet

z, 1

98871

7.0–

1394

×2

×M

anos

, 199

9725.

0–17

NR

××

2M

iche

lson,

200

1738.

0–18

NR

×2

×M

iche

lson,

200

2746.

0–16

NR

××

Mic

helso

n, 2

00475

6.0–

15N

R×

×M

TA, 1

99976

7–9.

910

1×

××

×Pe

lham

, 198

7776.

5–11

95×

××

Pelh

am, 1

99078

8.1–

13.2

106

××

××

×Pe

lham

, 199

3795.

4–9.

911

1×

××

××

×Pe

lham

, 199

9805.

8–12

.7N

R×

××

2Pe

lham

, 199

9816.

0–12

110

××

23

Pelh

am, 2

00182

6.0–

1210

5×

××

Plisz

ka, 2

00010

06.

0–11

NR

××

×

cont

inue

d



19


TA

BLE

2An

ove

rvie

w o

f tria

ls in

clud

ed in

the

revie

w –

and

com

para

tors

(co

nt’d

)

Stud

y

Age

(ye

ars)

IQ

Qui

nn, 2

00384

[Con

fiden

tial

info

rmat

ion

rem

oved

]Ra

ppor

t, 19

8985

5.0–

1510

0×

3×

Scha

char

, 199

7866.

0–12

108

××

Smith

, 199

88812

.0–1

710

1×

×2

Spen

cer,

2002

897.

0–13

105

××

Stee

le, 2

00490

[Con

fiden

tial

info

rmat

ion

rem

oved

]St

ein,

199

6916.

0–12

NR

×3

Stei

n, 2

00392

5.9–

1610

7×

××

×Sw

anso

n, 2

00432

6.0–

12N

R×

22

2Te

rvo,

200

2939.

9 (2

.9)

NR

××

×W

eiss

, 200

494[C

onfid

enti

al in

form

atio

n re

mov

ed]

Wer

nick

e, 2

00495

7.0–

12N

R×

×W

erry

, 198

0965.

5–12

.5N

R×

××

Wol

raic

h, 2

00197

6.0–

12N

R×

××

Zei

ner,

1999

987.

0–12

102

××

NR,

not

rep

orte

d; S

R, s

ingl

e re

leas

e, E

R, e

xten

ded

rele

ase.



20 TA

BLE

3M

PH lo

w d

ose

(≤15

mg/

day)

ver

sus

plac

ebo

Stud

yD

esig

nIn

terv

enti

on –

NA

ge

Dur

atio

n C

ore

outc

omes

(yea

rs)

(wee

ks)

Adm

inist

ered

onc

e da

ilyRa

ppor

t, 19

8985

C (5

×)M

PH (5

mg/

day,

o.d

.) –

455–

125

Cor

e: n

o hy

p; A

bbre

viat

ed C

TRS

: tot

al s

core

MPH

(10

mg/

day,

o.d

.) –

45Q

oL: n

ot r

epor

ted

MPH

(15

mg/

day,

o.d

.) –

45A

E: n

ot r

epor

ted

DuP

aul,

1993

49C

(5×)

MPH

(5 m

g/da

y, o

.d.)

– 31

6–11

6C

ore:

No

hyp;

Abb

revi

ated

CT

RS: t

otal

sco

reM

PH (1

0 m

g/da

y, o

.d.)

– 31

QoL

: not

rep

orte

dM

PH (1

5 m

g/da

y, o

.d.)

– 31

AE:

not

rep

orte

d

Wer

ry, 1

98096

C (3

×)M

PH (0

.40

mg/

kg, o

.d.)

– 30

5.5–

12.5

4C

ore:

Con

ners

’ Tea

cher

Que

stio

nnai

re: h

yper

activ

ity;

Con

ners

’ Par

ent

Que

stio

nnai

re: h

yper

activ

ityQ

oL: C

GI (

phys

icia

n)A

E: w

eigh

t

Adm

inist

ered

tw

o or

mor

e tim

es d

aily

Brow

n, 1

98843

C (4

×)M

PH (8

.76

mg/

day,

b.d

.) –1

113

–15

8C

ore:

CPR

S: H

yper

activ

ity In

dex;

Con

ners

’ Tea

cher

Hyp

erac

tivity

Inde

x; A

CTe

RS: h

yper

activ

ityQ

oL: n

ot r

epor

ted

AE:

SER

S (p

aren

ts);

wei

ght

Fisc

her,

1991

54C

(3×)

MPH

(0.4

0 m

g/kg

/day

, b.d

.) –

161

2.4–

17.2

3C

ore:

CPR

S-R:

Hyp

erac

tivity

Inde

x; C

TRS

-R: h

yper

activ

ityin

dex;

CT

RS-R

: hyp

erac

tivity

QoL

: not

rep

orte

dA

E: C

PRS-

R: p

sych

osom

atic

; SER

S (p

aren

ts, t

each

ers)

:nu

mbe

r of

sid

e-ef

fect

s, m

ean

seve

rity

ratin

g

Fitz

patr

ick,

199

255C

(4×)

MPH

(10–

15 m

g/da

y, b

.d.)

– 19

6.9–

11.5

8C

ore:

Con

ners

’ Hyp

erac

tivity

Inde

x (p

aren

ts a

nd t

each

er);

TO

TS:

hyp

erac

tivity

(par

ents

and

tea

cher

s)Q

oL: n

o C

GI;

com

men

ts r

atin

gs (p

aren

t/te

ache

r)A

E: S

TES

S (p

aren

ts);

wei

ght

Fine

, 199

352C

(3×)

MPH

[0.3

0 m

g/kg

/day

(unc

lear

), b.

d.] –

12

6–10

3C

ore:

not

rep

orte

dQ

oL: n

ot r

epor

ted

AE:

sid

e-ef

fect

s qu

estio

nnai

re

Hoe

ppne

r, 19

9761

C (3

×)M

PH (0

.30

mg/

kg/d

ay, b

.d.)

– 50

6.1–

18.2

4C

ore:

CPR

S: H

yper

activ

ity In

dex;

CT

RS: H

yper

activ

ityIn

dex

QoL

: not

rep

orte

dA

E: n

ot r

epor

ted

cont

inue

d


21


TA

BLE

3M

PH lo

w d

ose

(≤15

mg/

day)

ver

sus

plac

ebo

(con

t’d)

Stud

yD

esig

nIn

terv

enti

on –

NA

ge

Dur

atio

n C

ore

outc

omes

(yea

rs)

(wee

ks)

Han

den,

199

960C

(3×)

MPH

(12–

15 m

g/da

y, m

ax. 3

×) –

11

4–5.

13

Cor

e: C

TRS

: Hyp

erac

tivity

Inde

x; C

TRS

: hyp

erac

tivity

QoL

: not

rep

orte

dA

E: S

ide

Effe

cts

Che

cklis

t (t

each

ers,

par

ents

); m

ean

seve

rity

ratin

g 0–

6

Man

os, 1

99972

C (4

×)M

PH (1

0 m

g/da

y, b

.d.)

– 42

5–17

4C

ore:

no

hyp;

ASQ

(par

ents

and

tea

cher

s); A

RS (p

aren

t)Q

oL: n

o C

GI;

com

posit

e ra

tings

(clin

icia

n)A

E: S

ide

Effe

cts

Beha

viou

r M

onito

ring

Scal

e (p

aren

ts)

Bark

ley,

200

040C

(5×)

MPH

(10

mg/

day,

b.d

.) –

3812

–17

5C

ore:

no

hyp;

AD

HD

Tot

al P

aren

t/Te

ache

r ra

ting

QoL

: not

rep

orte

dA

E: n

umbe

r an

d se

verit

y of

sid

e-ef

fect

s (t

each

ers,

pare

nts,

sel

f)

Terv

o, 2

00293

C (3

×)M

PH (0

.10

mg/

kg/d

ay, b

.d.)

– 41

M=

9.9

(2.9

)3

Cor

e: n

o hy

p; C

BCL

(par

ent)

QoL

: not

rep

orte

dA

E: n

ot r

epor

ted

AC

TeRS

, AD

D-H

Com

preh

ensiv

e Te

ache

rs’ R

atin

g Sc

ale;

AE,

adv

erse

effe

cts;

ARS

, AD

HD

Rat

ing

Scal

e; A

SQ, A

bbre

viat

ed S

ympt

oms

Que

stio

nnai

re; b

.d.,

twic

e da

ily; C

, cro

ssov

ertr

ial (

num

ber

of c

ross

over

s); C

BCL,

Chi

ld B

ehav

iour

Che

cklis

t; C

GI,

Clin

ical

Glo

bal I

mpr

essio

n; C

PRS,

Con

ners

’ Par

ent

Ratin

g Sc

ale;

CT

RS, C

onne

rs’ T

each

er R

atin

g Sc

ale;

o.d

., on

ceda

ily; P

, par

alle

l tria

l; hy

p, h

yper

activ

ity; P

AC

S, P

aren

tal A

ccou

nt o

f Chi

ldho

od S

ympt

oms;

SER

S, S

ide

Effe

cts

Ratin

g Sc

ale.

Tervo and colleagues93 examined a low-dose MPHgroup and a placebo group (in addition to amedium-dose MPH group). The main outcomeexamined was Child Behaviour Checklist (CBCL)as rated by parents. Direct statistical comparisonswith placebo were not reported, although theauthors found a significant linear response tomedication (p = 0.001).

HyperactivityAll of the six studies that examined hyperactivityused a Conners’ scale (see Table 4). Five of thesestudies reported results separately for parents andteachers,43,54,55,61,96 and one reported results forteachers only.60 One of these studies also assessedhyperactivity using the ADD-H ComprehensiveTeacher’s Rating Scale (ACTeRS),43 and anotherstudy assessed hyperactivity using the Loney’sTime on Task Scale (TOTS).55

All of the above studies used a crossover design.They were not combined using meta-analysisbecause there appeared to be clinicalheterogeneity between some of the studies (e.g.the age groups varied) and because certain datawere not available that would facilitate meta-analysis of crossover trials. For instance, thesignificance values for paired analyses were notreported – evidence which is necessary to estimatea mean difference and standard error for eachstudy.

Overall, three studies reported that low-dose MPHwas better than placebo,54,55,61 and three reportedno significant difference between the groups.43,60,96

It is noted, however, that the study by Handen andcolleagues60 was conducted in younger children(4–5 years age) with a low mean IQ (60) and maynot be comparable to the other studies.


22

TABLE 4 Results for hyperactivity [MPH low-dose (≤ 15 mg/day) versus placebo]

Study Scale MPH low dose: Placebo: p-Value mean (SD) mean (SD) (if reported)a

2–17 years oldFischer, 199154 CPRS-R (Hyperactivity Index) 11.7 (6.4) 14.3 (6.8) S – NR

CTRS-R (Hyperactivity Index) 9.9 (6.6) 13.7 (7.6) S – NRCTRS-R (hyperactivity) 7.1 (5.5) 9.5 (5.8) S – NR

4–5 years oldHanden, 199960 CTRS (hyperactivity) 9.0 (5.1) 14.0 (3.7) NS

CTRS (Hyperactivity Index) 11.9 (5.7) 17.4 (6.0) NS

5–12 years/6–12 years oldWerry, 198096 Conners’ Teacher Questionnaire 2.22 (NR) 2.56 (NR) NS

Conners’ Parent Questionnaire 1.29 (NR) 1.27 (NR) NS

Fitzpatrick, 199255 (? Mean, ? SD) (? Mean, ? SD)Conners’ Hyperactivity Index (parents) 0.96 (0.50) 1.75 (0.67) <0.006Conners’ Hyperactivity Index (teacher) 0.73 (0.65) 1.36 (0.80) S – NR

Mean (SD) Mean (SD)TOTS (hyperactivity) (parents) 0.20 (0.31) 0.70 (0.48) S – NRTOTS (hyperactivity) (teachers) 0.16 (0.44) 0.36 (0.69) S – NR

6–18 years oldHoeppner, 199761 CPRS (Hyperactivity Index) 7.91 (7.21) 8.40 (6.59) S – NR

CTRS (Hyperactivity Index) 8.48 (7.42) 13.54 (8.66) S – NR

13–15 years oldBrown, 198843 CPRS-R (Hyperactivity Index) 9.33 (4.32) 12.66 (4.13) NS

Conners’ Teacher Hyperactivity Index 22.16 (3.12) 24.50 (2.81) NSACTeRS (hyperactivity) 13.66 (6.97) 8.00 (0.63) NS

a Note that owing to the overall poor reporting of study methodology, p-values should be interpreted with caution. Lower scores represent better behavioural outcome.ACTeRS, ADD/H Comprehensive Teacher Rating Scale; CPRS, Conners’ Parent Rating Scale; CPRS-R, Conners’ ParentRating Scale – Revised; CTRS, Conners’ Teacher Rating Scale; CTRS-R, Conners’ Teacher Rating Scale – Revised; NS, not significant; S – NR: significant (value not reported); TOTS, Loney’s Time on Task Scale.

Quality of lifeOnly one of the 12 studies examined physician-rated GGI (used as a proxy for QoL in this SR). This study, conducted by Werry andcolleagues,96 reported no significant differencebetween the MPH and placebo groups for thisoutcome.

Two other studies reported outcomes that could also be used to indicate QoL: Fitzpatrickand colleagues,55 presented parent and teacher comments ratings and Manos andcolleagues72 reported composite ratings asmeasured by a clinician (see Appendix 12 for results).

Adverse eventsOf the 12 studies comparing low-dose MPH withplacebo, only two presented usable data on adverseevents.55,60 The occurrence of headache was notsignificantly different between treatment arms ineither trial [relative risk (RR) = 1.00; 95% CI 0.16to 6.38; and RR = 3.00; 95% CI 0.14 to 66.53,respectively]. With regard to loss of appetite,neither study detected differences between thetreatment arms (RR = 3.00; 95% CI to 0.34 to26.33; and RR = 5.00; 95% CI 0.69 to 36.13,respectively). Similarly, incidence of stomach achedid not appear to differ between participantsassigned to low doses of MPH and those receivingplacebo (RR = 3.00; 95% CI 0.13 to 69.31; andRR = 3.00; 95% CI: 0.14 to 66.53). One trial55

reported the occurrence of insomnia which wasnot significantly different between treatment arms(RR = 2.67; 95% CI 0.83 to 8.55). Weight datawere not adequately reported in any trial.

SummaryIn summary, there seems to be variation in theresults for low-dose MPH compared with placebo.There were no differences in adverse events forboth groups. These studies did not score very wellin the quality assessment, and the results shouldbe interpreted with caution.

MPH medium dose (15–30 mg/day) versusplaceboTwenty-one studies examined medium-dose(15–30 mg/day) immediate-release MPHcompared with placebo (Table 5; with additionalinformation presented in Appendix 12). Twostudies examined medium-dose MPHadministered once daily and 19 examined MPHadministered two or more times daily.

MPH administered once dailyBoth Rapport and colleagues85 and DuPaul andRapport49 evaluated the effectiveness of medium-dose MPH administered once daily using theAbbreviated CTRS total score as a main outcomemeasure. Both studies reported a significantimprovement in the treatment group comparedwith placebo (p < 0.01).

MPH administered two or more times dailyOf the 19 studies that examined MPH administeredmore than once daily, nine measured hyperactivityand will be discussed below. In addition, fourstudies35,39,44,52 reported data only for adverseevents. These are also discussed separately below.

The first of the six remaining studies is that byManos and colleagues.72 In this study, twomedium doses of MPH were evaluated: 20 and30 mg/day. However, as reported above, Manosand colleagues did not report data separately forlow-dose MPH and the medium doses (seeAppendix 12), hence any results comparingmedium-dose MPH versus placebo cannot beextracted from this study. In their results, theystated that ‘best dose’ was better than placebo.

Kolko and colleagues69 presented results forinattention/overactivity using theInattention/Overactivity with Aggression (IOWA)CTRS. They reported that medium-dose MPHwas better than placebo (p < 0.0001). This scalewas also used by Pliszka and colleagues,83 wheremedium-dose MPH was also observed to improvebehaviour compared with placebo (p < 0.05), andby Pelham and colleagues,79 where results showedimprovement with treatment (see Appendix 12).

In the study by Barkley and colleagues,40 fivetreatment arms were examined, includingmedium-dose MPH and a placebo group. Themain core outcome examined was ADHD totalratings as evaluated by parents and teachers(mathematics and English teachers). Overallstatistical analyses resulted in no significantdifferences between the treatment arms.

Tervo and colleagues 200293 also compared amedium-dose MPH group and a placebo group. A main outcome examined was CBCL as rated byparents. Direct statistical comparisons with placebowere not reported, although the authors found asignificant linear response to medication(p = 0.001).


23



24 TA

BLE

5M

PH m

ediu

m d

ose

(15–

30 m

g/da

y) v

ersu

s pl

aceb

o

Stud

yD

esig

nIn

terv

enti

on –

NA

ge

Dur

atio

n C

ore

outc

omes

(yea

rs)

(wee

ks)

Adm

inist

ered

onc

e da

ilyRa

ppor

t, 19

8985

C (5

×)M

PH (2

0 m

g/da

y, o

.d.)

– 45

5–12

5C

ore:

no

hyp;

AC

TeRS

: tot

al s

core

QoL

: not

rep

orte

dA

E: n

ot r

epor

ted

DuP

aul,

1993

49C

(5×)

MPH

(20

mg/

day,

o.d

.) –

316–

116

Cor

e: n

o hy

p; A

CTe

RS: t

otal

sco

reQ

oL: n

ot r

epor

ted

AE:

not

rep

orte

d

Adm

inist

ered

tw

o or

mor

e tim

es d

aily

Con

ners

, 198

046P

MPH

(22.

0 m

g/da

y, b

.d.)

– 60

6–11

8C

ore:

Con

ners

’ Par

ent

Que

stio

nnai

re: h

yper

activ

ity;

Con

ners

’ Tea

cher

Que

stio

nnai

re: h

yper

activ

ityQ

oL: n

o C

GI;

pare

nt g

loba

l jud

gem

ents

of i

mpr

ovem

ent

AE:

Phy

sicia

n’s

Ratin

g Sh

eet

for

Side

Effe

cts

Stan

dard

ised

Form

: 50-

item

che

cklis

t (p

aren

ts);

wei

ght

Brow

n, 1

98642

PM

PH (2

0.08

mg/

day,

b.d

.) –

405.

7–13

.13

mon

ths

Cor

e: C

PRS:

hyp

erac

tivity

inde

x; A

CTe

Rs: h

yper

activ

ityQ

oL: n

ot r

epor

ted

AE:

not

rep

orte

d

Brow

n, 1

98843

C (4

×)M

PH (2

5.1

mg/

day,

b.d

.) –

1113

–15

8C

ore:

CPR

S-R:

Hyp

erac

tivity

Inde

x; C

onne

rs’ T

each

erH

yper

activ

ity In

dex;

AC

TeRS

: hyp

erac

tivity

QoL

: not

rep

orte

dA

E: S

ERS

(par

ents

); w

eigh

t

Bark

ley,

199

039C

(3×)

MPH

(0.6

0 m

g/kg

/day

, b.d

.) –

835–

133

Cor

e: n

ot r

epor

ted

QoL

: not

rep

orte

dA

E: S

timul

ant

Dru

g Si

de E

ffect

s Q

uest

ionn

aire

: 17-

item

list

Fisc

her,

1991

54C

(3×)

MPH

(0.8

0 m

g/kg

/day

, b.d

.) –

161

2.4–

17.2

3C

ore:

CPR

S-R:

Hyp

erac

tivity

Inde

x; C

TRS

-R: H

yper

activ

ityIn

dex;

CT

RS-R

: hyp

erac

tivity

QoL

: not

rep

orte

dA

E: C

PRS-

R: p

sych

osom

atic

; SER

S (p

aren

ts, t

each

ers)

:nu

mbe

r an

d se

verit

y of

sid

e-ef

fect

s

Ahm

ann,

199

335C

(3×)

MPH

(0.9

0 m

g/kg

/day

, t.d

.s.)

– 23

45–

154

Cor

e: n

ot r

epor

ted

QoL

: not

rep

orte

dA

E: B

SEQ

cont

inue

d


25


TA

BLE

5M

PH m

ediu

m d

ose

(15–

30 m

g/da

y) v

ersu

s pl

aceb

o (c

ont’d

)

Stud

yD

esig

nIn

terv

enti

on –

NA

ge

Dur

atio

n C

ore

outc

omes

(yea

rs)

(wee

ks)

Fine

, 199

352C

(3×)

MPH

[0.6

0 m

g/kg

/day

(unc

lear

), b.

d.] –

12

6–10

3C

ore:

not

rep

orte

dQ

oL: n

ot r

epor

ted

AE:

Sid

e Ef

fect

s Q

uest

ionn

aire

Pelh

am, 1

99379

C (3

×)M

PH (1

6.2

mg/

day,

b.d

.) –

315.

4–9.

96

Cor

e: n

o hy

p; IO

WA

CT

RS: i

natt

entio

n/ov

erac

tivity

QoL

: not

rep

orte

dA

E: n

ot r

epor

ted

Klo

rman

, 199

468C

(2×)

MPH

(22.

3 m

g/da

y, b

.d.)

– 11

45.

6–11

.96

Cor

e: A

bbre

viat

ed C

onne

rs’ H

yper

activ

ity Q

uest

ionn

aire

(par

ents

and

tea

cher

s); T

OT

S: h

yper

activ

ity (p

aren

ts a

ndte

ache

rs)

QoL

: no

CG

IA

E: s

omat

ic c

ompl

aint

s (p

aren

ts);

moo

d pr

oble

ms

(par

ents

); w

eigh

t

Buite

laar

, 199

644P

MPH

(20

mg/

day,

b.d

.) –

326–

134

Cor

e: d

ata

pres

ente

d in

gra

phs

cann

ot b

e ex

trac

ted

QoL

: not

rep

orte

dA

E: S

SERS

(mod

ified

ver

sion)

: fre

quen

cy a

nd s

ever

ity(p

aren

ts, p

sych

iatr

ist)

Stei

n, 1

99691

C (4

×)M

PH (1

7.6

mg/

day,

b.d

.) –

256–

124

Cor

e: C

PRS

– hy

pera

ctiv

ity in

dex;

AC

TeRS

: hyp

erac

tivity

MPH

(26.

4 m

g/da

y, t

.d.s

.) –

25Q

oL: n

ot r

epor

ted

MPH

(17.

6–26

.4 m

g/da

y 2

or 3

×tit

rate

d) –

25

AE:

SSE

RS (p

aren

ts);

wei

ght

Hoe

ppne

r, 19

9761

C (3

×)M

PH (0

.60

mg/

kg/d

ay, b

.d.)

– 50

6.1–

18.2

4C

ore:

CPR

S: H

yper

activ

ity In

dex;

CT

RS: H

yper

activ

ityIn

dex

QoL

: not

rep

orte

dA

E: n

ot r

epor

ted

Han

den,

199

960C

(3×)

MPH

(25–

30 m

g/da

y, m

ax. 3

×) –

11

4–5.

13

Cor

e: C

TRS

: Hyp

erac

tivity

Inde

x; C

TRS

: hyp

erac

tivity

QoL

: not

rep

orte

dA

E: S

ide

Effe

cts

Che

cklis

t (t

each

ers,

par

ents

); m

ean

seve

rity

ratin

g 0–

6

Kolk

o, 1

99969

C (4

×)M

PH (0

.60

mg/

kg/d

ay, b

.d.)

– 22

7–13

6C

ore:

no

hyp;

IOW

A C

TRS

: ina

tten

tion/

over

activ

ity.

QoL

: not

rep

orte

dA

E: S

SERS

Man

os, 1

99972

C (4

×)M

PH (2

0 m

g/da

y, b

.d.)

– 42

5–17

4C

ore:

no

hyp;

ASQ

(par

ents

) (te

ache

rs);

ARS

(par

ent)

MPH

(30

mg/

day,

b.d

.) –

42Q

oL: n

o C

GI;

com

posit

e ra

tings

(clin

icia

n)A

E: S

E/BM

S (p

aren

ts)

cont

inue

d


26 TA

BLE

5M

PH m

ediu

m d

ose

(15–

30 m

g/da

y) v

ersu

s pl

aceb

o (c

ont’d

)

Stud

yD

esig

nIn

terv

enti

on –

NA

ge

Dur

atio

n C

ore

outc

omes

(yea

rs)

(wee

ks)

Zei

ner,

1999

98C

(2×)

MPH

(22.

4 m

g/da

y, b

.d. o

r t.d

.s.)

– 36

7–12

6C

ore:

CT

RS: h

yper

activ

ity; P

AC

S: h

yper

activ

ityQ

oL: n

ot r

epor

ted

AE:

not

rep

orte

d

Plisz

ka, 2

00083

PM

PH (2

5.2

mg/

day,

b.d

. or

t.d.s

.) –

586–

113

Cor

e: n

o hy

p; IO

WA

CT

RS: i

natt

entio

n/ov

erac

tivity

QoL

: CG

I (im

prov

emen

t); C

onne

rs’ G

loba

l Ind

ex (p

aren

t)A

E: M

ulti-

Mod

ality

Tre

atm

ent

of A

DH

D S

ide

Effe

cts

Scal

e;w

eigh

t

Bark

ley,

200

040C

(5×)

MPH

(20

mg/

day,

b.d

.) –

3812

–17

5C

ore:

no

hyp;

AD

HD

Tot

al P

aren

t/Te

ache

r ra

ting

QoL

: not

rep

orte

dA

E: n

umbe

r an

d se

verit

y of

sid

e-ef

fect

s (t

each

er, p

aren

ts,

self)

Terv

o, 2

00293

C (3

×)M

PH (0

.60

mg/

kg/d

ay, b

.d.)

– 41

M=

9.9

(2.9

)3

Cor

e: n

o hy

p; C

BCL

(par

ent)

QoL

: not

rep

orte

dA

E: n

ot r

epor

ted

AC

TeRS

, AD

D-H

Com

preh

ensiv

e Te

ache

rs’ R

atin

g Sc

ale;

ARS

, AD

HD

Rat

ing

Scal

e; A

SQ, A

bbre

viat

ed S

ympt

oms

Que

stio

nnai

re; b

.d.,

twic

e da

ily; B

SEQ

, Bar

kley

Sid

e Ef

fect

sQ

uest

ionn

aire

; C, c

ross

over

tria

l (nu

mbe

r of

cro

ssov

ers)

; CBC

L, C

hild

Beh

avio

ur C

heck

list;

CG

I, C

linic

al G

loba

l Im

pres

sion;

CPR

S, C

onne

rs’ P

aren

t Ra

ting

Scal

e; C

TRS

, Con

ners

’Te

ache

r Ra

ting

Scal

e; IO

WA

, Ina

tten

tion/

Ove

ract

ivity

with

Agg

ress

ion;

o.d

., on

ce d

aily

; P, p

aral

lel t

rial;

PAC

S, P

aren

tal A

ccou

nt o

f Chi

ldho

od S

ympt

oms;

SE/

BMS,

Sid

e Ef

fect

sBe

havi

our

Mon

itorin

g Sc

ale;

SER

S, S

ide

Effe

cts

Ratin

g Sc

ale;

SSE

RS, S

timul

ant

Dru

g Si

de E

ffect

s Ra

ting

Scal

e; t

.d.s

., th

ree

times

dai

ly.

HyperactivityAll nine studies that examined hyperactivity useda Conners’ scale. Six used the Conners’Hyperactivity Index as an outcome measure –three of which reported results for both parentsand teachers,43,54,61 one that reported results forteachers only60 and two that reported results forparents only.42,91 In addition to these studies, tworeported measuring hyperactivity for parents andteachers using a Conners’ questionnaire,46,68 andone measured hyperactivity only using theCTRS.98 Five of the studies also measuredhyperactivity using additional scales.42,43,68,91,98

Table 6 presents the means and SD for the resultsof these studies (where reported). Two of thestudies used a parallel design.42,46 As with low-doseplacebo, none of the studies could be combinedbecause of clinical heterogeneity (e.g. age), and/orbecause the necessary data were not available tofacilitate a meta-analysis of the crossover studies.Two studies that could have been combined (byKlorman and colleagues68 and Conners andTaylor46) used different study designs. Table 6shows that most of the studies reported significantimprovements for medium-dose MPH comparedwith placebo. When assessing hyperactivity usingthe Conners’ scales, four of the studies reportedsome tests with no significant changes, most ofwhich were measures of hyperactivity as assessedby the parents.42,43,91

Quality of lifeOnly one of these 21 studies examined CGI(considered to be a proxy for QoL in this SR).Pliszka and colleagues83 assessed the effectivenessof medium-dose MPH compared with placebousing the Clinical Global Impressionimprovement subscale (CGS-I) and also theConners Global Index (see Appendix 12). Theyobserved a significant difference between thetreatment group and placebo group at 3 weeks(p < 0.05), favouring the treatment group [mean(SD): MPH group 2.35 (0.81) and placebo group3.22 (1.44)]. They did not, however, find anysignificant differences between groups using theConners’ Global Index as assessed by the parents.

Other scales could also be used as an indicator ofQoL. For example, Conners and Taylor46 assessedparent global judgements of improvement (howserious a problem does your child have?:no/minor/serious problem). After 8 weeks, 27.8%of children in the MPH group were deemed tohave a ‘serious problem’ compared with 50% inthe placebo group.

Adverse eventsOf 21 studies comparing a medium dose of MPHwith placebo, eight reported data which wereinformative to the analysis of adverse events. Thestudy by Stein and colleagues91 examined theeffectiveness of a medium dose of MPHadministered twice daily, three times daily or bytitration. Two trials detected a significantly higheroccurrence of headache during the MPH phase,with RRs of 1.43 and 2.33, respectively (Figure 4).

Participants in four of the five crossover trialsreporting loss of appetite displayed a higherincidence of this outcome when assigned to amedium dose of MPH. RRs ranged from 2.00 to3.86 (Figure 5). The direction of effect inHanden60 was similar, but not significant. Thismay be related to differences in the ages ofparticipants; Handen 199960 examined children ofpreschool age whereas the other studies examinedchildren 5–13 years old.

A significantly higher incidence of stomach acheduring the MPH treatment phase was found intwo crossover trials, with RRs of 1.84 and 2.13,respectively (Figure 6).

One parallel trial and three crossover trialsdetected significant differences in the occurrenceof insomnia between treatment arms, with RRsranging from 1.55 to 2.73 (Figure 7). Greaterproportions of participants suffered from thisadverse event when assigned to the active drug.

Only one trial presented informative data onparticipants’ weight.83 No significant differences inmean weight change were detected betweentreatment groups (mean difference = –0.70; 95%CI –6.16 to 4.76).

SummaryGenerally, the majority of studies that examinedhyperactivity reported that medium-dose MPHwas beneficial compared with placebo. Only onestudy evaluated the CGI-I subscale.83 In this study,the authors reported that behaviour wasstatistically improved with medium-dose MPHcompared with placebo. Where data on adverseevents were available, medium-dose MPH wasconsistently associated with higher incidences ofheadache, loss of appetite, stomach ache andinsomnia. It is noted that, in general, the studiesdid not score very well in the quality assessement,and the results should be interpreted with caution.

MPH high dose (>30 mg/day) versus placeboTen studies examined high-dose (>30 mg/day)


27



28 TA

BLE

6Re

sults

for h

yper

activ

ity [

MPH

med

ium

dos

e (1

5–30

mg/

day)

ver

sus

plac

ebo]

Stud

ySc

ale

MP

H m

ediu

m d

ose:

P

lace

bo: m

ean

(SD

)p-V

alue

(if

repo

rted

) or

m

ean

(SD

)m

ean

diffe

renc

ea

2–17

yea

rsFi

sche

r, 19

9154

CPR

S-R

(Hyp

erac

tivity

Inde

x)

11.1

(6.7

)14

.3 (6

.8)

S –

NR

CT

RS-R

(Hyp

erac

tivity

Inde

x)

8.4

(6.3

)13

.7 (7

.6)

S –

NR

CT

RS-R

(hyp

erac

tivity

)6.

0 (4

.9)

9.5

(5.8

)S

– N

R

4–5

year

sH

ande

n, 1

99960

CT

RS (h

yper

activ

ity)

6.2

(3.4

)14

.0 (3

.7)

<0.

05C

TRS

(Hyp

erac

tivity

Inde

x)9.

2 (3

.8)

17.4

(6.0

)<

0.05

5–11

yea

rs/5

–13

year

sK

lorm

an, 1

99468

Abb

revi

ated

Con

ners

’ Hyp

erac

tivity

Que

stio

nnai

re:

Pare

nts

0.99

(0.6

0)1.

41 (0

.60)

<0.

01Te

ache

rs0.

56 (0

.42)

1.09

(0.5

2)<

0.01

TO

TS

(hyp

erac

tivity

) (pa

rent

s)–0

.44

(0.5

1)–0

.11

(0.4

8)<

0.01

TO

TS

(hyp

erac

tivity

) (te

ache

rs)

–1.0

4 (0

.43)

–0.6

1 (0

.45)

<0.

01

Brow

n, 1

98642

MD

(95%

CI)

CPR

S (H

yper

activ

ity In

dex)

15.8

8 (6

.36)

17.2

5 (7

.50)

–0.2

4 (–

4.61

to

4.13

)A

CTe

Rs (h

yper

activ

ity)

14.2

5 (5

.6)

17.5

0 (4

.41)

4.12

(–7.

84 t

o –0

.40)

6–11

yea

rs/6

–12

year

sC

onne

rs, 1

98046

MD

(95%

CI)

Con

ners

’ Par

ent

Que

stio

nnai

re (h

yper

activ

ity)

0.46

(0.2

3)0.

75 (0

.36)

–0.3

0 (–

0.46

to

–0.1

4)C

onne

rs’ T

each

er Q

uest

ionn

aire

(hyp

erac

tivity

)1.

28 (0

.67)

1.45

(0.6

3)–0

.51

(–0.

75 t

o –0

.27)

Stei

n, 1

99691

CPR

S (H

yper

activ

ity In

dex)

titra

tion:

11.4

(4.0

)11

.7 (6

.4)

NS

b.i.d

.:9.

7 (5

.8)

t.i.d

.:9.

2 (7

.5)

titra

tion:

16.3

(4.6

)b.

i.d.:

15.5

(4.8

)A

CTe

RS (h

yper

activ

ity)

t.i.d

.:13

.5 (4

.8)

16.3

(5.7

)N

S

6–18

yea

rsH

oepp

ner,

1997

61C

PRS

(Hyp

erac

tivity

Inde

x)7.

13 (6

.37)

8.40

(6.5

9)S

– N

RC

TRS

(Hyp

erac

tivity

Inde

x)8.

20 (6

.85)

13.5

4 (8

.66)

S –

NR

7–12

yea

rsZ

eine

r, 19

9998

CT

RS (h

yper

activ

ity)

8.83

(6.4

9)14

.69

(6.1

7)<

0.00

01PA

CS

(hyp

erac

tivity

) 3.

08 (3

.70)

5.25

(5.0

1)

<0.

05

cont

inue

d


29


TA

BLE

6Re

sults

for h

yper

activ

ity [

MPH

med

ium

dos

e (1

5–30

mg/

day)

ver

sus

plac

ebo]

Stud

ySc

ale

MP

H m

ediu

m d

ose:

P

lace

bo: m

ean

(SD

)p-V

alue

(if

repo

rted

) or

m

ean

(SD

)m

ean

diffe

renc

ea

13–1

5 ye

ars

Brow

n, 1

98843

CPR

S-R

(Hyp

erac

tivity

Inde

x)9.

00 (3

.16)

12.6

6 (4

.13)

NS

Con

ners

’ Tea

cher

Hyp

erac

tivity

Inde

x 18

.33

(3.0

7)24

.50

(2.8

1)<

0.05

AC

TeRS

(hyp

erac

tivity

)9.

50 (5

.04)

8.00

(0.6

3)N

S

aN

ote

that

ow

ing

to t

he o

vera

ll po

or r

epor

ting

of s

tudy

met

hodo

logy

, p-v

alue

s sh

ould

be

inte

rpre

ted

with

cau

tion.

Lo

wer

sco

res

repr

esen

t a

bett

er b

ehav

iour

al o

utco

me.

AC

TeRS

, AD

D-H

Com

preh

ensiv

e Te

ache

r Ra

ting

Scal

e; C

PRS,

Con

ners

’ Par

ent

Ratin

g Sc

ale;

CPR

S-R,

Con

ners

’ Par

ent

Ratin

g Sc

ale

– Re

vise

d; C

TRS

, Con

ners

’ Tea

cher

Rat

ing

Scal

e;C

TRS

-R, C

onne

rs’ T

each

er R

atin

g Sc

ale

– Re

vise

d; M

D, m

ean

diffe

renc

e; N

S, n

ot s

igni

fican

t; PA

CS,

Par

enta

l Acc

ount

of C

hild

hood

Sym

ptom

s; S

– N

R, s

igni

fican

t (v

alue

not

repo

rted

).


30

Study

Parallel trialsBuitelaar 1996Conners 1980Pliszka 2000

Crossover trialsAhmann 1993Barkley 1990Handen 1999Stein 1996 (b.d.)Stein 1996 (t.d.s.)Stein 1996 (titrate)

MPH (medium dose)n/N

2/105/200/20

63/20621/82

0/115/254/255/25

Placebon/N

3/112/211/18

44/2069/820/112/252/252/25

RR (fixed)95% CI

RR (fixed)95% CI

0.732.630.30

1.432.33

2.502.002.50

(0.15 to 3.53)(0.57 to 12.02)(0.01 to 6.97)

(1.03 to 2.00)(1.14 to 4.79)Not estimable(0.53 to 11.70)(0.40 to 9.95)(0.53 to 11.70)

210.50.20.1 5Favours placeboFavours MPH

10

FIGURE 4 Relative risks of headache: MPH (medium dose) versus placebo

Study


Crossover trialsAhmann 1993Barkley 1990Handen 1999Klorman 1994Stein 1996 (b.d.)Stein 1996 (t.d.s.)Stein 1996 (titrate)


2/10 8/20 3/20

115/20643/82 6/11

27/11416/2518/2516/25

Placebon/N

3/115/210/18

52/20612/82

1/11 7/114

8/258/258/25

RR (fixed)95% CI

RR (fixed)95% CI

0.731.686.33

2.213.586.003.862.002.252.00

(0.15 to 3.53)(0.66 to 4.28)(0.35 to 114.81)

(1.70 to 2.88)(2.04 to 6.29)(0.86 to 41.96)(1.75 to 8.50)(1.05 to 3.80)(1.21 to 4.19)(1.05 to 3.80)


10

FIGURE 5 Relative risks of loss of appetite: MPH (medium dose) versus placebo

Study


Crossover trialsAhmann 1993Barkley 1990Handen 1999Stein 1996 (b.d.)Stein 1996 (t.d.s.)Stein 1996 (titrate)


1/109/201/20

70/20632/82

0/1110/25

5/259/25

Placebon/N

3/117/210/18

38/20615/82

0/114/254/254/25

RR (fixed)95% CI

RR (fixed)95% CI

0.371.352.71

1.842.13

2.501.252.25

(0.05 to 2.98)(0.62 to 2.93)(0.12 to 62.70)

(1.31 to 2.60)(1.25 to 3.63)Not estimable(0.90 to 6.92)(0.38 to 4.12)(0.80 to 6.36)


10

FIGURE 6 Relative risks of stomach ache: MPH (medium dose) versus placebo

immediate-release MPH compared with placebo(Table 7; with additional information presented inAppendix 12). All of these studies examined high-dose MPH administered more than once daily.

Of the studies that examined high-dose MPH andplacebo, two examined hyperactivity43,67 and oneevaluated hyperactivity/impulsivity using theSNAP-IV scale (a variation of the Conners’ scales).These studies will be discussed separately below.Three studies presented results for adverse eventsonly,35,39,57 and these will also be presented below.

Two of the remaining studies examinedinattention/overactivity using an IOWA CTRS asone of the main core outcomes: Kolko andcolleagues69 observed that high-dose MPH wasbetter than placebo (p < 0.0001). Pelham andcolleagues79 reported that behaviour improvedwith treatment, but did not present directstatistical comparisons (see Appendix 12).

[Confidential information removed].

HyperactivityOf the three studies that examined hyperactivity,one did not report enough data to be included inTable 8.67 One of the remaining studies used acrossover design43 and the other was a parallelstudy.97 The studies used different Conners’ scales,but both reported results for parents and teachers.The study by Brown and Sexson43 also examinedhyperactivity using the ACTeRS. Given thedifferent outcome scales, study designs and agegroups, data from these two studies were not pooled.

Results from the two studies show some variability(Table 8). Of the two results that were not found tobe significant, one used a scale that was assessed

by parents and the other used a scale assessed byteachers.

Quality of lifeWolraich and colleagues97 examined CGI andreported that 47.2% of participants on immediate-release high-dose MPH were ‘much or very muchimproved’ at the end of the study compared with16.7% of participants in the placebo group (nosignificance value was reported).


It is noted that Klorman and colleagues66,67 alsoreported on ratings of global outcome (resultspresented in Appendix 12) – data that could alsobe used to assess QoL but were not considered tobe a primary outcome measure in this SR.

Adverse eventsOf 10 trials comparing a high dose of MPH withplacebo, seven presented usable data on adverseevents (see Table 7). The occurrence of headachewas significantly higher in the treatment grouponly in Ahmann and colleagues35 (RR = 1.89, 95%CI 1.34 to 2.68) (Figure 8).

One parallel trial found significantly higherproportions of participants suffering from loss ofappetite in the treatment group, with an RR of22.54 (95% CI 3.18 to 159.56). Three crossovertrials also detected higher proportions in thetreatment groups with RRs of between 2.44 and4.67 (Figure 9).

Two crossover trials reporting occurrence ofstomach ache detected significant differencesbetween the treatment and placebo groups; RRs of2.11 and 1.93 are reported (Figure 10). Both


31


Study

Parallel trialsBuitelaar 1996Conners 1980

Crossover trialsAhmann 1993Barkley 1990Klorman 1994Stein 1996 (b.d.)Stein 1996 (t.d.s.)Stein 1996 (titrate)


4/1013/20

121/206 51/82 41/11415/25 15/25 14/25

Placebon/N

3/115/21

76/20633/82 21/11413/25 13/25 13/25

RR (fixed)95% CI

RR (fixed)95% CI

1.472.73

1.591.551.951.151.151.08

(0.43 to 5.01)(1.19 to 6.26)

(1.29 to 1.97)(1.13 to 2.11)(1.24 to 3.08)(0.70 to 1.89)(0.70 to 1.89)(0.65 to 1.08)


10

FIGURE 7 Relative risks of insomnia: MPH (medium dose) versus placebo


32 TA

BLE

7M

PH h

igh

dose

(>

30 m

g/da

y) v

ersu

s pl

aceb

o

Stud

yD

esig

nIn

terv

enti

on –

NA

ge

Dur

atio

n C

ore

outc

omes

(yea

rs)

(wee

ks)

Adm

inist

ered

tw

o or

mor

e tim

es d

aily

Gitt

elm

an-K

lein

, 197

657P

MPH

(max

. 60

mg/

day,

b.d

.) –

166

6–12

12C

ore:

no

core

sym

ptom

s re

port

edQ

oL: n

ot r

epor

ted

AE:

sid

e ef

fect

s st

anda

rdise

d fo

rm: s

ever

ity a

nd c

onsis

tenc

yK

lorm

an, 1

98766

C (2

×)M

PH (2

5–40

mg/

day,

t.d

.s.)

– 19

12–1

96

Cor

e: n

o hy

p; A

bbre

viat

ed C

onne

rs’ Q

uest

ionn

aire

Q

oL: n

o C

GI;

Lone

y an

d O

rdon

a Sc

ale:

res

pons

e (p

atie

nts

and

pare

nts)

AE:

ST

ESS

(bas

ed o

n in

terv

iew

with

par

ent

and

patie

nt)

Brow

n, 1

98843

C (4

×)M

PH (4

2.56

mg/

day,

b.d

.) –1

113

–15

8C

ore:

CPR

S: H

yper

activ

ity In

dex;

Con

ners

’ Tea

cher

Hyp

erac

tivity

Inde

x;A

CTe

RS: h

yper

activ

ityQ

oL: n

ot r

epor

ted

AE:

SER

S (p

aren

ts);

wei

ght

Bark

ley,

199

039C

(3×)

MPH

(1.0

mg/

kg/d

ay, b

.d.)

– 83

5–13

3C

ore:

not

rep

orte

dQ

oL: n

ot r

epor

ted

AE:

Stim

ulan

t D

rug

Side

Effe

cts

Que

stio

nnai

re: 1

7-ite

m li

stK

lorm

an, 1

99067

C (2

×)M

PH (3

5.21

mg/

day,

t.d

.) –

4812

–18

6C

ore:

Abb

revi

ated

Con

ners

’ Hyp

erac

tivity

Que

stio

nnai

re (p

aren

ts,

teac

hers

); IO

WA

inat

tent

ion/

over

activ

ity s

cale

(par

ents

, tea

cher

s);

TO

TS:

hyp

erac

tivity

and

att

entio

n sc

ales

(par

ents

, tea

cher

s)Q

oL: n

o C

GI;

ratin

gs o

f glo

bal o

utco

me

(par

ents

/pat

ient

s)A

E: S

TES

S (p

atie

nts,

par

ents

); N

owlis

Moo

d Sc

ale:

fatig

ue (p

atie

nts)

;w

eigh

t Pe

lham

, 199

379C

(3×)

MPH

(32

mg/

day,

b.d

.) –

315.

4–9.

96

Cor

e: n

o hy

p; IO

WA

CT

RS: i

natt

entio

n/ov

erac

tivity

QoL

: not

rep

orte

dA

E: n

ot r

epor

ted

Ahm

ann,

199

335C

(3×)

MPH

(1.5

mg/

kg/d

ay, t

.d.)

– 23

45–

154

Cor

e: n

ot r

epor

ted

QoL

: not

rep

orte

dA

E: B

SEQ

Kolk

o, 1

99969

C (4

×)M

PH (1

.2 m

g/kg

/day

, b.d

.) –

227–

136

Cor

e: n

o hy

p; IO

WA

CT

RS: i

natt

entio

n/ov

erac

tivity

QoL

: not

rep

orte

dA

E: S

SERS

Wol

raic

h, 2

00197

PM

PH (1

5, 3

0 or

45

mg/

day,

t.d

.) –

312

6–12

4C

ore:

SN

AP-

IV H

yper

activ

ity/Im

pulsi

vity

(tea

cher

rat

ed) (

pare

nt r

ated

)Q

oL: C

GI (

inve

stig

ator

rat

ed)

AE:

sol

icite

d an

d sp

onta

neou

s re

port

s: fo

cus

on s

leep

qua

lity,

tic

s an

dap

petit

e (p

aren

t)Q

uinn

, 200

384[C

onfid

enti

al in

form

atio

n re

mov

ed]

AC

TeRS

, AD

D-H

Com

preh

ensiv

e Te

ache

rs’ R

atin

g Sc

ale;

b.d

., tw

ice

daily

; BSE

Q, B

arkl

ey S

ide

Effe

cts

Que

stio

nnai

re; C

, cro

ssov

er t

rial (

num

ber

of c

ross

over

s); C

GI,

Clin

ical

Glo

bal

Impr

essio

n; C

PRS,

Con

ners

’ Par

ent

Ratin

g Sc

ale;

CT

RS, C

onne

rs’ T

each

er R

atin

g Sc

ale;

P, p

aral

lel t

rial;

SNA

P, Sw

anso

n, N

olan

and

Pel

ham

(sca

le);

SSER

S, S

timul

ant

Dru

g Si

de E

ffect

sRa

ting

Scal

e; t

.d.s

., th

ree

times

dai

ly.


33


TA

BLE

8Re

sults

for h

yper

activ

ity [

MPH

Hig

h do

se (

>30

mg/

day)

ver

sus

plac

ebo]

Stud

ySc

ale

MP

H h

igh

dose

: P

lace

bo: m

ean

(SD

)p-V

alue

(if

repo

rted

) or

m

ean

(SD

)m

ean

diffe

renc

ea

6–12

yea

rsW

olra

ich,

200

197M

D (9

5%C

I)SN

AP-

IV H

yper

activ

ity/Im

pulsi

vity

(tea

cher

)0.

93 (0

.79)

1.57

(0.8

9)–1

.26

(–1.

44 t

o –1

.08)

SN

AP-

IV H

yper

activ

e/Im

pulsi

vity

(par

ent)

1.10

(0.6

9)1.

83 (0

.89)

–0.5

8 (–

0.73

to

–0.4

3)

13–1

5 ye

ars

Brow

n, 1

98843

CPR

S (H

yper

activ

ity In

dex)

9.33

(2.0

6)12

.66

(4.1

3)N

SC

onne

rs’ T

each

er H

yper

activ

ity In

dex

17.3

3 (3

.72)

24.5

0 (2

.81)

<0.

05A

CTe

RS (h

yper

activ

ity)

10.0

0 (6

.69)

8.00

(0.6

3)N

S

aN

ote

that

ow

ing

to t

he o

vera

ll po

or r

epor

ting

of s

tudy

met

hodo

logy

, the

p-v

alue

s sh

ould

be

inte

rpre

ted

with

cau

tion.

Lo

wer

sco

res

repr

esen

t a

bett

er b

ehav

iour

al o

utco

me.

AC

TeRS

: AD

D-H

Com

preh

ensiv

e Te

ache

r Ra

ting

Scal

e; C

PRS,

Con

ners

’ Par

ent

Ratin

g Sc

ale;

MD

, mea

n di

ffere

nce;

Nol

an a

nd P

elha

m (s

cale

); N

S, n

ot s

igni

fican

t; SN

AP,

Swan

son.

Stud

y

Para

llel t

rials

Gitt

elm

an-K

lein

197

6Q

uinn

200

3W

olra

ich

2001

Cro

ssov

er tr

ials

Ahm

ann

1993

Bark

ley

1990

Klo

rman

198

7K

lorm

an 1

990

MPH

(hig

h do

se)

n/N

1/4

1

6

/107

70/

206

17/8

2 0

/19

3/4

8

Plac

ebo

n/N

3/42

10/9

9

37/2

06 9

/82

0/

192/

48

RR (f

ixed

)95

% C

IRR

(fix

ed)

95%

CI

0.34

0.56

1.89

1.89

1.50

(0.0

4 to

3.1

5)

(0.2

1 to

1.4

7)

(1.3

4 to

2.6

8)(0

.89

to 3

.99)

Not

est

imab

le(0

.26

to 8

.58)

21

0.5

0.2

0.1

5Fa

vour

s pl

aceb

oFa

vour

s M

PH10

[Con

fiden

tial

info

rmat

ion

rem

oved

]

FIG

UR

E 8

Rela

tive

risks

of h

eada

che:

MPH

(hi

gh d

ose)

ver

sus

plac

ebo

studies included participants spanning a wide agerange: 5–13 and 5–15 years, respectively.

One parallel and two crossover trials detected asignificantly higher incidence of insomnia in thetreatment groups, with the greatest RR in theparallel study (4.10) compared with RRs of 1.51and 1.70 in the crossover trials (Figure 11).


SummaryOnly two studies reported detailed data onhyperactivity and there was variability in theresults. The one non-confidential study thatreported data on CGI reported improvedbehaviour in the high-dose MPH group comparedwith placebo [Confidential information removed].Adverse event data showed that high-dose MPHseems to be associated with higher incidences of

headaches, loss of appetite, stomach ache andinsomnia, although not all differences weresignificant. Most of these studies did not scorevery well in the quality assessment and theirresults should be interpreted with caution.


MPH low dose (≤15 mg/day) plus non-drugintervention versus placeboNo studies reported on low-dose (≤15 mg/day)immediate-release MPH plus a non-drugintervention versus placebo.

MPH medium dose (15–30 mg/day) plus non-drugintervention versus placeboThree studies examined medium-dose(15–30 mg/day) immediate-release MPH plus non-drug intervention compared with placebo (Table 9;with additional information presented in


34

Study

Parallel trialsGittelman-Klein 1976Quinn 2003Wolraich 2001

Crossover trialsAhmann 1993Barkley 1990Klorman 1987Klorman 1990

MPH (high dose)n/N

22/41

20/107

132/20646/82 2/1914/48

Placebon/N

1/42

12/99

54/20612/82

6/193/48

RR (fixed)95% CI

RR (fixed)95% CI

22.54

1.54

2.44 3.83 0.33 4.67

(3.18 to 159.56)

(0.80 to 2.99)

(1.90 to 3.14)(2.20 to 6.69)(0.08 to 1.45)(1.43 to 15.20)


10

[Confidential information removed]

FIGURE 9 Relative risks of loss of appetite: MPH (high dose) versus placebo

Study

Parallel trialsGittelman-Klein 1976Quinn 2003Wolraich 2001


MPH (high dose)n/N

2/41

6/107

74/20629/82 1/19 4/48

Placebon/N

1/42

1/99

35/20615/82

1/192/48

RR (fixed)95% CI

RR (fixed)95% CI

2.05

5.55

2.111.931.002.00

(0.19 to 21.73)

(0.68 to 45.30)

(1.49 to 3.01)(1.12 to 3.33)(0.07 to 14.85)(0.38 to 10.41)


10


FIGURE 10 Relative risks of stomach ache: MPH (high dose) versus placebo

Appendix 12). All of the studies examinedmedium-dose MPH administered more than oncedaily.

Only one of these studies evaluated MPHtreatment plus non-drug intervention usinghyperactivity as an outcome variable (Table 10).42 Inthis parallel study, the non-drug intervention wascognitive therapy. Statistical comparisons betweenthe treatment arms for the (CPRS) or ACTeRs werenot clearly reported by the authors; however, meandifferences (MDs) and 95% CIs show no significantdifferences between the groups.

Rating scaleThe remaining two studies evaluated medium-dose MPH plus non-drug treatment by measuringinattention/overactivity (as one of the coreoutcomes) using the IOWA Conners’ Rating Scales(IOWA-C).69,79 Both studies examined MPH plus abehavioural modification intervention, and bothreported that combined treatment was better thanplacebo; however, the statistical results were notclearly presented (see Appendix 12).

Adverse eventsNone of the trials presented informative adverseevent data.

SummaryOnly one study evaluated the effectiveness ofMPH plus a non-drug intervention compared withplacebo using hyperactivity as an outcome,42 andthe results from these comparisons were notsignificant. None of the studies evaluated CGI.

MPH high dose (>30 mg/day) plus non-drugintervention versus placeboTwo studies evaluated high-dose (>30 mg/day)

immediate-release MPH plus non-drugintervention compared with placebo (Table 11; withadditional information presented in Appendix 12).

Both studies in this category used the IOWA-C toevaluate inattention/overactivity. As presented inTable 11, both examined MPH plus a behaviouralmodification intervention and a placebo group.Both reported that the MPH plus non-intervention group differed from the placebogroup; however, direct statistical comparisons werenot clearly reported (see Appendix 12).

Adverse eventsNone of the trials presented informative adverseevent data.

SummaryNo studies in this category examined hyperactivityor CGI or reported informative adverse eventdata.

ER-MPH low dose (�20 mg/day) versus placeboTwo studies examined low-dose (≤ 20 mg/day)extended-release MPH (ER-MPH) compared withplacebo (Table 12; with additional informationpresented in Appendix 12). Both studiesexamined low-dose MPH administered once daily.

Only one of these studies measured hyperactivityas an outcome.55 Using the Conners’ HyperactivityIndex, this crossover study reported that low-doseER-MPH resulted in improved behaviourcompared with placebo, when assessed by bothparents and teachers (Table 13). The other study inthis category measured other behaviouraloutcomes using scales such as the ADHD RatingScale IV.92 Details from these results are presentedin Appendix 12.


35


Study

Parallel trialsGittelman-Klein 1976Quinn 2003


MPH (high dose)n/N

28/41

110/20656/82 3/19 8/48

Placebon/N

7/42

73/20633/82

4/196/48

RR (fixed)95% CI

RR (fixed)95% CI

4.10

1.511.700.751.33

(2.02 to 8.32)

(1.20 to 1.89)(1.25 to 2.30)(0.19 to 2.91)(0.50 to 3.55)


10


FIGURE 11 Relative risks of insomnia: MPH (high dose) versus placebo


36 TA

BLE

10

Resu

lts fo

r hyp

erac

tivity

[M

PH m

ediu

m d

ose

(15–

30 m

g/da

y) p

lus

non-

drug

inte

rven

tion

vers

us p

lace

bo]

Stud

ySc

ale

MP

H m

ediu

m d

ose

+ n

on-d

rug:

mea

n (S

D)

Pla

cebo

: mea

n (S

D)

Mea

n di

ffere

nce

5–13

yea

rsBr

own,

198

642M

D (9

5% C

I)C

PRS

(Hyp

erac

tivity

Inde

x)

13.7

8 (8

.14)

17.2

5 (7

.50)

–0.3

1 (–

4.47

to

3.85

)A

CTe

Rs (h

yper

activ

ity)

19.6

0 (2

.63)

17.5

0 (4

.41)

0.68

(–1.

89 t

o 3.

25)

Low

er s

core

s re

pres

ent

a be

tter

beh

avio

ural

out

com

e.A

CTe

RS, A

DD

-H C

ompr

ehen

sive

Teac

hers

’ Rat

ing

Scal

e; C

PRS,

Con

ners

’ Par

ent

Ratin

g Sc

ale.

TA

BLE

9M

PH m

ediu

m d

ose

(15–

30 m

g/da

y) p

lus

non-

drug

inte

rven

tion

vers

us p

lace

bo

Stud

yD

esig

nIn

terv

enti

on –

NA

ge

Dur

atio

n C

ore

outc

omes

(yea

rs)

(wee

ks)

Adm

inist

ered

tw

o or

mor

e tim

es d

aily

Brow

n, 1

98642

PM

PH (2

0.08

mg/

day,

b.d

.) –

405.

7–13

.13

mon

ths

Cor

e: C

PRS:

Hyp

erac

tivity

Inde

x; A

CTe

Rs: h

yper

activ

ityQ

oL: n

ot r

epor

ted

AE:

not

rep

orte

d.

Pelh

am, 1

99379

C (3

×)M

PH (1

6.2

mg/

day,

b.d

.) –

315.

4–9.

96

Cor

e: n

o hy

p; IO

WA

CT

RS: i

natt

entio

n/ov

erac

tivity

QoL

: not

rep

orte

dA

E: n

ot r

epor

ted

Kolk

o, 1

99969

C (4

×)M

PH (0

.60

mg/

kg/d

ay, b

.d.)

– 22

7–13

6C

ore:

no

hyp;

IOW

A C

TRS

: ina

tten

tion/

over

activ

ityQ

oL: n

ot r

epor

ted

AE:

SSE

RS

AC

TeRS

, AD

D-H

Com

preh

ensiv

e Te

ache

rs’ R

atin

g Sc

ale;

C, c

ross

over

tria

l (nu

mbe

r of

cro

ssov

ers)

; CPR

S, C

onne

rs’ P

aren

t Ra

ting

Scal

e; C

TRS

, Con

ners

’ Tea

cher

Rat

ing

Scal

e; P,

para

llel t

rial;

SSER

S, S

timul

ant

Dru

g Si

de E

ffect

s Ra

ting

Scal

e.


37


TA

BLE

11

MPH

hig

h do

se (

>30

mg/

day)

plu

s no

n-dr

ug in

terv

entio

n ve

rsus

pla

cebo

Stud

yD

esig

nIn

terv

enti

on –

NA

ge

Dur

atio

n C

ore

outc

omes

(yea

rs)

(wee

ks)

Adm

inist

ered

tw

o or

mor

e tim

es d

aily

Pelh

am, 1

99379

C (3

×)M

PH (3

2 m

g/da

y, b

.d.)

– 31

5.4–

9.9

6C

ore:

no

hyp;

IOW

A C

TRS

: ina

tten

tion/

over

activ

ityQ

oL: n

ot r

epor

ted

AE:

not

rep

orte

d

Kolk

o, 1

99969

C (4

×)M

PH (1

.2 m

g/kg

/day

, b.d

.) –

227–

136

Cor

e: n

o hy

p; IO

WA

CT

RS: i

natt

entio

n/ov

erac

tivity

QoL

: not

rep

orte

dA

E: S

SERS

C, c

ross

over

tria

l (nu

mbe

r of

cro

ssov

ers)

; CT

RS. C

onne

rs’ T

each

er R

atin

g Sc

ale;

SSE

RS, S

timul

ant

Dru

g Si

de E

ffect

s Ra

ting

Scal

e.

TA

BLE

12

ER-M

PH lo

w d

ose

(≤20

mg/

day)

ver

sus

plac

ebo

Stud

yD

esig

nIn

terv

enti

on –

NA

ge

Dur

atio

n C

ore

outc

omes

(yea

rs)

(wee

ks)

Adm

inist

ered

onc

e da

ilyFi

tzpa

tric

k, 1

99255

C (4

×)Su

stai

ned-

rele

ase

MPH

(slo

w-r

elea

se

6.9–

11.5

8C

ore:

Con

ners

’ Hyp

erac

tivity

Inde

x (p

aren

ts a

nd t

each

er)

Rita

lin, S

R-20

) (20

mg/

day,

o.d

.) –

19Q

oL: n

o C

GI;

com

men

ts r

atin

gs (p

aren

t an

d te

ache

r)A

E: S

TES

S (p

aren

ts);

wei

ght

Stei

n, 2

00392

C (4

×)O

ROS

MPH

(Con

cert

a) (1

8 m

g/da

y,

5.9–

164

Cor

e: n

o hy

p; A

DH

D R

atin

g Sc

ale

IV (p

aren

ts)

o.d.

) – 4

7Q

oL: C

GI

AE:

SER

S (p

aren

ts)

C, c

ross

over

tria

l (nu

mbe

r of

cro

ssov

ers)

; CG

I, C

linic

al G

loba

l Im

pres

sion;

SER

S, S

ide

Effe

cts

Ratin

g Sc

ale;

ST

ESS,

Sub

ject

’s T

reat

men

t Em

erge

nt S

ympt

om S

cale

.

Quality of lifeStein and colleagues92 reported results for CGIseverity scores. They observed that overallimpairment decreased with increasing doses ofMPH (p < 0.001). Results were presentedseparately for children with inattentive andcombined subtypes, and are presented inAppendix 12.

Adverse eventsBoth studies reported data on adverse events.Neither study detected differences in theincidence of headache or stomach ache betweentreatment arms. A significantly higher number ofparticipants suffered from decreased appetitewhen assigned to a low dose of ER-MPH in onestudy;92 similar results were observed with regardto incidence of insomnia (Figures 12 and 13).

Neither study adequately reported data on theweight of participants.

SummaryThe study that assessed hyperactivity reported thatlow-dose ER-MPH was better than placebo.55 Theother study assessed CGI, but direct statisticalcomparisons between low dose ER-MPH andplacebo were not reported, although there was atrend towards improvement with drug treatment.Both studies showed differences between groupsfor loss of appetite and insomnia, but not forheadache or stomach ache. The studies did notscore very well in the quality assessment, andalthough the study by Fitzpatrick and colleagues55

scored slightly better, the results should be treatedwith caution.

ER-MPH medium dose (20–40 mg/day) versusplaceboSix studies examined medium-dose(20–40 mg/day) ER-MPH compared with placebo(Table 14; with additional information presented inAppendix 12). All of the studies examined


38

TABLE 13 Results for hyperactivity [ER-MPH low dose (≤ 20 mg/day) versus placebo]

Study Scale ER-MPH low dose: Placebo: p-Value (if mean (SD) mean (SD) reported)a

6–11 yearsFitzpatrick, 199255 Conners’ Hyperactivity Index

(parents) 0.98 (0.72) 1.75 (0.67) <0.006(teachers) 0.77 (0.63) 1.36 (0.80) <0.006

a Note that owing to the overall poor reporting of study methodology, the p-values should be interpreted with caution. Lower scores represent a better behavioural outcome.

Study

Fitzpatrick 1992Stein 2003

ER-MPH (low dose)n/N

7/1930/47

Placebon/N

1/1916/47

RR (fixed)95% CI

RR (fixed)95% CI

7.001.88

(0.95 to 51.54)(1.19 to 2.95)

210.50.20.1 5Favours placeboFavours ER-MPH

10

FIGURE 12 Relative risks of loss of appetite: ER-MPH (low dose) versus placebo

Study

Fitzpatrick 1992Stein 2003

ER-MPH (low dose)n/N

7/1931/47

Placebon/N

3/1921/47

RR (fixed)95% CI

RR (fixed)95% CI

2.331.48

(0.71 to 7.70)(1.01 to 2.16)


10

FIGURE 13 Relative risks of insomnia: ER-MPH (low dose) versus placebo


39


TA

BLE

14

ER-M

PH m

ediu

m d

ose

(20–

40 m

g/da

y) v

ersu

s pl

aceb

o

Stud

yD

esig

nIn

terv

enti

on –

NA

ge

Dur

atio

n C

ore

outc

omes

(yea

rs)

(wee

ks)

Adm

inist

ered

onc

e da

ilyW

olra

ich,

200

197P

ORO

S M

PH (C

once

rta)

(18,

36

or

6–12

4C

ore:

SN

AP-

IV: h

yper

activ

ity/im

pulsi

vity

(par

ent,

teac

her)

54 m

g/da

y, o

.d.)

– 31

2Q

oL: C

GI i

mpr

ovem

ent

(inve

stig

ator

s)A

E: s

olic

ited

and

spon

tane

ous

repo

rts:

focu

s on

sle

epqu

ality

, tic

s an

d ap

petit

e (p

aren

t)

Gre

enhi

ll, 2

00259

PM

PH-M

odifi

ed R

elea

se (M

etad

ate

CD

) 5–

153

Cor

e: n

ot r

epor

ted

(mea

n 40

.7 m

g/da

y, o

.d.)

– 32

1Q

oL: C

GI (

teac

her;

par

ent)

; CG

I im

prov

emen

t ra

tings

AE:

Tea

cher

and

Par

ent

Side

-Effe

ct Q

uest

ionn

aire

s

Qui

nn, 2

00384

[Con

fiden

tial

info

rmat

ion

rem

oved

]

Dop

fner

, 200

348P

MPH

-Ret

ard

(Med

ikin

et) (

20–6

0 m

g/da

y,

6–16

4C

ore:

no

hyp;

Pee

r A

sses

smen

t fo

r H

yper

kine

tic D

isord

ers

o.d.

) – 8

5(t

each

ers)

; effe

ctiv

enes

s sc

ale

(par

ents

, phy

sicia

ns)

QoL

: not

rep

orte

dA

E: n

o sp

ecifi

c sc

ale

repo

rted

Stei

n, 2

00392

C (4

×)O

ROS

MPH

(Con

cert

a) (3

6 m

g/da

y,

5.9–

164

Cor

e: n

o hy

p; A

DH

D R

atin

g Sc

ale

IV (p

aren

ts)/

AC

TeRS

o.d.

) – 4

7Q

oL: C

GI

AE:

SER

S (p

aren

ts)

Swan

son,

200

432C

(3×)

ORO

S M

PH (C

once

rta)

(18,

36

or

6–12

3C

ore:

no

hyp;

SK

AM

P: d

epor

tmen

t; at

tent

ion

(mea

sure

d by

54

mg/

day,

o.d

.) –

184

two

trai

ned

obse

rver

s)Q

oL: n

ot r

epor

ted

MPH

-Mod

ified

Rel

ease

(Met

adat

e C

D)

AE:

sid

e-ef

fect

s on

the

Bar

kley

Sca

le(2

0, 4

0 or

60

mg/

day,

o.d

.) –1

84

C, c

ross

over

tria

l (nu

mbe

r of

cro

ssov

ers)

; CG

I, C

linic

al G

loba

l Im

prov

emen

t; P,

para

llel t

rial;

SKA

MP,

Swan

son,

Kot

kin

Agl

er, M

-Fly

nn, P

elha

m (s

cale

); SN

AP,

Swan

son,

Nol

an a

ndPe

lham

(sca

le).

medium-dose MPH administered once daily. Fourstudies were parallel in design and two used acrossover design. Although participants inSwanson and colleagues’ trial32 were assigned tovarying dosages of ER-MPH according to pre-existing requirements, data were presented by typeof treatment only (Metadate MPH, Concerta MPHor placebo); hence it is included in the medium-dose category. Similarly, data were presentedacross varying dosages in the trials of Wolraichand colleagues97 and Quinn.84

Most of the studies did not report results forhyperactivity, but did measure other coreoutcomes including [Confidential informationremoved], the ADHD Rating Scale92 and theSwanson, Kotkin, Agler, M-Flynn and Pelham(SKAMP Scale).32 In addition, Dopfner andcolleagues48 used the Peer Assessment forHyperkinetic Disorders as assessed by teachers,and a five-point effectiveness scale as assessed byparents, and also by physicians.


The statistical results from the other studies wereless clearly presented (see Appendix 12), althoughStein and colleagues92 reported that as the ER-MPH dose was increased from 0 to 54 mg, parent-rated ADHD symptoms decreased in a linearmanner (p < 0.001).

HyperactivityOne study measured hyperactivity/impulsivityusing the Swanson, Nolan and Pelham (SNAP) IVScale (a variation of the Conners’ scales).97

Wolraich and colleagues97 observed significantimprovements in the ER-MPH group comparedwith the placebo group (Table 15). This result wasconsistent when rated by teachers and parents.

Quality of lifeFour studies in this category reported on CGI.Wolraich and colleagues97 reported that 46.7% of

participants on medium-dose ER-MPH were‘much or very much improved’ at the end of thestudy compared with 16.7% of participants in theplacebo group (no significance value wasreported). Greenhill and colleagues59 observedthat 81% assigned to active drug were significantlyimproved compared with 50% in the placebogroup (p < 0.001).


Finally, Stein and colleagues92 reported CGIseverity scores in children with ADHD inattentivesubtype and combined subtype. The resultsincreased with medication dose, and were betterthan placebo; however, direct statisticalcomparisons were not reported.

Adverse eventsFive of the six studies reported usable dataregarding adverse events. Differences betweenparticipants in the incidence of headache were notdetected in four of these trials. [Confidentialinformation removed].

Participants in the trials of Stein and colleagues92

and Greenhill and colleagues59 suffered fromsignificantly decreased appetite when assigned tothe extended-release form of MPH. Thoseassigned to the Concerta form of MPH in Swansonand colleagues’ trial32 also suffered fromsignificantly decreased appetite (Figure 14).

One crossover trial92 reported higher numbers ofparticipants suffering from stomach ache duringthe ER-MPH phase compared with placebo(RR = 2.09; 95% CI 1.15 to 3.79). The other threenon-confidential trials did not find significantlydifferent frequencies between treatment arms. Of the three trials reporting incidence ofinsomnia,32,59,84,92 no trial detected significantdifferences between participants assigned to ER-MPH compared with placebo. No trial in thiscomparison group reported data on weight.


40

TABLE 15 Results for hyperactivity [ER-MPH medium dose (20–40 mg/day) versus placebo]

Study Scale ER-MPH medium Placebo: Mean dose: mean (SD) mean (SD) difference

6–12 yearsWolraich, 200197 SNAP-IV

(hyperactivity/impulsivity) MD (95%CI)(teacher) 0.96 (0.79) 1.57 (0.89) –1.21 (–1.40 to –1.02)(parent) 1.11 (0.65) 1.83 (0.89) –0.75 (–0.89 to –0.61)

Lower scores represent a better behavioural outcome.

SummaryOnly one study examined hyperactivity/impulsivityusing the SNAP-IV scale and reported significantimprovements in the ER-MPH group comparedwith the placebo group when assessed by teachersand parents.97 Four studies reported results forCGI, two of which did not report direct statisticalcomparisons between treatment and placebo.92,97

One study reported a significant improvement inthe treatment group compared with placebo.59

[Confidential information removed]. Adverseevents data showed that medium dose ER-MPHseems to be associated with a higher incidence ofdecreased appetite. Most of these studies did notscore very well in the quality assessment and theirresults should be interpreted with caution.[Confidential information removed].

ER-MPH high dose (>40 mg/day) versus placeboOnly one study examined high-dose (>40 mg/day)ER-MPH compared to placebo (Table 16; withadditional information presented in Appendix 12).No hyperactivity outcomes were reported,although the study did report on CGI and adverseevents. Although no direct statistical comparisons

were made with placebo, Stein and colleagues92

reported that overall impairment, as measured byCGI severity scores, decreased with increasingdose of MPH (p < 0.001).


Adverse eventsParticipants suffered from a significantly higherincidence of decreased appetite and insomniaduring the ER-MPH phase compared with placebo(RR = 2.31; 95% CI 1.51 to 3.54 and RR = 1.62;95% CI 1.13 to 2.33, respectively). No significantdifferences in the incidence of headache orstomach ache were detected. Data on weight werenot given.

SummaryThe one study included in this category reportedresults for CGI only. The authors reported thatoverall impairment decreased with increasing doseof ER-MPH. In addition, they reported asignificant decrease in appetite and increasedinsomnia with treatment. However, this study didnot score very well in the quality assessment, andthe results should be treated with caution.


41


Study

Crossover trialsStein 2003Swanson 2004 (CON)Swanson 2004 (MCD)

Parallel trialsGreenhill 2002Quinn 2003Wolraich 2001

ER-MPH (medium dose)n/N

35/47 11/181 8/174

73/158

24/106

Placebon/N

16/47 3/183 3/183

33/163

12/99

RR (fixed)95% CI

RR (fixed)95% CI

2.193.712.80

2.28

1.87

(1.42 to 3.37)(1.05 to 13.07)(0.76 to 10.40)

(1.61 to 3.23)

(0.99 to 3.53)


10


FIGURE 14 Relative risks of loss of appetite: ER-MPH (medium dose) versus placebo

TABLE 16 ER-MPH high dose (>40 mg/day) versus placebo

Study Design Intervention – N Age Duration Core outcomes(years) (weeks)

Administered once dailyStein, 200392 C (4×) OROS MPH (Concerta) 5.9–16 4 Core: no hyp; ADHD Rating

(54 mg/day, o.d.) – 47 Scale IV (parents)/ACTeRSQoL: CGIAE: side-effect rating scale(parents)

ACTeRS, ADD-H Comprehensive Teachers’ Rating Scale; AE, adverse events; C, crossover trial (number of crossovers);CGI, Clinical Global Impression.

MPH versus non-drug interventionMPH low dose (��15 mg/day) versus non-druginterventionOnly Brown and colleagues41 evaluated low-dose(≤15 mg/day) immediate-release MPH comparedwith a non-drug intervention (cognitive training)(Table 17; with additional information in Appendix 12). In this study, hyperactivity was notevaluated, although the authors measured overallbehaviour using the CPRS and the AbbreviatedCTRS. Results from comparisons betweentreatment groups were not clearly presented inthis study.

SummaryNo studies in this category examined hyperactivity,CGI or adverse events.

MPH medium dose (15–30 mg/day) versus non-drug interventionFour studies examined medium-dose(15–30 mg/day) immediate-release MPHcompared with a non-drug intervention (Table 18;with additional information presented inAppendix 12). All studies examined MPHadministered two or more times daily.

Only two of the studies measured hyperactivity(Table 19). Firestone and colleagues53 observedthat medium-dose MPH was significantly betterthan parent training. This involved sessions andgroup meetings on child management andlearning how to cooperate efficiently with schoolpersonnel. In the study by Brown andcolleagues,42 results from comparisons between the


42

TABLE 17 MPH low dose (≤ 15 mg/day) versus non-drug intervention

Study Design Intervention – N Age Duration Core outcomes(years) (months)

Administered two or more times dailyBrown, 198541 P MPH (5–15 mg/day, b.d.) vs 6.3–11.8 6 Core: no hyp; CPRS; ACTRS

cognitive training (12 weeks, QoL: not reported24 sessions) – 30 AE: not reported

ACTRS, Abbreviated Conners’ Teacher Rating Scale; CPRS, Conners’ Parent Rating Scale; P, Parallel trial.

TABLE 18 MPH medium dose (15–30 mg/day) versus non-drug intervention


Administered two or more times dailyBrown, 198642 P MPH (20.08 mg/day, b.d.) vs 5.7–13.1 3 months Core: CPRS: Hyperactivity Index;

cognitive therapy (11 weeks, ACTeRS: hyperactivity22 sessions) – 40 AE: not reported

Firestone, 198653 P MPH (22.0 mg/day, b.d.) vs 5–9 3 months Core: CTRS: Hyperactivity Indexparent training (3 months) – 134 QoL: not reported

AE: not reported

Pelham, 199379 C (3×) MPH (16.2 mg/day, b.d.) vs 5.4–9.9 6 Core: no hyp; IOWA CTRS: behavioural modification inattention/overactivityintervention (STP, 8 weeks) – 31 QoL: not reported

AE: not reported

Kolko, 199969 C (4×) MPH (1.2 mg/kg/day, b.d.) vs 7–13 6 Core: no hyp; IOWA CTRS: behavioural modification inattention/overactivity(6 weeks) – 22 QoL: not reported

AE: no details reported

ACTeRS, ADD-H Comprehensive Teachers’ Rating Scale; C, crossover trial (number of crossovers); CPRS, Conners’ ParentRating Scale; CTRS, Conners’ Teacher Rating Scale; P, parallel trial.

MPH group and those receiving cognitive therapywere not clearly presented; however, calculatedMD and 95% CI values demonstrate significantdifferences.

The two remaining studies measuredinattention/overactivity using the IOWA CTRS as amain outcome variable.69,79 However, directstatistical comparisons between medium-doseMPH and behaviour modification were not clearlypresented (see Appendix 12).

Adverse eventsNo study reported adequate data on adverseevents.

SummaryBoth studies that compared medium-dose MPH(with either parent training or cognitive therapy)demonstrated a significant difference in favour ofMPH. None of the studies examined CGI. Thestudies did not score very well in the qualityassessment, and the results should be interpretedwith caution.

MPH high dose (>30 mg/day) versus non-druginterventionThree studies examined high-dose (>30 mg/day)immediate-release MPH compared with a non-drug intervention (Table 20; with additionalinformation presented in Appendix 12). Allstudies examined MPH administered two or moretimes daily.

The two other studies measuredinattention/overactivity using the IOWA CTRS as amain outcome variable.69,79 Statistical significanceof comparisons between high-dose MPHcompared with behaviour modification was notreported in these studies (see Appendix 12).

One study measured hyperactivity using threedifferent measures (Table 21). Klein and Abikoff65

compared high-dose MPH with a behaviouraltherapy that involved parent and teachereducation. They observed that the behaviour ofthe children was better in the MPH groupcompared with non-drug therapy when evaluatedusing the CTRS, but not when evaluated byparents using either the CPRS or the HomeHyperactivity Scale. However, when the MD andCIs are calculated, these results are significant.

In addition, Klein and Abikoff65 assessed CGI.When evaluated by psychiatrists, 79% in the MPHgroup compared with 50% in the non-drug groupimproved after 8 weeks of treatment.

Adverse eventsNo study reported adequate data on adverseevents.

SummaryOnly one study examined high-dose MPHcompared with a non-drug intervention (parentand teacher education) using hyperactivity as anoutcome,65 with significant effects in favour of thedrug treatment group. This study also examinedCGI, and found a higher percentage of improvedchildren in the MPH group compared with thenon-drug intervention group. This study did notscore very well in the quality assessment, and theresults should be interpreted with caution.

MPH low dose (�15 mg/day) plus non-drugintervention versus non-drug interventionThree studies examined low-dose (≤15 mg/day)immediate-release MPH plus non-drugintervention compared with a non-drugintervention (Table 22; with additional informationpresented in Appendix 12). Two studies


43


TABLE 19 Results for hyperactivity [MPH medium dose (15–30 mg/day) versus non-drug intervention]

Study Scale MPH medium dose: Non-drug: Mean mean (SD) mean (SD) difference

5–9 yearsFirestone, 198653 MD (95% CI)

CTRS (Hyperactivity Index) 0.91 (0.58) 1.37 (0.57) –0.49 (–0.65 to –0.33)

5–13 yearsBrown, 198642 MD (95% CI)

CPRS (Hyperactivity Index) 15.88 (6.36) 21.10 (5.65) –5.27 (–9.72 to –0.82)ACTeRS (hyperactivity) 14.25 (5.60) 19.60 (2.63) –4.80 (–7.86 to –1.74)

Lower scores represent a better behavioural outcome.CTRS, Conners’ Teacher Rating Scale; CPRS, Conners’ Parent Rating Scale; ACTeRS, ADD-H Comprehensive Teachers’Rating Scale.


44 TA

BLE

20

MPH

hig

h do

se (

>30

mg/

day)

ver

sus

non-

drug

inte

rven

tion

Stud

yD

esig

nIn

terv

enti

on –

NA

ge

Dur

atio

n C

ore

outc

omes

(yea

rs)

(wee

ks)

Adm

inist

ered

tw

o or

mor

e tim

es d

aily

Pelh

am, 1

99379

C (3

×)M

PH (3

2 m

g/da

y, b

.d.)

vs b

ehav

iour

al

5.4–

9.9

6 C

ore:

no

hyp;

IOW

A C

TRS

: ina

tten

tion/

over

activ

itym

odifi

catio

n in

terv

entio

n (S

TP,

8 w

eeks

) – 3

1Q

oL: n

ot r

epor

ted

AE:

not

rep

orte

d

Kle

in, 1

99765

PM

PH (1

.55

mg/

kg/d

ay, b

.d.)

vs

6–12

12

Cor

e: C

TRS

and

CPR

S: h

yper

activ

ity; H

ome

Hyp

erac

tivity

be

havi

oura

l int

erve

ntio

n +

par

ent

Scal

e (p

aren

ts)

and

teac

her

educ

atio

n (4

wee

ks) –

86

QoL

: CG

I (te

ache

r/m

othe

rs/p

sych

iatr

ists)

AE:

not

rep

orte

d

Kolk

o, 1

99969

C (4

×)M

PH (1

.2 m

g/kg

/day

, b.d

.) vs

beh

avio

ural

7–

136

Cor

e: n

o hy

p; IO

WA

CT

RS: i

natt

entio

n/ov

erac

tivity

mod

ifica

tion

(6 w

eeks

) – 2

2Q

oL: n

ot r

epor

ted

AE:

no

deta

ils r

epor

ted

C, c

ross

over

tria

l (nu

mbe

r of

cro

ssov

ers)

; CG

I, C

linic

al G

loba

l Im

prov

emen

t; C

PRS,

Con

ners

’ Par

ent

Ratin

g Sc

ale;

CT

RS, C

onne

rs’ T

each

er R

atin

g Sc

ale;

P, p

aral

lel t

rial;

STP,

Sum

mer

Trea

tmen

t Pr

ogra

mm

e.

TA

BLE

21

Resu

lts fo

r hyp

erac

tivity

[M

PH h

igh

dose

(>

30m

g/da

y) v

ersu

s no

n-dr

ug in

terv

entio

n]

Stud

ySc

ale

MP

H h

igh

dose

: mea

n (S

D)

Non

-dru

g: m

ean

(SD

)p-V

alue

(if

repo

rted

)a

6–12

yea

rsK

lein

, 199

765M

D (9

5% C

I)C

TRS

(hyp

erac

tivity

)1.

1 (0

.4)

1.5

(0.6

)0.

01 –

0.60

(–0.

76 t

o –0

.44)

CPR

S (h

yper

activ

ity)

0.7

(0.5

)0.

8 (0

.4)

NS

–0.2

0 (–

0.32

to

–0.0

8)H

ome

Hyp

erac

tivity

Sca

le (p

aren

ts)

2.1

(0.4

6)2.

4 (0

.86)

NS

–0.6

0 (–

0.92

to

–0.2

8)

aN

ote

that

ow

ing

to t

he o

vera

ll po

or r

epor

ting

of s

tudy

met

hodo

logy

, the

p-v

alue

s sh

ould

be

inte

rpre

ted

with

cau

tion.

Lo

wer

sco

res

repr

esen

t a

bett

er b

ehav

iour

al o

utco

me.

CPR

S, C

onne

rs’ P

aren

t Ra

ting

Scal

e; C

TRS

, Con

ners

’ Tea

cher

Rat

ing

Scal

e; N

S, n

ot s

igni

fican

t; M

D, m

ean

diffe

renc

e.


45


TA

BLE

22

MPH

low

dos

e (≤

15 m

g/da

y) p

lus

non-

drug

inte

rven

tion

vers

us n

on-d

rug

inte

rven

tion

Stud

yD

esig

nIn

terv

enti

on –

NA

ge

Dur

atio

n C

ore

outc

omes

(yea

rs)

(wee

ks)

Adm

inist

ered

onc

e da

ilyKu

piet

z, 1

98871

PM

PH (0

.3 m

g/kg

, o.d

.) vs

one

-to-

one

7–13

6 m

onth

sC

ore:

CT

RS: h

yper

activ

ityre

adin

g th

erap

y (6

mon

ths)

– 5

8Q

oL: n

ot r

epor

ted

AE:

not

rep

orte

dM

PH (0

.5 m

g/kg

, o.d

.) vs

one

-to-

one

read

ing

ther

apy

(6 m

onth

s) –

58

Pelh

am, 1

99981

C (7

×)M

PH (0

.3 m

g/kg

/day

, o.d

.) vs

beh

avio

ural

6–

126

Cor

e: n

o hy

p; IO

WA

-C: i

natt

entio

n/ov

erac

tivity

(tea

cher

s pr

ogra

mm

e –

21an

d pa

rent

s)Q

oL: n

ot r

epor

ted

AE:

Pitt

sbur

gh S

ide

Effe

ct R

atin

g Sc

ale

(cou

nsel

lors

, tea

cher

s,pa

rent

s)

Adm

inist

ered

tw

o or

mor

e tim

es d

aily

Brow

n, 1

98541

PM

PH (5

–15

mg/

day,

b.d

.) vs

cog

nitiv

e tr

aini

ng

6.3–

11.8

6 m

onth

sC

ore:

no

hyp;

CPR

S; A

CT

RS

(12

wee

ks, 2

4 se

ssio

ns) –

30

QoL

: not

rep

orte

dA

E: n

ot r

epor

ted

AC

TRS

, Abb

revi

ated

Con

ners

’ Tea

cher

Rat

ing

Scal

e; C

, cro

ssov

er t

rial (

num

ber

of c

ross

over

s); C

PRS,

Con

ners

’ Par

ent

Ratin

g Sc

ale;

CT

RS, C

onne

rs’ T

each

er R

atin

g Sc

ale;

P, p

aral

lel

tria

l.

TA

BLE

23

Resu

lts fo

r hyp

erac

tivity

MPH

low

dos

e (≤

15 m

g/da

y) p

lus

non-

drug

inte

rven

tion

vers

us n

on-d

rug

inte

rven

tion

Stud

ySc

ale

MP

H lo

w d

ose

+ n

on-d

rug:

mea

n (S

D)

Non

-dru

g: m

ean

(SD

)M

ean

diffe

renc

e

7–13

yea

rsKu

piet

z, 1

98871

CT

RS (h

yper

activ

ity)

0.3

mg/

kg:

MD

(95%

CI)

2.21

(0.5

3)2.

89 (0

.69)

–0.4

7 (–

0.84

to

–0.1

0)0.

5 m

g/kg

:2.

04 (0

.43)

–0.7

6 (–

1.06

to

–0.4

6)

Low

er s

core

s re

pres

ent

a be

tter

beh

avio

ural

out

com

e.C

TRS

, Con

ners

’ Tea

cher

Rat

ing

Scal

e.

examined MPH administered once daily and oneexamined MPH given two or more times daily.

Only Kupietz and colleagues71 evaluatedhyperactivity as an outcome measure using theCTRS (Table 23). In this parallel study, the non-drug intervention involved a one-to-one readingtherapy programme during weeks 3–14 and weeks16–27 of the study. While the results of theseanalyses are not clearly reported, the MD and 95%CI values show significant differences in favour ofcombined treatment.

Pelham and colleagues81 did not report onhyperactivity or QoL outcomes, but did assessinattention/overactivity (using the IOWA Conners’Rating Scale as assessed by teachers and parents),and also other core outcomes (see Appendix 12).They reported that once-daily low-dose MPH (plusnon-drug intervention) was better forinattention/overactivity than non-drugintervention (p < 0.05).

Brown and colleagues41 examined MPH pluscognitive training compared with cognitivetraining alone. Outcome measures assessedincluded the CPRS and the Abbreviated CTRS. Itappears that MPH plus cognitive training wasmore beneficial than cognitive training alone,although the significance of these comparisons wasnot clear.

Adverse eventsOne study presented data on adverse events.81

Incidence of headache, stomach ache, insomniaand appetite loss did not differ significantlybetween treatment groups. Data on weight werenot reported.

SummaryOnly one study examined hyperactivity as anoutcome measure,71 with significant results infavour of combined treatment. No studiesexamined CGI. Another study examined adverseevents,81 but reported no differences between thetreatment groups. The studies did not score verywell in the quality assessment, and any resultsshould be interpreted with caution.

MPH medium dose (15–30 mg/day) plus non-drugintervention versus non-drug interventionEleven studies evaluated medium-dose(15–30 mg/day) immediate-release MPH plus non-drug intervention compared with a non-drugintervention (Table 24; with additional informationin Appendix 12). One of the studies examinedmedium-dose MPH administered once daily and

10 examined MPH administered two or moretimes daily.

HyperactivityThree studies reported on hyperactivity as anoutcome, all of which used a Conners’ scale(Table 25).42,53,71 Brown and colleagues42 alsoevaluated hyperactivity using the ACTeRS.Although all were parallel studies, the differenttypes of Conners’ scales (two reported results forteachers and one reported result for parents) andthe lack of detailed statistical informationprecluded the studies from being combined in ameta-analysis.

Firestone and colleagues53 reported that MPHplus parent training was superior to parenttraining alone. Similarly, Kupietz and colleagues71

reported that MPH plus one-to-one readingtherapy was significantly better than one-to-onetherapy without medication. Brown andcolleagues42 did not present clear results forcomparisons between MPH plus cognitive therapycompared with cognitive therapy; however, theMD and 95% CI values demonstrate significantdifferences in favour of combined treatment.

Two of the remaining eight studies measuredoverall behaviour using the Abbreviated CTRS,77,78

and six evaluated inattention/overactivity using aIOWA Rating Scale.69,79–82,88 A number of thesestudies were conducted by Pelham and colleaguesand involved behaviour modification that tookplace over a Summer Treatment Programme(STP).77–81 The results from direct comparisonsbetween treatment groups were not always clearlypresented in these studies (see Appendix 12 formore detail). One study by Pelham and colleagues82

was conducted at home and in a Saturday school. In this study, MPH plus a behavioural programmewas significantly better than the behaviouralprogramme alone (p < 0.001).

Similarly, Smith and colleagues88 evaluatedmedium-dose MPH plus behavioural treatmentcompared with behavioural treatment alone. Theyreported that a significant difference was observedbetween all comparisons at the p < 0.05 level, butno further detail was given. Lastly, in the study byKolko and colleagues,69 medium-dose MPH plusbehaviour modification was reported to be betterfor scores on inattention/overactivity thanbehaviour modification alone (see Appendix 12).

Quality of lifeNo studies reported data on CGI. Pelham andcolleagues82 assessed global effectiveness as


46


47


TA

BLE

24

MPH

med

ium

dos

e (1

5–30

mg/

day)

plu

s no

n-dr

ug in

terv

entio

n ve

rsus

non

-dru

g in

terv

entio

n

Stud

yD

esig

nIn

terv

enti

on –

NA

ge

Dur

atio

n C

ore

outc

omes

(yea

rs)

(wee

ks)

Adm

inist

ered

onc

e da

ilyKu

piet

z, 1

98871

PM

PH (0

.7 m

g/kg

, o.d

.) vs

one

-to-

one

7–13

6 m

onth

sC

ore:

CT

RS: h

yper

activ

ityre

adin

g th

erap

y (6

mon

ths)

– 5

8Q

oL: n

ot r

epor

ted

AE:

not

rep

orte

d

Adm

inist

ered

tw

o or

mor

e tim

es d

aily

Brow

n, 1

98642

PM

PH (2

0.08

mg/

day,

b.d

.) vs

cog

nitiv

e 5.

7–13

.13

mon

ths

Cor

e: C

PRS:

hyp

erac

tivity

Inde

x; A

CTe

Rs: h

yper

activ

ityth

erap

y (1

1 w

eeks

, 22

sess

ions

) – 4

0Q

oL: n

ot r

epor

ted

AE:

not

rep

orte

d

Fire

ston

e, 1

98653

PM

PH (2

2 m

g/da

y, b

.d.)

vs p

aren

t tr

aini

ng

5–9

3 m

onth

sC

ore:

CT

RS: H

yper

activ

ity In

dex

(3 m

onth

s) –

134

QoL

: not

rep

orte

dA

E: n

ot r

epor

ted

Pelh

am, 1

98777

C (3

×)M

PH (2

0 m

g/da

y, b

.d.)

vs b

ehav

iour

6.

5–11

7C

ore:

no

hyp;

Abb

revi

ated

CT

RSm

odifi

catio

n (S

TP,

7 w

eeks

) – 1

3Q

oL: n

ot r

epor

ted

AE:

Sid

e Ef

fect

s C

heck

lists

(par

ents

, tea

cher

s, c

ouns

ello

rs)

Pelh

am, 1

99078

C (5

×)M

PH (2

0 m

g/da

y, b

.d.)

vs b

ehav

iour

8.

1–13

.28

Cor

e: n

o hy

p; A

bbre

viat

ed C

TRS

mod

ifica

tion

(ST

P, 8

wee

ks) –

22

QoL

: not

rep

orte

dA

E: S

ide

Effe

cts

Che

cklis

ts (p

aren

ts, t

each

ers,

cou

nsel

lors

)

Pelh

am, 1

99379

C (3

×)M

PH (1

6.2

mg/

day,

b.d

.) vs

beh

avio

ur

5.4–

9.9

8C

ore:

no

hyp;

IOW

A C

TRS

: ina

tten

tion/

over

activ

itym

odifi

catio

n (S

TP,

8 w

eeks

) – 3

1Q

oL: n

ot r

epor

ted

AE:

not

rep

orte

d

Smith

, 199

888C

(4×)

MPH

(25

mg/

day,

t.d

.s.)

vs b

ehav

iour

al

12–1

78

Cor

e: n

o hy

p; IO

WA

CT

RS: i

natt

entio

n/ov

erac

tivity

trea

tmen

t (S

TP,

8 w

eeks

) – 4

9Q

oL: n

ot r

epor

ted

AE:

sid

e-ef

fect

s ra

ting

from

0 (n

ot t

roub

ling)

to

3 (s

ever

e)(c

ouns

ello

r, pa

rent

s)

Pelh

am, 1

99981

C (7

×)M

PH (0

.75

mg/

kg/d

ay, t

.d.s

,) –

216–

126

Cor

e: n

o hy

p; IO

WA

-C: i

natt

entio

n/ov

erac

tivity

(tea

cher

san

d pa

rent

s)M

PH (0

.90

mg/

kg/d

ay, t

.d.s

.) –

21Q

oL: n

ot r

epor

ted

AE:

Pitt

sbur

gh S

ide

Effe

ct R

atin

g Sc

ale

(cou

nsel

lors

, tea

cher

s,pa

rent

s)

Pelh

am, 1

99980

C (5

×)M

PH (2

0 m

g/da

y, b

.d.)

– 26

5.8–

12.7

6C

ore:

no

hyp;

IOW

A-C

: ina

tten

tion/

over

activ

ity (t

each

ers

and

pare

nts)

QoL

: not

rep

orte

dA

E: S

ide

Effe

cts

Che

cklis

ts (t

each

ers,

cou

nsel

lors

, par

ents

)

cont

inue

d


48 TA

BLE

24

MPH

med

ium

dos

e (1

5–30

mg/

day)

plu

s no

n-dr

ug in

terv

entio

n ve

rsus

non

-dru

g in

terv

entio

n (c

ont’d

)

Stud

yD

esig

nIn

terv

enti

on –

NA

ge

Dur

atio

n C

ore

outc

omes

(yea

rs)

(wee

ks)

Kolk

o, 1

99969

C (4

×)M

PH (0

.60

mg/

kg/d

ay, b

.d.)

vs b

ehav

iour

7–

136

Cor

e: n

o hy

p; IO

WA

CT

RS: i

natt

entio

n/ov

erac

tivity

mod

ifica

tion

(ST

P, 3

wee

ks) –

22

QoL

: not

rep

orte

dA

E: S

SERS

Pelh

am, 2

00182

C (3

×)M

PH (1

5, 3

0 or

45

mg/

day,

t.d

.s.)

vs

6–12

3C

ore:

no

hyp;

IOW

A-C

: ina

tten

tion/

over

activ

ity (t

each

ers

pare

nt t

rain

ing,

tea

cher

con

sulta

tion

and

and

pare

nts)

poin

t sy

stem

s (3

wee

ks) –

70

QoL

: no

CG

I; gl

obal

effe

ctiv

enes

s (p

aren

t/te

ache

r)A

E: q

uest

ions

reg

ardi

ng a

dver

se e

vent

s, s

leep

qua

lity,

appe

tite

and

tics

(par

ents

)Sp

onta

neou

s re

port

s of

adv

erse

eve

nts

C, c

ross

over

tria

l (nu

mbe

r of

cro

ssov

ers)

; CG

I, C

linic

al G

loba

l Im

pres

sion;

CPR

S, C

onne

rs’ P

aren

t Ra

ting

Scal

e; C

TRS

, Con

ners

’ Tea

cher

Rat

ing

Scal

e; P,

par

alle

l tria

l; ST

P, Su

mm

erTr

eatm

ent

Prog

ram

me.

TA

BLE

25

Resu

lts fo

r hyp

erac

tivity

[M

PH m

ediu

m d

ose

(15–

30m

g/da

y) p

lus

non-

drug

inte

rven

tion

vers

us n

on-d

rug

inte

rven

tion]

Stud

ySc

ale

MP

H m

ediu

m d

ose

+ n

on-d

rug:

mea

n (S

D)

Non

-dru

g: m

ean

(SD

)M

ean

diffe

renc

e

5–9

year

sFi

rest

one,

198

653M

D (9

5% C

I)C

TRS

(Hyp

erac

tivity

Inde

x)0.

89 (0

.49)

1.37

(0.5

7)–0

.40

(–0.

56 t

o –0

.24)

5–13

yea

rsBr

own,

198

642C

PRS

(Hyp

erac

tivity

Inde

x)

13.7

8 (8

.14)

21.1

0 (5

.65)

–5.3

4 (–

9.58

to

–1.1

0)A

CTe

RS (h

yper

activ

ity)

14.3

3 (4

.66)

19.6

0 (2

.63)

–2.5

8 (–

5.09

to

–0.0

7)

7–13

yea

rsKu

piet

z, 1

98871

CT

RS (h

yper

activ

ity)

2.03

(0.6

4)2.

89 (0

.69)

–0.6

4 (–

0.93

to

–0.3

5)

Low

er s

core

s re

pres

ent

a be

tter

beh

avio

ural

out

com

e.A

CTe

RS, A

DD

-H C

ompr

ehen

sive

Teac

her

Ratin

g Sc

ale;

CPR

S, C

onne

rs’ P

aren

t Ra

ting

Scal

e; C

TRS

, Con

ners

’ Tea

cher

Rat

ing

Scal

e.


49


assessed by parents and teachers. In this study, thenon-drug intervention involved a behaviouralprogramme incorporating parent training, teacherconsultation and a point system. They observedsignificant differences in favour of the MPH plusnon-drug intervention group (p < 0.001 for bothparent and teacher ratings).

Adverse eventsOf the 11 trials, six contributed data to theanalysis of adverse events. Pelham81 examined thetolerability of two medium doses of MPH, 0.75and 0.90 mg/kg/day. No significant differences inthe incidence of headache,80–82,88 stomachache80–82,88 or insomnia77,78,80–82,88 were detectedbetween treatment phases in any trial reportingthese outcomes. Participants in the Pelham82 trialdisplayed significantly reduced appetite in theMPH treatment phase compared with the placebophase (RR = 5.67; 95% CI 1.74 to 18.48). None ofthe other three trials80,81,88 reporting this outcomedetected differences in incidence. No studyreported data on weight.

SummaryMPH plus non-drug interventions (one-to-onereading therapy, parent training and cognitivetherapy) were found to be better than the non-drug intervention alone. None of the studies inthis category evaluated CGI as a core outcome. No significant differences in the incidence ofheadache, stomach ache or insomnia werereported between treatment groups, although onestudy reported significantly reduced appetite inthe MPH group. Overall, these studies did notscore very well in the quality assessment, and anyresults should be interpreted with caution.

MPH high dose (>30 mg/day) plus non-drugintervention versus non-drug interventionSeven trials (presented in eight papers) evaluatedhigh-dose (>30 mg/day) immediate-release MPHplus non-drug intervention compared with a non-drug intervention (Table 26; with additionalinformation in Appendix 12). All of these studiesexamined MPH administered two or more timesdaily.

The majority of studies in this category did notassess hyperactivity, but examinedinattention/overactivity using a Conners’ scale as amain outcome measure.69,79,80,88 The results ofcomparisons between treatment groups were notclearly presented in most of these studies.

HyperactivityThree of the studies reported results onhyperactivity,51,65,86 but one of them presented

data in graphs only and was not included inTable 27.51 The remaining two studies bothassessed hyperactivity using Conners’ scales andalso other measures. In Schachar and colleagues’study,86 the non-drug intervention was parenttraining or support, and in the Klein andAbikoff65 study, the non-drug interventioninvolved parent and teacher education. Theseparallel studies both reported significantdifferences in favour of MPH plus non-drugintervention compared with non-drug interventionwhen assessed by teachers, but not when assessedby parents.

Quality of lifeElia and colleagues51 assessed CGI as assessed byphysicians. They reported that children receivingMPH plus a behaviour modification programmeand a low monoamine diet had better behaviourthan children in the non-drug intervention group(p < 0.05). In addition, Klein and Abikoff65

evaluated Clinical Global Improvement as assessedby teachers, mothers and psychiatrists. After8 weeks of treatment, the physicians rated 97% ofchildren in the MPH plus parent and teachereducation group to be improved compared with50% in the non-drug intervention group.

Adverse eventsOf the seven trials comparing a high dose of MPHplus a non-drug intervention with a non-drugintervention alone, three informed analysis ofadverse events. Smith and colleagues88 examinedthe tolerability of two high doses of MPHcombined with a non-drug intervention – 50 and75 mg daily. No differences in the incidence ofheadache80,88 or stomach ache80,88 were detected.Participants in Elia and colleagues’ trial51 sufferedfrom loss of appetite and insomnia to asignificantly greater extent in the MPH treatmentphase of the trial (RR = 83.00; 95% CI 5.25 to1311.65 and RR = 2.92; 95% CI 1.80 to 4.75,respectively). The remaining two trials did notdetect differences in the incidence of insomnia orreduced appetite. Heterogeneity of results did notappear to be explained by participant age,outcome measurement or trial duration. Data onweight were not presented in any of the trials inthis comparison group.

SummaryTwo studies presented results for hyperactivityusing different scales, including those ofConners.65,86 In these studies, the non-druginterventions were parent training or support, andparent and teacher education, respectively. Forboth studies, the results were generally significant


50 TA

BLE

26

MPH

hig

h do

se (

>30

mg/

day)

plu

s no

n-dr

ug in

terv

entio

n ve

rsus

non

-dru

g in

terv

entio

n

Stud

yD

esig

nIn

terv

enti

on –

NA

ge

Dur

atio

n C

ore

outc

omes

(yea

rs)

(wee

ks)

Adm

inist

ered

tw

o or

mor

e tim

es d

aily

Elia

, 199

151C

(3×)

MPH

(25–

90 m

g/da

y, b

.d.)

vs b

ehav

iour

6–

129

Cor

e: r

esul

ts p

rese

nted

in g

raph

s on

lym

odifi

catio

n +

die

t (9

wee

ks) –

48

QoL

: CG

I (ph

ysic

ian)

; C-G

AS

AE:

ST

ESS

(phy

sicia

n, p

aren

ts)

Chi

ldre

n’s

Psyc

hiat

ric R

atin

g Sc

ale:

ner

vous

man

neris

ms,

obse

ssiv

e th

inki

ng

Pelh

am, 1

99379

C (3

×)M

PH (3

2 m

g/da

y, b

.d.)

vs b

ehav

iour

al

5.4–

9.9

8C

ore:

no

hyp;

IOW

A C

TRS

: ina

tten

tion/

over

activ

itym

odifi

catio

n (S

TP,

8 w

eeks

) – 3

1Q

oL: n

ot r

epor

ted

AE:

not

rep

orte

d

Scha

char

, 199

786an

d P

MPH

(31.

4 m

g/da

y, b

.d.)

vs p

aren

t tr

aini

ng

6–12

4 m

onth

sC

ore:

IOW

A-C

: hyp

erac

tivity

(par

ent)

; IO

WA

-C:

Dia

mon

d, 1

99987

or p

aren

t su

ppor

t (4

mon

ths)

– 9

1hy

pera

ctiv

ity–i

natt

entiv

enes

s (t

each

er);

TIP

: hyp

erac

tivity

–im

pulsi

vene

ss (t

each

er a

nd p

aren

t)Q

oL: n

ot r

epor

ted

AE:

Bar

kley

10-

poin

t sc

ale:

phy

siolo

gica

l, af

fect

ive,

ove

rfoc

us,

tics

(par

ents

, tea

cher

s)

Kle

in, 1

99765

PM

PH (1

.48

mg/

kg/d

ay, b

.d.?)

vs

BI +

par

ent

6–12

12C

ore:

CT

RS/C

PRS:

hyp

erac

tivity

; Hom

e H

yper

activ

ity S

cale

an

d te

ache

r ed

ucat

ion

(4 w

eeks

) – 8

6(p

aren

ts)

QoL

: CG

I (te

ache

r/m

othe

rs/p

sych

iatr

ists)

AE:

not

rep

orte

d

Smith

, 199

888C

(4×)

MPH

(50

mg/

day,

t.d

.s.)

vs b

ehav

iour

al

12–1

78

Cor

e: n

o hy

p; IO

WA

CT

RS: i

natt

entio

n/ov

erac

tivity

trea

tmen

t (S

TP,

8 w

eeks

) – 4

9Q

oL: n

ot r

epor

ted

AE:

sid

e-ef

fect

s ra

ting

from

0 (n

ot t

roub

ling)

to

3 (s

ever

e)

MPH

(75

mg/

day

MPH

) beh

avio

ural

(c

ouns

ello

r, pa

rent

s)tr

eatm

ent

(ST

P, 8

wee

ks) –

49

Kolk

o, 1

99969

C (4

×)M

PH (1

.2 m

g/kg

/day

, b.d

.) vs

beh

avio

ur

7–13

6C

ore:

no

hyp;

IOW

A C

TRS

: ina

tten

tion/

over

activ

itym

odifi

catio

n (S

TP,

3 w

eeks

) – 2

2Q

oL: n

ot r

epor

ted

AE:

SSE

RS

Pelh

am, 1

99980

C (5

×)M

PH (3

5 m

g/da

y, b

.d.)

– 26

5.8–

12.7

6C

ore:

no

hyp;

IOW

A-C

: ina

tten

tion/

over

activ

ity (t

each

ers

and

pare

nts)

QoL

: not

rep

orte

dA

E: S

ide

Effe

cts

Che

cklis

t (t

each

ers,

cou

nsel

lors

, par

ents

)

C, c

ross

over

tria

l (nu

mbe

r of

cro

ssov

ers)

; C-G

AS,

Chi

ldre

n’s

Glo

bal A

sses

smen

t Sc

ale;

CG

I, C

linic

al G

loba

l Im

pres

sion;

CPR

S, C

onne

rs’ P

aren

t Ra

ting

Scal

e; C

TRS

, Con

ners

’ Tea

cher

Ratin

g Sc

ale;

IOW

A-C

, IO

WA

Con

ners

’ Rat

ing

Scal

e; P,

par

alle

l tria

l; SS

ERS,

Stim

ulan

t D

rug

Side

Effe

cts

Ratin

g Sc

ale;

TIP,

Tel

epho

ne In

terv

iew

Pro

be.


51


TA

BLE

27

Resu

lts fo

r hyp

erac

tivity

[M

PH h

igh

dose

(>

30 m

g/da

y) p

lus

non-

drug

inte

rven

tion

vers

us n

on-d

rug

inte

rven

tion]

Stud

ySc

ale

MP

H h

igh

dose

+ n

on-d

rug:

N

on-d

rug:

mea

n (S

D)

Mea

n di

ffere

nce

mea

n (S

D)

6–12

yea

rsSc

hach

ar, 1

99786

MD

(95%

CI)

IOW

A-C

(hyp

erac

tivity

) (pa

rent

s)

1.2

(0.7

)1.

3 (0

.7)

–0.1

0 (–

0.33

to

0.13

)IO

WA

-C (h

yper

activ

ity–i

natt

entiv

enes

s) (t

each

ers)

0.9

(0.7

)1.

7 (0

.7)

–0.8

0 (–

1.05

to

–0.5

5)T

IP (h

yper

activ

ity–i

mpu

lsive

ness

) (te

ache

rs)

0.8

(0.7

)1.

3 (1

.1)

–0.9

0 (–

1.29

to

–0.5

1)T

IP (h

yper

activ

ity–i

mpu

lsive

ness

) (pa

rent

s)1.

2 (1

.1)

0.9

(0.9

) 0.

20 (–

0.22

to

0.62

)

Kle

in, 1

99765

CT

RS (h

yper

activ

ity)

0.8

(0.4

)1.

5 (0

.6)

–0.7

0 (–

0.84

to

–0.5

6)C

PRS

(hyp

erac

tivity

)0.

6 (0

.3)

0.8

(0.4

)0.

20 (0

.11

to 0

.29)

Hom

e H

yper

activ

ity S

cale

(par

ents

)1.

8 (0

.51)

2.4

(0.8

6)0.

20 (–

0.15

to

0.55

)

Low

er s

core

s re

pres

ent

a be

tter

beh

avio

ural

out

com

e.C

PRS,

Con

ners

’ Par

ent

Ratin

g Sc

ale;

CT

RS, C

onne

rs’ T

each

er R

atin

g Sc

ale;

TIP,

Tel

epho

ne In

terv

iew

Pro

be.

TA

BLE

28

ER-M

PH lo

w d

ose

(≤20

mg/

day)

plu

s no

n-dr

ug in

terv

entio

n ve

rsus

non

-dru

g in

terv

entio

n

Stud

yD

esig

nIn

terv

enti

on –

NA

ge

Dur

atio

n C

ore

outc

omes

(yea

rs)

(wee

ks)

Adm

inist

ered

onc

e da

ilyPe

lham

, 198

777C

(3×)

Sust

aine

d-re

leas

e M

PH (S

low

Rel

ease

Rita

lin,

6.5–

117

Cor

e: n

o hy

p; A

bbre

viat

ed C

onne

rs’ R

atin

g Sc

ale

(tea

cher

SR

-20)

(20

mg/

day,

o.d

.) vs

beh

avio

ur

ratin

g)

mod

ifica

tion

(ST

P, 7

wee

ks) –

13

QoL

: not

rep

orte

dA

E: S

ide

Effe

cts

Che

cklis

ts (p

aren

ts, t

each

ers,

cou

nsel

lors

)

Pelh

am, 1

99078

C (5

×)Su

stai

ned-

rele

ase

MPH

(Slo

w R

elea

se R

italin

, 8.

1–13

.28

Cor

e: n

o hy

p; A

bbre

viat

ed C

onne

rs’ R

atin

g Sc

ale

(tea

cher

s SR

-20)

(20

mg/

day,

o.d

.) vs

beh

avio

ur

and

coun

sello

rs)

mod

ifica

tion

(ST

P, 8

wee

ks) –

22

QoL

: not

rep

orte

dA

E: S

ide

Effe

cts

Che

cklis

ts (p

aren

ts, t

each

ers,

cou

nsel

lors

)

C, c

ross

over

tria

l (nu

mbe

r of

cro

ssov

ers)

; ST

P, Su

mm

er T

reat

men

t Pr

ogra

mm

e.

(in favour of the MPH group) when assessed by teachers and non-significant when assessed byparents. One study reported results for CGI51

and one measured Clinical Global Improvement.65

Both studies reported that children receivingMPH in addition to a behavioural modificationprogramme, or parent and teacher education,showed improvement compared with children receiving only non-drug treatments.Regarding adverse events, no differences were observed between treatment groups forheadache or stomach ache. One study reported a higher incidence of loss of appetite andinsomnia in the MPH group, whereas the othertwo did not. However, these studies did not score very well in the quality assessment, and any results should be interpreted with caution.

ER-MPH low dose (�20 mg/day) plus non-drugintervention versus non-drug interventionTwo studies evaluated low-dose (≤ 20 mg/day) ER-MPH plus non-drug intervention comparedwith a non-drug intervention (Table 28; withadditional information in Appendix 12).77,78 Bothstudies were crossover trials conducted by Pelhamand colleagues that examined the effectiveness ofER-MPH in association with a behaviourmodification programme. Neither reportedhyperactivity as a core outcome measure, but bothmeasured behaviour using the AbbreviatedConners’ Rating Scale as measured by teachers(and counsellors). Only one significant result wasreported: Pelham and colleagues78 foundbehaviour was improved in children receiving ER-MPH plus behaviour modification in comparisonwith children receiving only behaviourmodification (p < 0.05) when assessed bycounsellors.

Adverse eventsNeither study displayed significant differencesbetween treatments in the incidence of insomnia.Data regarding other adverse events or weightwere not adequately reported to be included in theanalysis.

SummaryNo studies in this category measured hyperactivityor CGI as outcome measures. Only data oninsomnia could be evaluated, and no differenceswere observed between treatment groups. Thesestudies did not score very well in the qualityassessment, and the results should be interpretedwith caution.

ER-MPH medium dose (20–40 mg/day) plus non-drug intervention versus non-drug interventionOnly one study evaluated medium-dose(20–40 mg/day) extended-release MPH plus non-drug intervention compared with a non-drugintervention (Table 29; with additional informationin Appendix 12). In this crossover trial, the non-drug intervention involved a behaviouralprogramme incorporating parent training, teacherconsultation and a point system.82 One of the mainoutcomes examined was inattention/overactivity asmeasured using the IOWA-C. Behaviour wassignificantly improved in the combined treatmentgroup compared with the non-drug interventiongroup when assessed by teachers, parents andcounsellors (see Appendix 12).

Although Pelham and colleagues82 did notexamine CGI, they did measure globaleffectiveness (as assessed by parents and teachers).They observed that consistently higherpercentages of children rated better in thecombined treatment compared with the non-


52

TABLE 29 ER-MPH medium dose (20–40 mg/day) plus non-drug intervention versus non-drug intervention


Administered once dailyPelham, 200182 C (3×) OROS MPH (Concerta) (18, 36 6–12 3 Core: no hyp; IOWA-C

or 54 mg/day, o.d.) vs parent (inattention/overactivity) (teacher, training, teacher consultation and parent, counsellor)point systems (3 weeks) – 70 QoL: no CGI; global effectiveness

(parent/teacher)AE: questions regarding adverseevents, sleep quality, appetite,and tics (parents) plusspontaneous reporting

C, crossover trial (number of crossovers); CGI, Clinical Global Impression.

intervention group (p < 0.001 for parent andteacher scores) (see Appendix 12).

Adverse eventsParticipants suffered from a significantly greaterloss of appetite in the MPH group compared withplacebo (RR = 4.33; 95% CI 1.29 to 14.54).Differences in the incidence of headache, stomachache or insomnia were not detected. Data onweight were not reported.

SummaryNo studies in this category measured hyperactivityor CGI as outcome measures. Of the four adverseevents, only loss of appetite was observed to besignificantly greater in the MPH group than in theplacebo group. This study did not score very wellin the quality assessement, and the results shouldbe interpreted with caution.

DEX versus placeboDEX medium dose (10–20 mg/day) versusplaceboTwo studies evaluated medium-dose(10–20 mg/day) DEX compared with placebo(Table 30; with additional information in Appendix 12). In both studies, DEX wasadministered twice per day and evaluated using aparallel design.

HyperactivityBoth studies examined hyperactivity as one of thecore outcome measures. However, Gillberg andcolleagues56 presented their results in graph formonly, and their results could not be included inTable 31. Conners and colleagues45 reported asignificant difference between DEX and placebowhen using a symptom checklist, but not whenmeasured using the parent questionnaire.


53


TABLE 30 DEX medium dose (10–20 mg/day) versus placebo


Administered two or more times dailyConners, 197245 P DEX (20 mg/day, b.d.) – 28 6–12 8 Core: symptom checklist:

hyperactivity; parentquestionnaire: hyperactivityQoL: CGIAE: no specific scale reported

Gillberg, 199756 P DEX (17 mg/day, b.d.) – 32 6–11 15 months Core: CPRS: impulsivity/hyperactivity; CTRS: hyperactivityQoL: no CGIAE: incidence of 20 adverseevents; weight

P, parallel, trial; CPRS, Conners’ Parent Rating Scale; CGI, Clinical Global Impression.

TABLE 31 Results for hyperactivity [DEX medium dose (10–20 mg/day) versus placebo]

Study Scale DEX medium dose: Placebo: p-Value mean (SD) mean (SD) (if reported)a

6–12 yearsConners, 197245 Symptom checklist 6.2 (SD not reported) 13.3 (SD not reported) S – NR

(hyperactivity)Parent questionnaire 10.6 13.4 NS(hyperactivity)

a Note that owing to the overall poor reporting of study methodology, p-Values should be interpreted with caution. Lower scores represent a better behavioural outcome.NS, not significant; S – NR, significant (value not reported).

Quality of lifeConners and colleagues45 evaluated ClinicalGlobal Improvement as measured by a clinician.The authors reported that after 8 weeks, 33% weremuch improved in the DEX group compared with9% in the placebo group.

Adverse eventsNo significant differences in the incidence ofheadache, loss of appetite, stomach ache orinsomnia were detected between participants ofthe trial adequately reporting these outcomes.56

Data on weight were not reported separately forthose on DEX versus those on placebo.

SummaryOne study presented reproducible results forhyperactivity.45 In this study, the results weresignificant when assessed using the symptomchecklist, but not when assessed using the parentquestionnaire. This study also reported on ClinicalGlobal Improvement, and reported that childrenin the DEX group were more often improvedcompared with the placebo group. No significantdifferences in the incidence of headache, loss ofappetite, stomach ache or insomnia were reportedin the one study that presented data on theseadverse events.56 The studies did not score verywell in the quality assessment, and the resultsshould be interpreted with caution.

DEX high dose (>20 mg/day) versus placeboThree studies evaluated high dose (>20 mg/day)DEX compared with placebo (Table 32; withadditional information in Appendix 12). Of these,two appeared to have examined DEXadministered once daily and one examined DEXadministered twice per day.

Only one of the studies evaluated hyperactivity asa core outcome, using a number of differentscales.36 In this study, children in the DEX grouphad consistently better scores than children in theplacebo group; however, the authors did notreport results for any statistical comparisons (seeTable 33). Another study reported on behaviourratings as assessed by teachers and parents47 (seeAppendix 12) and the final study reported only onadverse events (discussed separately below).58

Quality of lifeArnold and colleagues36 reported on globalratings as assessed by clinicians. They reportedthat children in the DEX group rated better thanthose in the placebo group (p < 0.01).

Adverse eventsOf the three trials comparing a high dose of DEXwith placebo, one reported adequate data for theanalysis of adverse events.58 In this trial,participants assigned to DEX suffered from loss of


54

TABLE 32 DEX high dose (>20 mg/day) versus placebo


Administered once dailyArnold, 197636 C (3×) DEX (mean 21.75 mg/day, 4.6–12 12 Core: Parents’ Behaviour

o.d.?) – 31 Checklist: hyperactivity; Conners’ Teachers’ Behaviour Checklist: hyperactivityQoL: global ratings (clinicians)AE: Parents’ Behaviour Checklist:somatic complaints; Conners’Teachers’ Behaviour Checklist:lack of health; weight

Conrad, 197147 P DEX (10–20 mg/day, o.d.?) – 17 4–6 4–6 months Core: no hyp; behaviour ratings (teacher and parent) QoL: not reportedAE: not reported

Administered two or more times dailyGreenberg, P DEX (mean 25 mg/day, b.d.) – 17 6.5–11 8 Core: not reported197258 QoL: not reported

AE: incidence of side-effects

C, crossover trial (number of crossovers); P, parallel trial.

appetite to a significantly greater extent thanthose assigned to placebo (RR = 3.82; 95% CI1.08 to 13.58). Participants did not, however,suffer from headache, stomach ache or insomniasignificantly more often when assigned to DEX.Data on weight were inadequately reported.

SummaryOne study that evaluated hyperactivity reportedthat children in the DEX group had improvedbehaviour compared with the placebo group.36

However, no statistical analyses were presented.The same study presented results for globalratings (as assessed by clinicians) and reportedimprovements with medication. Generally, thisstudy rated well in the quality assessement,whereas the other studies did not score very well.One of these other studies reported usable data onadverse events.58 Loss of appetite was significantlygreater in the DEX group compared with placebo,but this was not observed for headache, stomachache or insomnia.

DEX-TR versus placeboOne study evaluated 10–15 mg/day time-releaseDEX (DEX-TR) administered once daily (Table 34;with additional information in Appendix 12).

HyperactivityArnold and colleagues38 reported that DEX time-release capsules were significantly better thanplacebo using the Hyperactivity Index of theCTRS (Table 35).

Quality of lifeArnold and colleagues38 also reported that global ratings, as assessed by a psychiatrist, were improved in the treatment group comparedwith the placebo group (p < 0.05) (see Appendix 12).

Adverse eventsData on weight were not adequately presented inthis trial.

SummaryOne study evaluated DEX-TR compared withplacebo.38 This study reported significantimprovements in the treatment group comparedwith the placebo group for both hyperactivity andpsychiatrist-assessed global ratings. Owing to thepoor reporting of some methodological criteria,the results from this study should be interpretedwith caution.


55


TABLE 34 DEX-TR versus placebo


Administered once dailyArnold, 198938 C (3×) DEX-TR (10–15 mg/day, 6–12 3 Core: CTRS Hyperactivity Index

o.d.) – 18 QoL: global ratings (psychiatrist)AE: weight

C, crossover trial (number of crossovers); CTRS, Conners’ Teacher Rating Scale.

TABLE 33 Results for hyperactivity [DEX high dose (>20 mg/day) versus placebo]

Study Scale DEX high dose: Placebo: p-Value mean (SD) mean (SD) (if reported)

4–12 yearsArnold, 197636 Parents’ Behaviour 16.68 (6.59) 20.81 (6.83) NR

Checklist (hyperactivity)Conners’ Teachers 16.80 (5.54) 21.27 (5.63) NR

Behaviour Checklist (hyperactivity)

Davids’ Hyperkinetic Rating Scale (hyperactivity)

(parents) 3.97 (1.40) 4.58 (1.26) NR(teachers) 3.70 (1.49) 4.90 (1.30) NR

Lower scores represent a better behavioural outcome.NR, not reported.

DEX medium dose (10–20 mg/day) plus non-drugintervention versus placeboOne study evaluated medium-dose (10–20 mg/day)DEX plus non-drug intervention compared withplacebo (Table 36; with additional information inAppendix 12). In this study, DEX wasadministered once daily, and the interventioninvolved tutoring sessions twice per week. Conradand colleagues47 did not evaluate hyperactivity orCGI, but did report on behaviour ratings byteachers and parents. The authors reportedimprovements in the combined treatment groupcompared with placebo (see Appendix 12).

SummaryNo studies in this category measured hyperactivity,CGI or adverse events as outcome measures.

DEX versus non-drug interventionDEX high dose (>20 mg/day) versus non-druginterventionOne study evaluated high-dose (>20 mg/day) DEX compared with a non-drug intervention(Table 37; with additional information in Appendix 12). As presented above, thisintervention involved tutoring sessions twice perweek. Conrad and colleagues47 did not evaluatehyperactivity or CGI, but reported on behaviourratings by teachers and parents. The authorsreported greater improvements in the treatmentgroup than in the tutoring group, but the twogroups were not directly compared (see Appendix 12).


56

TABLE 35 Results for hyperactivity (DEX-TR versus placebo)

Study Scale DEX-TR: mean (SD) Placebo: p-Value mean (SD) (if reported)a

6–12 yearsArnold, 198938 CTRS (Hyperactivity Index) 1.39 (0.76) 2.10 (0.47) <0.05

a Note that owing to poor reporting for some aspects of the study methodology, p-Values should be interpreted withcaution.

Lower scores represent a better behavioural outcome.CTRS, Conners’ Teacher Rating Scale.

TABLE 36 DEX medium dose (10–20 mg/day) plus non-drug intervention versus placebo


Administered once dailyConrad, 197147 P DEX (10–20 mg/day, o.d.?) plus 4–6 4–6 Core: no hyp; behaviour ratings

prescriptive tutoring – 17 (teacher, parent)QoL: not reportedAE: not reported

P, parallel trial.

TABLE 37 DEX high dose (>20 mg/day) versus non-drug intervention


Administered once dailyConrad, 197147 P DEX (10–20 mg/day, o.d.?) vs 4–6 4–6 Core: no hyp; behaviour ratings


P, parallel trial.

SummaryNo studies in this category measured hyperactivity,CGI or adverse events as outcome measures.

DEX medium dose (10–20 mg/day) plus non-drugintervention versus non-drug interventionTwo studies evaluated medium-dose(10–20 mg/day) DEX plus non-drug interventioncompared with a non-drug intervention (Table 38;with additional information in Appendix 12). The study by Conrad and colleagues47 did notreport on hyperactivity, CGI or adverse events;however, they did report on behavioural ratings(see Appendix 12).

HyperactivityJames and colleagues62 evaluated hyperactivity

using three different scales. In this crossoverstudy, immediate-release DEX was administeredonce weekly in addition to formal academicinstruction and therapeutic recreation (e.g. sports,art therapy, structured social skills sessions).James and colleagues62 observed significantresults in favour of combined treatment whenhyperactivity was measured using the CTRS andthe Children’s Psychiatric Rating Scale, but notwhen measured using the CPRS (Table 39). Thisstudy did not examine CGI as an outcomemeasure.

Adverse eventsNo data reported in James and colleagues’ trial62

could usefully contribute to the analysis of adverseevents.


57


TABLE 39 Results for hyperactivity [DEX medium dose (10–20 mg/day) plus non-drug intervention versus non-drug intervention]

Study Scale DEX medium dose + Non-drug: p-Value non-drug: mean (SD) mean (SD) (if reported)a

6–12 yearsJames, 200162 CTRS (hyperactivity) 50.5 (5.4) 63.1 (12.6) S – NR

Children’s Psychiatric Rating Scale 2.5 (1.1) 3.8 (1.1) <0.001(hyperactivity)

CPRS (hyperactivity) 60.5 (14.7) 68.0 (14.5) 0.053

a Note that owing to poor reporting for some aspects of the study methodology, p-values should be interpreted withcaution.

Lower scores represent a better behavioural outcome.S – NR, significant (value not reported); CTRS, Conners’ Teacher Rating Scale; CPRS, Conners’ Parent Rating Scale.

TABLE 38 DEX medium dose (10–20 mg/day) plus non-drug intervention versus non-drug intervention


Administered once dailyConrad, 197147 P DEX (10–20 mg/day, o.d.?) vs 4–6 4–6 Core: no hyp; behaviour ratings


Administered once weeklyJames, 200162 C (4×) DEX (5–30 mg/kg once per 6.9–12.2 10 Core: CTRS: hyperactivity;

week for 2 weeks) vs formal Children’s Psychiatric Rating academic instruction (a.m.) and Scale: hyperactivity (recreation therapeutic recreation (p.m.) therapist rated); CPRS: (full time, 10 weeks) – 35 hyperactivity

QoL: not reportedAE: Stimulant Side Effect RatingScale (nurse); Barkley Side EffectRating Scale (parent); weight

C, Crossover trial (number of crossovers); CPRS, Conners’ Parent Rating Scale; CTRS, Conners’ Teacher Rating Scale; P, parallel trial.

SummaryOnly one study in this category examinedhyperactivity as an outcome.62 In this study,children who received DEX in combination withacademic instruction and therapeutic recreationhad better behaviour than those in the non-intervention group when assessed using the CTRSand the Children’s Psychiatric Rating Scale, butnot when assessed using the CPRS. No studies inthis category evaluated CGI. Owing to the poorreporting of some methodological criteria, theresults from this study should be interpreted withcaution.

DEX high dose (>20 mg/day) plus non-drugintervention versus non-drug interventionOne study evaluated high-dose (>20 mg/day) DEXplus non-drug intervention compared with a non-drug intervention (Table 40; with additionalinformation in Appendix 12). In this crossoverstudy, drug treatment was combined with amultidisciplinary behaviour modificationprogramme and low monoamine diet. Althoughthe authors evaluated hyperactivity and CGI (withresults favouring the combined treatment group),data were presented in graph form only and couldnot be reproduced.

Adverse eventsA significantly higher incidence of reducedappetite and insomnia was detected during theactive drug phase of this trial (RR = 93.00; 95%CI 5.90 to 1467.03 and RR = 3.00; 95% CI 1.85to 4.87, respectively). Data on headache, stomachache or weight were not reported.

SummaryNo studies in this category reported reproducibledata on hyperactivity or CGI. The one study inthis category reported significantly greaterincidence of loss of appetite and insomnia in thecombined treatment group.

DEX-SR plus non-drug intervention versus non-drug interventionTwo studies examined sustained-release DEX(DEX-SR) plus non-drug intervention comparedwith a non-drug intervention (Table 41; withadditional information in Appendix 12). In thestudy by Pelham and colleagues 1990,78

hyperactivity was not assessed, although theauthors did evaluate behaviour using theAbbreviated CTRS. They reported that behaviourwas improved in the combined treatment groupcompared with the behaviour modification group(p < 0.050). This study also reported data onadverse events (see below).

In addition to evaluating immediate-release DEX,James and colleagues62 examined DEX-SRadministered once weekly in combination withformal academic instruction and therapeuticrecreation. Hyperactivity was improved in thecombined treatment group compared with thenon-drug treatment group when assessed usingthe Children’s Psychiatric Rating Scale and theCPRS. Statistical comparisons for the CTRS werenot reported (Table 42). This study did notexamine CGI as an outcome measure.

Adverse eventsOf the two trials comparing DEX-SR and non-drug intervention with a non-drug interventionalone, one provided adequate data regarding theadverse events of interest.78 Differences in theincidence of insomnia were not found betweentreatment phases. Incidence of headache, stomachache or loss of appetite was not rated by patientsor parents. Neither study reported data on weight.

SummaryOf the two studies included in this category, oneevaluated hyperactivity as an outcome measure.62

In this study, DEX-SR in combination with formalacademic instruction and therapeutic recreation


58

TABLE 40 DEX high dose (>20 mg/day) plus non-drug intervention versus non-drug intervention


Administered two or more times a dayElia, 199151 C (3×) DEX (mean 10–45 mg/day, b.d.) 6–12 9 Core: CTRS: hyperactivity;

vs multidisciplinary behaviour CPQ: hyperactivitymodification programme and diet QoL: CGI (physician); C-GAS(twice daily, 11 weeks) – 48 AE: STESS (physician, parents)

C, crossover trial (number of crossovers); C-GAS, Children’s Global Assessment Scale; CGI, Clinical Global Impression;CPQ, Conners’ Parent Questionnaire; CTRS, Conners’ Teacher Rating Scale; STESS, Subject Treatment Emergent SymptomScale.

resulted in improved behaviour compared withnon-drug intervention alone. Owing to the poorreporting of some methodological criteria, theresults from this study should be interpreted withcaution. The other study reported on adverseevents of interest.78 In this study, no significantdifferences in the incidence of insomnia wereobserved. Neither study assessed CGI.

ATX versus placeboATX low/medium dose (<1.5 mg/kg/day) versusplaceboThree parallel studies evaluated low/medium-dose(<1.5 mg/kg/day) ATX compared with placebo(Table 43; with additional information in Appendix 12).

HyperactivityAll three studies examined hyperactivity/impulsivity using the ADHD Rating Scale (Table 44). One study also evaluated hyperactivityusing the Conners’ Parent Rating Scale – Revised(CPRS-R).73 The results from these studies werestatistically significant in favour of the drugtreatment group, with one exception. Michelsonand colleagues73 observed no difference betweenATX (0.5 mg/kg/day) and placebo when assessedusing the ADHD Rating Scale. The studies werenot pooled; however, given that all used the same scale and were parallel studies, meandifferences were calculated and are presented inFigure 15.


59


TABLE 41 DEX-SR plus non-drug intervention versus non-drug intervention


Administered once dailyPelham, 199078 C (5×) DEX-SR (10 mg/day, o.d.) vs 8.1–13.2 6.5 Core: no hyp; Abbreviated CTRS

broad spectrum behaviour QoL: not reportedmodification intervention (STP, AE: Side Effects Checklists 8 weeks) – 22 (parents/teachers/counsellors)

Administered once weeklyJames, 200162 C (4×) DEX-SR (5–30 mg/kg in two 6.9–12.2 10 Core: CTRS: hyperactivity;

doses, once per week) vs formal Children’s Psychiatric Rating academic instruction (a.m.) and Scale: hyperactivity (recreation therapeutic recreation (p.m.) therapist rated); (full time, 10 weeks) – 35 CPRS: hyperactivity

QoL: not reportedAE: Stimulant Side Effect RatingScale (nurse);Barkley Side Effect Rating Scale(parent); weight

C, crossover trial (number of crossovers); CPRS, Conners’ Parent Rating Scale; CTRS, Conners’ Teacher Rating Scale.

TABLE 42 Results for hyperactivity (DEX-SR plus non-drug intervention versus non-drug intervention)

Study Scale DEX medium dose + Non-drug: p-Value non-drug: mean (SD) mean (SD) (if reported)a

6–12 yearsJames, 200162 CTRS (Hyperactivity) 53.7 (9.1) 63.1 (12.6) NR

Children’s Psychiatric Rating Scale 2.3 (1.0) 3.8 (1.1) <0.001(hyperactivity)

CPRS (hyperactivity) 60 (15.6) 68.0 (14.5) <0.007

a Note that owing to poor reporting for some aspects of the study methodology, p-Values should be interpreted withcaution.

Lower scores represent a better behavioural outcome.CPRS, Conners’ Parent Rating Scale; CTRS, Conners’ Teacher Rating Scale; NR, not reported.


60

TABLE 43 ATX low/medium dose (<1.5 mg/kg/day) versus placebo


Administered once dailyMichelson, P ATX (1.0–1.5 mg/kg/day, o.d.) – 6–16 6 Core: ADHD-RS-IV: 200274 171 hyperactive/impulsive

QoL: CGIAE: 16 types of adverse effectsreported assessed by open-endedquestioning; weight

Kelsey, 200463 P ATX (mean 1.3 mg/kg/day, 6–12 8 Core: ADHD-RS: o.d.) – 197 hyperactive/impulsive

QoL: CGIAE: open-ended questioning

Administered twice dailyMichelson, P ATX (max. 0.5 mg/kg/day, 8–18 8 Core: CPRS-R: hyperactive; 200173 b.d.) – 297 ADHD-RS-IV:

hyperactive/impulsiveATX (max. 1.2 mg/kg/day, QoL: CGIb.d.) – 297 AE: open-ended questioning;

weight

ADHD-RS-IV, Attention Deficit/Hyperactivity Disorder Rating Scale IV – Parent Version: investigator-administered andscored; CGI, Clinical Global Impression; CPRS-R, Conners’ Parent Rating Scale – Revised; P, parallel trial.

TABLE 44 Results for hyperactivity [ATX low/medium dose (<1.5 mg/kg/day) versus placebo]

Study Scale ATX low/medium dose: Placebo: Mean mean (SD) mean (SD) differencea

6–12 yearsKelsey, 200463 ADHD-RS (hyperactive/impulsive) 11.0 (7.7) 16.3 (7.5) MD (95% CI)

See Figure 15

6–16 yearsMichelson, 200274 ADHD-RS-IV (hyperactive/ Baseline/change from Baseline/change See Figure 15

impulsive) baseline 15.7 (8.0)/ from baseline–5.7 (6.8) 15.3 (7.1)/

–2.1 (5.7)

8–18 yearsMichelson, 200173 ADHD-RS-IV (hyperactive/ Baseline/change from Baseline/change See Figure 15

impulsive) baseline from baselineATX 0.5: 16.9 (6.6)/–3.2 17.8 (7.4)/–4.8 (7.9) (5.6)ATX 1.2: See Figure 1516.9 (7.1)/–6.6 (7.1)

CPRS-R (hyperactivity) ATX 0.5: 10.3 (4.9)/–1.1 –3.00 (–4.73 to 12.0 (4.8)/–4.1 (4.4) (3.9) –1.27)ATX 1.2: –3.00 (–4.47 to 10.2 (5.1)/–4.1 (4.9) –1.53)

a [Confidential information removed].Lower scores represent a better behavioural outcome.ADHD-RS-IV, Attention Deficit/Hyperactivity Disorder Rating Scale IV – Parent Version: investigator-administered andscored; CPRS-R, Conners’ Parent Rating Scale – Revised.

Quality of lifeAll three studies reported CGI severity as anoutcome measure. Michelson and colleagues74 andKelsey and colleagues63 both reportedimprovements in the ATX group compared withplacebo (p < 0.001 and p < 0.05, respectively).Michelson and colleagues73 observed a significantdifference when 1.2 mg/kg/day was administered(p = 0.002), but not when 0.5 mg/kg/day wasadministered.

Adverse eventsAll studies in this category contributed to theanalysis of adverse events. No significantdifferences were detected between treatmentgroups in the incidence of headache,63,73,74

stomach ache63,73,74 or insomnia.73 However,participants in the ATX treatment groupsdisplayed significant reductions in appetite in twoof the three trials (Figure 16).

In the one trial that reported weight data,73 asignificantly different mean change was detectedin the 0.5 and 1.2 mg/kg/day ATX treatmentgroups compared with placebo (RR = -–1.40; 95%CI –1.88 to –0.92 and RR = –2.10; 95% CI –2.56to –1.64, respectively). The 0.5 mg/kg/day groupincreased in overall mean weight to a lesserdegree than the placebo group; the 1.2 mg/kg/daygroup decreased in overall mean weight.

SummaryThree studies in this category examinedhyperactivity and CGI,63,73,74 with almost allresults favouring ATX over placebo. However, thelowest dose of ATX examined (0.5 mg/kg/day) didnot significantly improve behaviour whenmeasured using the ADHD Rating Scale or CGI.73

No differences were observed between treatmentgroups in the incidence of headache, stomachache or insomnia, whereas there was someevidence for a reduction in appetite with ATX.Whereas two of the studies did not score very wellin the quality assessement, the study by Michelsonand colleagues73 was deemed to be of goodquality. Hence the results from this study are likelyto be reliable.

ATX high dose (�1.5 mg/kg/day) versus placeboSix trials (published in five papers) evaluatedhigh-dose (≥1.5 mg/kg/day) ATX compared with placebo (Table 45; with additional informationin Appendix 12). Five trials used a parallel design and, with the exception of one study,95

all evaluated hyperactivity and CGI. The study by Wernicke and colleagues95 did evaluate ADHD Rating Scale Total Score andreported significant improvements in the ATX group compared with placebo (see Appendix 12).


61


Study

Michelson 2001 [0.5]Michelson 2001 [1.2]Michelson 2002Kelsey 2004

44 84 84126

–4.80 (7.90)–6.60 (7.10)–5.70 (6.80)–8.50 (7.50)

ATX (low/medium dose) WMD (fixed)95% CI

WMD (fixed)95% CI

–1.60–3.40–3.60–5.60

(–4.22 to 1.02)(–5.33 to –1.47)(–5.50 to –1.70)(–7.57 to –3.63)

0–5–10 5Favours placeboFavours ATX

10

N Mean (SD)

84 84 83 60

–3.20 (5.60)–3.20 (5.60)–2.10 (5.70)–2.90 (5.80)

PlaceboN Mean (SD)

FIGURE 15 Mean differences: ATX (low/medium dose) versus placebo

Study

Kelsey 2004Michelson 2001 [0.5]Michelson 2001 [1.2]Michelson 2002

23/131 3/44

10/84 17/85

ATX (low/medium)n/N

RR (fixed)95% CI

RR (fixed)95% CI

2.771.432.503.40

(1.00 to 7.66)(0.34 to 6.12)(0.82 to 7.66)(1.31 to 8.80)

10.50.20.1 52Favours placeboFavours ATX

10

4/63 4/84 4/84 5/85

Placebon/N

FIGURE 16 Relative risks of loss of appetite: ATX (low/medium dose) versus placebo


62 TA

BLE

45

ATX

hig

h do

se (

≥1.

5 m

g/kg

/day

) ve

rsus

pla

cebo

Stud

yD

esig

nIn

terv

enti

on –

NA

ge

Dur

atio

n C

ore

outc

omes

(yea

rs)

(wee

ks)

Adm

inist

ered

onc

e da

ilyW

eiss

, 200

494P

ATX

(max

. 1.8

mg/

kg/d

ay, o

.d.)

– 15

38–

127

Cor

e: A

DH

D-R

S-IV

– T

each

er: I

nv: h

yper

activ

e/im

pulsi

ve;

CPR

S-R:

hyp

erac

tivity

QoL

: CG

I (te

ache

r/in

vest

igat

or)

AE:

ope

n-en

ded

disc

ussio

n

Adm

inist

ered

tw

ice

daily

Mic

helso

n, 2

00173

PAT

X (m

ax. 1

.8 m

g/kg

/day

, b.d

.) –

297

8–18

8C

ore:

CPR

S-R:

hyp

erac

tive;

AD

HD

-RS-

IV–P

aren

t: In

v hy

pera

ctiv

ity/im

pulsi

vity

QoL

: CG

IA

E: o

pen-

ende

d qu

estio

ning

; wei

ght

Spen

cer,

2002

89P

ATX

(max

. 2.0

mg/

kg/d

ay, b

.d.)

– 14

77–

139

Cor

e: A

DH

D-R

S: h

yper

activ

e/im

pulsi

veQ

oL: C

GI

AE:

Uns

olic

ited

adve

rse

even

ts

Spen

cer,

2002

89P

ATX

(max

. 2.0

mg/

kg/d

ay, b

.d.)

– 14

47–

139

Cor

e: A

DH

D-R

S: h

yper

activ

e/im

pulsi

veQ

oL: C

GI

AE:

Uns

olic

ited

adve

rse

even

ts

Mic

helso

n, 2

00475

PAT

X (m

ean

1.56

mg/

kg/d

ay, b

.d.)

– 41

66–

159

mon

ths

Cor

e: A

DH

D-R

S: h

yper

activ

e/im

pulsi

ve;

CPR

S/C

TRS

: hyp

erac

tivity

QoL

: CG

IA

E: S

ome

adve

rse

even

ts r

epor

ted

Wer

nick

e, 2

00495

PAT

X (m

ax. 2

.0 m

g/kg

/day

, b.d

.) –

194

7–12

9C

ore:

no

hyp;

AD

HD

-RS:

tot

al s

core

QoL

: not

rep

orte

dA

E: B

erkl

ey B

ehav

iour

and

Adv

erse

Eve

nts

Que

stio

nnai

re –

Mod

ified

; ope

n-en

ded

ques

tions

AD

HD

-RS-

IV, A

tten

tion

Def

icit/

Hyp

erac

tivity

Diso

rder

Rat

ing

Scal

e IV

–Par

ent

(or

Teac

her)

Ver

sion:

inve

stig

ator

-adm

inist

ered

and

sco

red;

CG

I, C

linic

al G

loba

l Im

pres

sion;

CPR

S,C

onne

rs’ P

aren

t Ra

ting

Scal

e; C

PRS-

R, C

onne

rs’ P

aren

t Ra

ting

Scal

e –

Revi

sed;

CT

RS, C

onne

rs’ T

each

er R

atin

g Sc

ale;

P, p

aral

lel t

rial.

HyperactivityFive trials (published in four papers) evaluatedhyperactivity/impulsivity using the ADHD RatingScale (Table 46). Three of these studies also used aConners’ scale to measure hyperactivity.73,75,94 Allresults were significantly in favour of ATX, exceptwhen the CTRS was used in the study byMichelson and colleagues.75 The mean differencesfor ADHD Rating Scale are presented for each ofthe five studies in Figure 17.

Quality of lifeWeiss and colleagues94 presented results for CGISeverity and Clinical Global Improvement. For bothmeasures, children in the ATX group had betterbehaviour than children in the placebo group(p < 0.001). Similar results were observed in thestudy by Michelson and colleagues73 (CGI ADHDSeverity: p < 0.05), Spencer and colleagues89 (CGIADHD Severity: p = 0.003) and Michelson andcolleagues75 (CGI Severity: p = 0.003).


63


TABLE 46 Results for hyperactivity [ATX high dose (≥ 1.5 mg/kg/day) versus placebo]

Study Scale ATX high dose: Placebo: Mean mean (SD) mean (SD) differencea

6–15 yearsMichelson, 200475 Baseline/change from Baseline/change

baseline from baseline MD (95% CI)ADHD-RS-IV 7.2 (5.5)/3.1 (7.0) 7.1 (5.5)/5.9 (7.4) See Figure 17(hyperactive/impulsive)CPRS (hyperactivity) 4.5 (3.8)/1.5 (4.7) 4.6 (4.2)/3.1 (4.9) –1.60 (–2.62 to –0.58)CTRS (hyperactivity) 7.7 (5.1)/0.4 (5.2) 8.1 (5.5)/1.4 (4.6) –1.00 (–2.15 to 0.15)

7–13 yearsSpencer, 200289 ADHD-RS 19.3 (6.1)/–8.0 (7.4) 19.2 (5.5)/–2.5 (5.9) See Figure 17

(hyperactive/impulsive)

Spencer, 200289 ADHD-RS 16.8 (6.5)/–6.9 (6.6) 16.5 (6.1)/–2.9 (7.1) See Figure 17(hyperactive/impulsive)

8–12 yearsWeiss, 200494 ADHD-RS-IV–Teacher: Inv: [Confidential information removed]

(hyperactive/impulsive) CPRS-R (hyperactivity) subscale mean score (SD)

8–18 yearsMichelson, 200173 ADHD-RS-IV 17.6 (6.2)/–6.7 (7.5) 16.9 (6.6)/–3.2 (5.6) see Figure 17

(hyperactive/impulsive) CPRS-R (hyperactivity) 10.6 (4.6)/–4.3 (4.6) 10.3 (4.9)/–1.1 (3.9) –3.20 (–4.59 to –1.81)

a Note that owing to the overall poor reporting of study methodology, values should be interpreted with caution (with theexception of Spencer and colleagues;89 [Confidential information removed].

Lower scores represent a better behavioural outcome.ADHD-RS-IV: Attention Deficit/Hyperactivity Disorder Rating Scale IV–Parent Version: investigator-administered and scored;CPRS, Conners’ Parent Rating Scale; CTRS, Conners’ Teacher Rating Scale; CPRS-R, Conners’ Parent Rating Scale – Revised.

Study

Michelson 2001Spencer 2002aSpencer 2002bMichelson 2004Weiss 2004

82 64 63290

–6.70 (7.50)–8.00 (7.40)–6.90 (6.60) 3.10 (7.00)

ATX (high dose) WMD (fixed)95% CI

WMD (fixed)95% CI

–3.50–5.50–4.00–2.80

(–5.52 to –1.48)(–7.84 to –3.16)(–6.43 to –1.57)(–4.34 to –1.26)

0–5–10 5Favours placeboFavours ATX

10

N Mean (SD)

83 61 60123

–3.20 (5.60)–2.50 (5.90)–2.90 (7.10) 5.90 (7.40)

PlaceboN Mean (SD)


FIGURE 17 Mean differences: ATX high dose versus placebo

Adverse eventsOf the six trials comparing a high dose of ATXwith placebo, five presented data informative tothe analysis of adverse events. Adverse eventsresults were combined in the trials reported bySpencer and colleagues.89 No significantdifferences in the incidence of headache,73,89

stomach ache73,89 or insomnia73 were found.Participants in the ATX treatment groupsdisplayed a significantly greater loss of appetite intwo trials (Figure 18).

A detrimental effect of a high dose of ATX onmean change in weight was observed in all fourtrials (Figure 19).

SummaryHigh-dose ATX improved hyperactivity/impulsivityand QoL (as measured by CGI) compared withplacebo. No significant differences in theincidence of headache, stomach ache or insomniawere observed between groups, although childrenin the ATX had greater loss of appetite in two ofthe trials. The studies by Michelson andcolleagues73 and Spencer and colleagues89 ratedwell in the quality assessement, hence their resultsare likely to be reliable. The other studies did notscore as well, and their results should beinterpreted with caution.

ATX versus non-drug interventionNo studies evaluated ATX compared with non-drug intervention.

IR-MPH versus ER-MPHIR-MPH low dose (≤ 15 mg/day) versus ER-MPHlow dose (≤ 20 mg/day)One study evaluated low-dose (≤ 15 mg/day)immediate-release MPH (IR-MPH) compared withlow-dose (≤ 20 mg/day) extended-release MPH(ER-MPH) (Table 47; with additional informationin Appendix 12). This study by Fitzpatrick andcolleagues55 examined hyperactivity using twodifferent scales as assessed by both parents andteachers, and reported no significant differencesbetween IR-MPH and ER-MPH (Table 48). This study did not evaluate CGI, but did present comment ratings by parents and teachers(see Appendix 12). Generally, the authors found no differences between the MPHconditions.

Adverse eventsNo significant differences in the incidence ofheadache, loss of appetite, stomach ache orinsomnia were detected between treatments. Dataon weight were not adequately reported.

SummaryThe one study included in this category reportedno differences between low-dose IR-MPH and low-dose ER-MPH for hyperactivity or any adverse events.55 This study did not report onCGI. The quality of reporting for some aspects of study methodology was poor for this study,hence the results should be interpreted withcaution.


64

Study

Michelson 2001Spencer 2002Weiss 2004

10/85 28/129 24/101

ATX (high dose)n/N

RR (fixed)95% CI

RR (fixed)95% CI

2.472.996.18

(0.81 to 7.57)(1.47 to 6.08)(1.52 to 25.13)

10.50.20.1 52Favours placeboFavours ATX

10

4/84 9/124

2/52

Placebon/N

FIGURE 18 Relative risks of loss of appetite: ATX (high dose) versus placebo

Study

Michelson 2001Michelson 2004Spencer 2002Weiss 2004

81292127101

–0.50 (1.70) 1.20 (2.40)–0.50 (1.40)–0.67 (1.21)

ATX (high dose) WMD (fixed)95% CI

WMD (fixed)95% CI

–2.20–2.10–1.90–1.88

(–2.71 to –1.69)(–2.79 to –1.41)(–2.25 to –1.55)(–2.32 to –1.44)

0–5–10 5Favours placebo Favours ATX

10

N Mean (SD)

83124122 52

1.70 (1.60)3.30 (3.60)1.40 (1.40)1.21 (1.38)

PlaceboN Mean (SD)

FIGURE 19 Differences between mean weight change: ATX (high dose) versus placebo

IR-MPH high dose (>30 mg/day) versus ER-MPH medium dose (20–40 mg/day)Three studies evaluated high-dose (>30 mg/day)IR-MPH compared with medium-dose(20–40 mg/day) ER-MPH (Table 49; with additionalinformation in Appendix 12). Only one of thesestudies reported data on hyperactivity (Table 50).In this study by Wolraich and colleagues,97 nosignificant differences were reported between thetwo MPH treatment groups.


Quality of lifeAll three studies examined CGI. In the study byWolraich and colleagues,97 the percentages ofthose ‘much or very much improved’ were verysimilar between treatment groups (IR-MPH 47.2%vs ER-MPH 46.7%).

[Commercial information removed].

Adverse eventsIn the trial by Wolraich and colleagues,97

participants assigned to the extended release formof MPH suffered from a significantly higherincidence of headache (Figure 20).

In the trial by Wolraich and colleagues,97 nosignificant differences were observed with regardto loss of appetite or stomach ache. Data oninsomnia or weight were not reported in this trial.[Confidential information removed].

SummaryOne study in this category assessed hyperactivitywith no significant differences reported betweenhigh-dose IR-MPH and medium-dose ER-MPH.97 All three studies evaluated CGI; the


65


TABLE 47 IR-MPH low-dose (≤ 15 mg/day) versus ER-MPH low dose (≤ 20 mg/day)


Administered twice dailyFitzpatrick, C (4×) MPH (10–15 mg/day, b.d.) vs 6.9–11.5 8 Core: Conners’ Hyperactivity 199255 sustained-release MPH (Slow Index (parents/teacher); TOTS:

Release Ritalin, SR-20) (20 mg/day, hyperactivity (parents and o.d.) – 19 teachers)

QoL: no CGI; comments ratings(parent/teacher)AE: STESS (parents); weight

C, crossover trial (number of crossovers); CGI, Clinical Global Impression; STESS, Subject’s Treatment Emergent SymptomScale; TOTs, Loney’s Time on Task Scale.

TABLE 48 Results for hyperactivity [IR-MPH low dose (≤ 15 mg/day) versus ER-MPH low dose (≤ 20 mg/day)]

Study Scale IR-MPH low dose: ER-MPH low dose: p-Value mean (SD) mean (SD) (if reported)a

6–11 yearsFitzpatrick, 199255 (? Mean, ? SD) (? Mean, ? SD)

Conners’ Hyperactivity Index 0.96 (0.50) 0.98 (0.72) NS(parents)Conners’ Hyperactivity Index (teacher) 0.73 (0.65) 0.77 (0.63) NS

Mean (SD) Mean (SD)TOTS (hyperactivity) (parents) 0.20 (0.31) 0.22 (0.50) NSTOTS (hyperactivity) (teachers) 0.16 (0.44) 0.12 (0.51) NS

a Note that owing to poor reporting for some aspects of the study methodology, p-Values should be interpreted withcaution. Lower scores represent a better behavioural outcome.NS, not significant; TOTS, Loney’s Time on Task Scale.

non-confidential study showed similarimprovement between treatment groups.Regarding adverse events, the non-confidentialtrial found no differences between the treatmentgroups with regard to loss of appetite or stomachache. However, this trial reported a higherincidence of headache with ER-MPH. The onestudy that examined hyperactivity rated poorly inthe quality assessment.


IR-MPH medium dose (15–30 mg/day) plus non-drug intervention versus ER-MPH low dose(≤ 20 mg/day) plus non-drug interventionTwo studies evaluated medium-dose(15–30 mg/day) IR-MPH plus non-drugintervention compared with low-dose(≤ 20 mg/day) ER-MPH plus non-drugintervention (Table 51; with additional informationin Appendix 12). These studies were bothconducted by Pelham and colleagues and involved behaviour modification during an


66

TABLE 49 IR-MPH high dose (>30 mg/day) versus ER-MPH medium dose (20–40 mg/day)


Administered twice dailyWolraich, 200197 P MPH (15, 30 or 45 mg/day, 6–12 4 Core: SNAP-IV: hyperactivity/

t.d.s.) vs OROS MPH (Concerta) impulsivity (parent, teacher)(18, 36 or 54 mg/day, o.d.) – 312 QoL: CGI improvement

(investigators)AE: solicited and spontaneousreports: focus on sleep quality,tics and appetite (parent)

Quinn, 200384 [Confidential information removed]

Steele, 200490 [Confidential information removed]

P, parallel trial.

TABLE 50 Results for hyperactivity [IR-MPH high dose (>30 mg/day) versus ER-MPH medium dose (20–40 mg/day)]

Study Scale IR-MPH high dose: ER-MPH medium Mean mean (SD) dose: mean (SD) difference

6–12 yearsWolraich, 200197 MD (95% CI)

SNAP-IV (hyperactivity/ 0.93 (0.79) 0.96 (0.79) –0.05 (–0.24 to 0.14)impulsivity) (teacher rated)SNAP-IV (hyperactivity/ 1.10 (0.69) 1.11 (0.65) 0.17 (0.03 to 0.31)impulsivity) (parent rated)

Lower scores represent a better behavioural outcome.

Study

Quinn 2003Steele 2004Wolraich 2001 6/107

MPH (high dose)n/N

RR (fixed)95% CI

RR (fixed)95% CI

0.40 (0.16 to 0.98)

10.50.20.1 52Favours ER-MPHFavours MPH-IR

10

15/106

ER-MPH (medium dose)n/N

[Confidential information removed][Confidential information removed]

FIGURE 20 Relative risks of headache: MPH high dose versus ER-MPH medium dose

STP.77,78 Neither examined hyperactivity or CGI as core outcomes. Behaviour was assessedusing the teacher-rated Abbreviated CTRS, and both studies reported no significantdifferences between the treatment groups (seeAppendix 12).

Adverse eventsBoth trials reported on the incidence of insomnia;neither detected significant differences betweentreatments. Pelham and colleagues77 reported the incidence of anorexia; however, no significant differences were found betweentreatments. No further data of interest werereported.

SummaryNo studies in this category evaluated hyperactivityor CGI as outcomes. Two studies did report some

adverse event data, and found no difference ininsomnia or anorexia.

IR-MPH high dose (>40 mg/day) plus non-drugintervention versus ER-MPH medium dose(20–40 mg/day) plus non-drug interventionOne crossover study evaluated high-dose(>40 mg/day) IR-MPH plus non-drug interventioncompared with medium-dose (20–40 mg/day) ER-MPH plus non-drug intervention (Table 52;with additional information in Appendix 12). The non-drug intervention involved a behaviouralprogramme incorporating parent training, teacherconsultation and a point system. This study byPelham and colleagues82 did not evaluatehyperactivity or CGI, but did assess a number ofother core outcomes. For some outcomes, ER-MPH was superior to IR-MPH (see Appendix 12),but this was not the case when inattention/


67


TABLE 51 IR-MPH medium dose (15–30 mg/day) plus non-drug intervention versus ER-MPH low dose (≤ 20 mg/day) plus non-drugintervention


Administered twice dailyPelham, 198777 C (3×) MPH (20 mg/day, b.d.) vs 6.5–11 7 Core: no hyp; Abbreviated CTRS

sustained-release MPH (Slow (teacher)Release Ritalin, SR-20) (20 mg/day, QoL: not reportedo.d.) plus behaviour modification AE: Side Effects Checklists (STP, 7 weeks) – 13 (parents, teachers, counsellors)

Pelham, 199078 C (5×) MPH (20 mg/day, b.d.) vs 8.1–13.2 8 Core: no hyp; Abbreviated CTRS sustained-release MPH (Slow (teachers/counsellors)Release Ritalin, SR-20) (20 mg/day, QoL: not reportedo.d.) plus behaviour modification AE: Side Effects Checklists (STP, 8 weeks) – 22 (parents, teachers, counsellors)

C, crossover trial (number of crossovers); STP, Summer Treatment Programme.

TABLE 52 MPH high dose (>40 mg/day) plus non-drug intervention versus ER-MPH medium dose (20–40 mg/day) plus non-drugintervention


Administered twice or more dailyPelham, 200182 C (3×) MPH (15, 30 or 45 mg/day, t.d.s.) 6–12 3 Core: no hyp; IOWA-C:

vs OROS MPH (Concerta) inattention/overactivity (teacher (18, 36 or 54 mg/day, o.d.) plus and parent)parent training, teacher QoL: no CGI; global effectiveness consultation and point systems (parent and teacher)(3 weeks) – 70 AE: questions regarding adverse

events, sleep quality, appetite,and tics (parents); spontaneousreports of adverse events

C, crossover trial (number of crossovers); CGI, Clinical Global Impression.

overactivity was measured using the IOWA-C,except when assessed by parents. No significantdifferences between these two treatment arms werereported for global effectiveness (parent or teacherrated).

Adverse eventsThere were no significant differences betweentreatments in the incidence of headache, stomachache, loss of appetite or insomnia. No data onweight were reported.

SummaryNo studies in this category evaluated hyperactivityor CGI as outcomes. One study that reportedadverse events data found no differences betweenthe treatment groups.

MPH versus DEXMPH medium dose (15–30 mg/day) versus DEX low dose (<10 mg/day)One study evaluated medium-dose(15–30 mg/kg/day) IR-MPH compared with low-dose (<10 mg/day) DEX (Table 53; with additionalinformation in Appendix 12). In this crossoverstudy by Efron and colleagues,50 hyperactivity wasmeasured using four different Conners’ scales.The authors reported significant differences infavour of MPH for the teacher-rated scales, butnot for the parent-rated scales (Table 54). Efronand colleagues did not examine CGI, but didreport on Parental Global Perceptions and GlobalRatings of Response (by the child and parent).They reported no significant differences betweenthe treatment groups for these outcomes.


68

TABLE 53 MPH medium dose (15–30 mg/day) versus DEX low dose (<10 mg/day)


Administered two or more times dailyEfron, 199750 C (2×) MPH (0.60 mg/kg/day, 5–14.9 4 Core: CPRS-R: impulsive–

b.d.) vs DEX (0.3 mg/kg/day, hyperactive factor; composite b.d.) – 125 hyperactivity index; CTRS-R:

hyperactivity factor; HyperactivityIndexQoL: no CGI; Parental GlobalPerceptions Questionnaire:overall perceptions; Child GlobalPerceptions QuestionnaireAE: SERS (parents)

C, crossover trial (number of crossovers); CGI, Clinical Global Impression; CPRS-R, Conners’ Parent Rating Scale – Revised;CTRS-R: Conners’ Teacher Rating Scale – Revised; SERS, Side Effects Rating Scale.

TABLE 54 Results for hyperactivity [MPH medium dose (15–30 mg/day) versus DEX low dose (<10 mg/day)]

Study Scale MPH medium dose: DEX low dose: p-Value (if mean (SD) mean (SD) reported)a

5–14 yearsEfron, 199750 CPRS-R (Composite Hyperactivity 64.28 (13.46) 64.89 (13.74) 0.51

Index)CPRS-R (impulsive/hyperactive) 57.39 (10.53) 57.33 (11.22) 0.87CTRS-R (Hyperactivity Index) 56.14 (10.17) 58.76 (10.57) <0.01CTRS-R (hyperactivity) Baseline: 71.26 (13.24) 56.20 (11.02) 58.88 (11.08) <0.01

a Note that owing to poor reporting for some aspects of the study methodology, p-values should be interpreted withcaution.

Lower scores represent a better behavioural outcome.CPRS-R, Conners’ Parent Rating Scale – Revised; CTRS-R, Conners’ Teacher Rating Scale – Revised.

Adverse eventsNo significant differences were detected betweenDEX and MPH treatments in the incidence ofheadache, stomach ache, loss of appetite orinsomnia. Data on weight were not reported.

SummaryOne study evaluated medium-dose MPHcompared with low-dose MPH.50 This studyreported significant effects in favour of MPH whenhyperactivity was assessed by teachers, but notwhen assessed by parents. They reported nosignificant differences between treatments in theincidence of headache, stomach ache, loss ofappetite or insomnia. Owing to the poor reportingof some methodological criteria, the results fromthis study should be interpreted with caution.

MPH medium dose (15–30 mg/day) versus DEXmedium dose (10–20 mg/day)One study evaluated medium-dose(15–30 mg/kg/day) immediate release IR-MPH

compared with medium-dose (10–20 mg/day) DEX (Table 55; with additional information inAppendix 12). In this crossover study by Arnoldand colleagues,37 hyperactivity was assessed usingfour scales. The authors reported no differencesbetween medium doses of MPH and DEX(Table 56). This study did not report on CGI. It generally scored well in the quality assessment.

Adverse eventsAdverse event data were not reported adequatelyto be included in the analysis.

SummaryThe one study that evaluated medium-dose MPHcompared with medium-dose DEX reported nodifference in hyperactivity scores between thetreatment groups.37 This study did not examineCGI or adequately report on adverse events data.It rated well in the quality assessment, hence the results for hyperactivity are likely to bereliable.


69


TABLE 55 MPH medium dose (15–30 mg/day) versus DEX medium dose (10–20 mg/day)


Administered two or more times dailyArnold, 197837 C (3×) MPH (1.25 mg/kg, 1 or 5–12 3 Core: Conners’ Teachers’

2×) vs DEX (0.63 mg/kg/day, Behaviour Problem Checklist: 1 or 2×) – 29 hyperactivity; Problem Behaviour

Checklist (parents): hyperactivity;Davids’ Hyperkinetic Rating Scale(parents and teachers)QoL: not reported AE: Problem Behaviour Checklist(parents): side-effects; weight loss

C, Crossover trial (number of crossovers).

TABLE 56 Results for hyperactivity [MPH medium dose (15–30 mg/day) versus DEX medium dose (10–20 mg/day)]

Study Scale MPH medium dose: DEX medium p-Value (if mean (SD) dose: mean (SD) reported)

5–12 yearsArnold, 197837 Problem Behaviour Checklist (parents) 18.21 (5.61) 17.21 (5.45) NR

(hyperactivity)Conners’ Teachers’ Behaviour Problem 16.83 (5.50) 16.17 (4.64) NSChecklist (hyperactivity)Davids’ Hyperkinetic Rating Scale 4.31 (1.23) 4.28 (1.07) NS(parents) (hyperactivity)Davids’ Hyperkinetic Rating Scale 3.83 (1.49) 3.90 (1.54) NS(teachers) (hyperactivity)

NR, not reported; NS, not significant.

MPH high dose (>30 mg/day) plus non-drugintervention versus DEX high dose (>20 mg/day)plus non-drug interventionOne study evaluated high-dose (>30 mg/day) IR-MPH compared with high-dose (>20 mg/day)DEX (Table 57; with additional information inAppendix 12). Although this study by Elia andcolleagues51 examined hyperactivity and CGI asoutcomes, the results were presented in graphform only and could not be reproduced in a table.The authors did report that both drugs werefound to be equally efficacious.

Adverse eventsNo significant differences were detected in theincidence of decreased appetite or insomniabetween treatment periods (RR = 1.12; 95% CI0.98 to 1.28 and RR = 1.03; 95% CI 0.84 to 1.25).No further data were reported.

SummaryOne study evaluated high-dose MPH plus non-drug intervention compared with high-dose DEXplus non-drug intervention,51 but no data could beextracted for hyperactivity or CGI. The trial didnot report any differences between groups forappetite or insomnia (the only adverse events that

were examined). This study did not score very wellin the quality assessment, and any results shouldbe interpreted with caution.

MPH medium dose (15–30 mg/day) plus non-drugintervention versus DEX-SR plus non-druginterventionOne study evaluated medium-dose (15–30 mg/day)IR-MPH compared with DEX-SR plus non-drugintervention (Table 58; with additional informationin Appendix 12). This study did not report anyhyperactivity or QoL outcomes. They did measurebehaviour using the Abbreviated CTRS (asassessed by teachers and counsellors). The scoresbetween treatment groups were similar (seeAppendix 12).

Adverse eventsNo significant differences were detected in theincidence of insomnia between DEX-SR and IR-MPH. Occurrences of headache, stomach acheand loss of appetite were not recorded by patientsor parents. No weight data were reported.

SummaryOne study was included in this category,78 but itdid not evaluate hyperactivity or CGI. This study


70

TABLE 57 MPH high dose (>30 mg/day) plus non-drug intervention versus DEX high dose (>20 mg/day) plus non-drug intervention


Administered two or more times dailyElia, 199151 C (3x) MPH (25–90 mg/day, b.d.) vs 6–12 9 Core: CTRS: hyperactivity;

DEX (mean 10–45 mg/day, b.d.) CPQ: hyperactivityplus behaviour modification + diet QoL: CGI (physician); C-GAS(11 weeks) – 48 AE: STESS (physician, parents)

C, crossover trial (number of crossovers); CGI, Clinical Global Impression; C-GAS, Children’s Global Assessment Scale;CPQ, Conners’ Parent Questionnaire; CTRS, Conners’ Teacher Rating Scale; STESS, Subject Treatment Emergent SymptomScale.

TABLE 58 MPH medium dose (15–30 mg/day) plus non-drug intervention versus DEX-SR plus non-drug intervention


Administered two or more times dailyPelham, 199078 C (5×) MPH (20 mg/day, b.d.) vs 8.1–13.2 8 Core: no hyp; Abbreviated CTRS

DEX-SR (10 mg/day, o.d.) plus (teachers/counsellors)behaviour modification (STP, QoL: not reported8 weeks) – 22 AE: Side Effects Checklists

(parents/teachers/counsellors)


did report that there were no significantdifferences between treatments in the incidence ofinsomnia (other adverse events of interest werenot examined). This study did not score very wellin the quality assessment, and any results shouldbe interpreted with caution.

ER-MPH low dose (≤ 20 mg/day) plus non-drugintervention versus DEX-SR plus non-druginterventionOne study evaluated low-dose (≤ 20 mg/day) ER-MPH compared with DEX-SR plus non-drugintervention (Table 59; with additional informationin Appendix 12). As presented above, this studydid not report any hyperactivity or QoL outcomes.The scores between treatment groups were similarwhen assessed using the Abbreviated CTRS (seeAppendix 12).

Adverse eventsNo significant differences were detected in theincidence of insomnia between DEX-SR and ER-MPH. Occurrences of headache, stomach acheand loss of appetite were not recorded by patientsor parents. No weight data were reported.

SummaryAs above, one study was included in thiscategory,78 but it did not evaluate hyperactivity orCGI. This study reported that there were nosignificant differences between treatments in theincidence of insomnia (other adverse events ofinterest were not examined). This study did notscore very well in the quality assessment, and anyresults should be interpreted with caution.

MPH versus ATXMPH high dose (>30 mg/day) versus ATX highdose (≥1.5 mg/kg/day)One study evaluated high-dose (>30 mg/day) IR-MPH compared with high-dose (≥1.5 mg/kg/day)ATX (Table 60; with additional information inAppendix 12).

In this parallel study by Kratochvil and colleagues,70

hyperactivity was measured using two scales. Nodifferences were reported between the treatmentgroups using either scale (Table 61). This study alsoreported on CGI – Severity. Again, no difference wasreported in children in the MPH group comparedwith children in the DEX group (p = 0.663).


71


TABLE 59 ER-MPH low dose (≤ 20 mg/day) plus non-drug intervention versus DEX-SR plus non-drug intervention


Administered once dailyPelham, 199078 C (5×) Sustained-release MPH (Slow 8.1–13.2 8 Core: no hyp; Abbreviated CTRS

Release Ritalin, SR-20) (20 mg/day, (teachers/counsellors)o.d.) vs DEX-SR (10 mg/day, o.d.) QoL: not reportedplus behaviour modification (STP, AE: Side Effects Checklists 8 weeks) – 22 (parents/teachers/counsellors)


TABLE 60 MPH high dose (>30 mg/day) versus ATX high dose (≥ 1.5 mg/kg/day)


Administered twice dailyKratochvil, P MPH (31.3 mg/day 1–3×) vs ATX 7–15 10 Core: ADHD-RS-IV–Parent 200270 (max. 2.0 mg/kg/day, b.d.) – 228 Version (investigator administered

and scored):hyperactivity/impulsivity; CPRS-R: hyperactivityQoL: CGIAE: open-ended questions;weight

CPRS-R, Conners’ Parent Rating Scale – Revised; CGI, Clinical Global Impression; P, parallel trial.

Adverse eventsNo significant differences in the incidence ofheadache, loss of appetite, stomach ache orinsomnia were detected between the ATX andMPH treatment groups. In addition, nodifferences in participants’ mean change in weightwere observed.

SummaryOne study was included in this category.70 Nodifferences were reported between the two drugsfor hyperactivity, CGI or adverse events (ofinterest). This study did not score very well in thequality assessment, and any results should beinterpreted with caution.

ER-MPH medium dose (20–40 mg/day) versusATX low/medium dose (<1.5 mg/kg/day)One study evaluated medium-dose (20–40 mg/day)ER-MPH compared with low/medium-dose

(<1.5 mg/kg/day) ATX (Table 62; with additionalinformation in Appendix 12). This parallel studyby Kemner and colleagues64 reported a significantimprovement in favour of MPH when measuredusing the ADHD Rating Scale for hyperactivity(Table 63). Similarly, the authors reported that CGI– Improvement responder rates were significantlybetter for the MPH group than for the ATX group(68.6 versus 52.8%, p < 0.001).

Adverse eventsIncidence of headache and stomach ache did notdiffer significantly between the ATX and ER-MPHtreatment groups. Participants assigned to ER-MPH suffered from a significantly higherincidence of reduced appetite and insomniacompared with those assigned to ATX (RR = 1.95;95% CI 1.09 to 3.49 and RR = 2.88; 95% CI 1.57to 5.28, respectively). No data on weight werereported in this study.


72

TABLE 61 Results for hyperactivity [MPH high dose (>30 mg/day) versus ATX high dose (≥ 1.5 mg/kg/day)]

Study Scale MPH high dose: ATX high dose: Mean mean (SD) mean (SD) difference

7–15 yearsKratochvil, 200270 ADHD-RS-IV Baseline/change from Baseline/change from MD (95% CI)

(hyperactivity/impulsivity) baseline baselineCPRS-R (hyperactivity) 16.95 (7.07)/–8.48 (7.08) 17.77 (6.31)/–9.50 (6.99) 1.02 (–1.40 to 3.44)

10.05 (5.35)/–4.78 (4.49) 10.25 (4.39)/–5.56 (4.74) 0.78 (–0.82 to 2.38)

Lower scores represent a better behavioural outcome.ADHD-RS-IV, ADHD Rating Scale–Parent Version (investigator administered and scored); CPRS-R, Conners’ Parent RatingScale – Revised.

TABLE 62 ER-MPH medium dose (20–40 mg/day) versus ATX low/medium dose (<1.5 mg/kg/day)


Administered once dailyKemner, 200464 P OROS MPH (Concerta) 6–12 3 Core: ADHD-RS: hyperactivity

[Confidential information (investigator rated)removed] vs ATX [Confidential QoL: CGI Improvementinformation removed] – 1323 AE: as reported by

participants/parents to clinicians

ADHD-RS, ADHD Rating Scale; CGI, Clinical Global Impression; P, parallel trial.

TABLE 63 Results for hyperactivity [ER-MPH medium dose (20–40 mg/day) versus ATX low/medium dose (<1.5 mg/kg/day)]

Study Scale ER-MPH medium dose: ATX low/medium dose: p-Value mean (SD) mean (SD) (if reported)

6–12 yearsKemner, 200464 [Commercial information removed]

SummaryThe one study in this category reported asignificant improvement in favour of MPHcompared with ATX for both hyperactivity andCGI. However, this study also reported a higherincidence of reduced appetite and insomnia in theMPH group, but no differences in headache orstomach ache.

[Commercial information removed].

ATX versus DEXNo studies directly compared ATX and DEX.

The MTA trial(This section of the report was copied from theoriginal NICE review: Lord J, Paisley S. The clinicaleffectiveness and cost-effectiveness of methylphenidate forhyperactivity in childhood. London: National Institutefor Clinical Excellence, Version 2; August 2000.Results of recently published papers were added.)

The Multimodal Treatment Study of Children withADHD (MTA) trial does not strictly fall within theremit of this review, since ‘medical management’included the option to use various drugs, not justmethylphenidate. However, given the importanceof this study and its relevance to practice, its keyresults are summarised below.

Children between the ages of 7 and 9 years with adiagnosis of ADHD combined type (DSM-IV) wererecruited through six centres. They had a range ofco-morbid conditions, although children withconditions thought likely to prevent fullparticipation in the treatments or assessmentswere excluded. Participants were randomised(n = 579) to one of four groups:

1. Medication management. Children had aninitial 28-day double-blind, placebo-controlleddose titration of MPH (‘n of 1’ trial). This wasfollowed by open titration of other medicationsfor children with inadequate response to MPH.Children were maintained on optimalmedication (including ‘no medication’ whereappropriate) for 13 months, with half-hourmonthly medication maintenance visits to apharmaco-therapist, who offered ‘supportencouragement, and practical advice (but notbehavioural treatment)’. Further algorithm-guided dose adjustments were allowed.

2. Behavioural treatment. This included threemain components: first, a parent-trainingprogramme with 27 group and eight individualsessions per family; second, a child-focusedtreatment programme, which comprised an

8-week, 5 days per week, 9 hours per daysummer camp; third, a school-basedprogramme, which included 10–16 sessions ofbiweekly teacher consultation and 12 weeks of apart-time, behaviourally trained classroomaide. Daily report cards were completed byteachers, to link school and home. Thesebehavioural interventions were tapered, withintensive initial inputs fading to once-monthlycontacts by the end of the 14-month treatmentperiod.

3. Combined treatment. This included both ofthe above treatment programmes, but was notthe simple addition of the other two strategies.To coordinate treatment, information wasshared between the teacher, consultant andpharmaco-therapist. Average medication dosesreceived also varied between the medicationmanagement and combined treatment groups.

4. Community care. Here children were providedwith a report of their initial study assessmentsand a list of community mental healthresources, then discharged to their ownprovider. In accordance with US practice, mostof the children in this group receivedpharmaceutical therapy. The level ofpsychotherapeutic interventions in this grouphas not yet been reported.

Outcomes were measured across six major domains:ADHD symptoms, oppositional/aggressivesymptoms, social skills, internalising symptoms(anxiety and depression), parent–childrelationships and academic achievement. Openparent and teacher ratings for these dimensionswere augmented with blinded observationalratings of classroom behaviour. Assessments wereconducted at baseline and 3, 9 and 14 months.Further follow-up assessments are planned.Analyses were conducted on an ITT basis usingrandom-effects regression methods.

The study was designed to assess the relativeeffectiveness of alternative treatment strategies.These strategies met ‘good practice’ ideals,although the children were intended to berepresentative of ‘real-world’ patients.103 It isimportant to note that this trial did not include aplacebo or ‘no treatment’ control group. Hence,the MTA trial cannot be used to assess the efficacyof the single treatment modalities (medication orbehavioural therapy alone). Also, the communitycare group is of little direct relevance in the UK,because of the large differences between currentpractice in the USA and UK.104 Most of thechildren in the community care group (97/146)received stimulant medication.


73


Results of the MTA trialThe MTA trial was designed to answer threequestions:76

1. Do medication and behavioural treatments result incomparable levels of improvement in pertinentoutcomes at the end of treatment?Medication management was superior tobehavioural treatment for three (of five)measures of ADHD core symptoms. Nosignificant differences were observed across theother key dimensions.

2. Do participants assigned to combined treatment showhigher levels of improvement in overall functioningin pertinent outcome domains than those assigned toeither medication management or behaviouraltreatment at the end of treatment (one-tailedhypotheses)?(a) Combined treatment and medication

management do not differ significantlyacross any domain.

(b) Combined treatment was superior tobehavioural management on three (of five)measures of ADHD core symptoms, for one(of three) measures ofaggression/oppositional behaviour, for one(of three) measure of anxiety depressionand for one (of three) measure of academicachievement. No significant differenceswere observed in social skills orparent–child relations.

3. Do participants assigned to each of the three MTAtreatments (medication management, behaviouraltreatment and combined treatment) show greaterimprovement over 14 months than those assigned tocommunity care (one-tailed)?(a) Medication management was superior to

community care for three (of five) measuresof ADHD symptoms, for two (of three)measures of aggression/oppositionalbehaviour and for one (of two) measures ofsocial skills. No significant differences wereobserved in anxiety/depression orparent–child relations.

(b) No significant differences betweenbehavioural management and communitycare were observed for any outcomedomains.

(c) Combined treatment was superior tocommunity care for four (of five) measuresof ADHD symptoms, for two (of three)measures of aggression/oppositionalbehaviour, for one (of three) measures ofanxiety/depression, for both measures ofsocial skills, for one (of two) measures ofparent–child relations and for one (of three)measures of academic achievement.

The MTA Cooperative Group conducted furtheranalysis to identify patient subgroups with betteror worse response to the various treatmentstrategies.105 This analysis should be seen asexploratory, because of the danger of repeatedstatistical testing with a sample not designed forthis purpose. There was no difference in treatmentresponse by sex, prior treatment or presence of co-morbid disruptive disorders. Behaviouraltreatment appeared to be more effective inchildren with anxiety disorders and children fromdeprived backgrounds.

Since the publication of the previous NICEreport,4 the authors of the MTA trial havepublished a number of related papers.106–113 Mostof these are cross-sectional analyses, overviewstudies or studies focusing on covariates, such associo-economic status or ethnicity, which will notbe discussed in this report.106,107,110,111

Other papers focused on subgroupanalyses.108,109,112 As the samples were not designedfor this purpose, results should be interpreted withcaution. Vitiello and colleagues examined thetrajectory of MPH dosage over time, following acontrolled titration.108 The aim was to ascertainhow accurately the titration was able to predicteffective long-term treatment in children withADHD. They concluded that for most children,initial titration found a dose of MPH in the generalrange of the effective maintenance dose, but didnot prevent the need for subsequent maintenanceadjustments.108 Greenhill and colleagues109

examined whether the trial identified the best MPHdose for each child with ADHD. They found thatthe MTA titration protocol validated the efficacy ofweekend MPH dosing and established a total dailydose limit of 35 mg of MPH for children weighing<25 kg. It replicated previously reported MPHresponse rates (77%), distribution of best doses(10–50 mg/day) across children, effect sizes onimpairment and deportation and dose-relatedadverse events.109 Galanter and colleaguesexamined the response to MPH in children withADHD and some manic symptoms.112 Theirfindings suggested that these children respondrobustly to MPH during the first month oftreatment and that they are not more likely to havean adverse response to MPH.112

In 2004, the authors of the MTA trial publishedresults of a follow-up 10 months beyond the14 months of intensive intervention.113 Of the 579children who entered the study, 540 (93%)participated in this first follow-up. Resultsindicated that the MTA medication strategy


74

showed persisting significant superiority overbehavioural treatment and community care forADHD and oppositional–defiant symptoms at24 months, although not as great as at 14 months.Significant additional benefits of combinedtreatment over medication management and ofbehavioural treatment over community care werenot found. The groups differed significantly inmean dose (MPH equivalents: 30.4, 37.5, 25.7 and24.0 mg/day, respectively).113

Summary of clinical effectivenessdataMPH versus placeboStudies that evaluated low-dose MPH comparedwith placebo demonstrated variable results forhyperactivity.43,54,55,60,61,96 No differences in CGIwere reported between the groups96 (no otherstudies measured this outcome). With medium-dose MPH, the majority of studies demonstratedthat MPH was superior to placebo forhyperactivity;42,43,46,54,60,61,68,98 (no results weresignificant in the study by Stein and colleagues91),and one study reported that MPH improved CGIcompared with placebo83 (no other studiesmeasured this outcome). The two studies thatevaluated hyperactivity with high-dose MPHdemonstrated variable results,43,77 although therewas evidence that high-dose MPH improvedCGI.97 [Confidential information from one studythat evaluated the effectiveness of high-doseMPH on CGI removed].

There was a paucity of studies that examinedMPH (any dose) plus a non-drug intervention (e.g. cognitive therapy) compared with placebo.Only one study reported data for hyperactivity,and the results were not significant.42 In addition,very few studies examined the effectiveness of ER-MPH compared with placebo. Of these studies,the majority reported that ER-MPH (low, mediumor high) was superior to placebo for all outcomesof interest (hyperactivity55;97 and CGI59,93,97).[Confidential information from one study thatreported on CGI removed].

Again, very few studies compared MPH with anon-drug intervention (e.g. parent training,cognitive training and behavioural therapy). Three studies evaluated hyperactivity, two of whichdemonstrated that children receiving MPH weresignificantly improved compared with childrenreceiving a non-drug intervention42,53 and onewhich demonstrated variable results depending onthe scale used.65

A larger number of studies examined MPH incombination with a non-drug intervention (e.g.parent training, one-to-one reading therapy,cognitive therapy and parent and teachereducation) compared with a non-drug intervention.Of these, five presented reproducible results onhyperactivity and two reported results on CGS.Generally, combined treatment was superior tonon-drug treatment for hyperactivity42,53,71 andClinical Global Impression.51,65 In two studies,hyperactivity was improved when assessed byteachers, but not by parents65,86 (in these cases thenon-drug interventions were parent training orsupport and parent and teacher education).

No studies that compared ER-MPH plus non-drugintervention with a non-drug interventionevaluated hyperactivity or CGI.

Generally, the studies that evaluated MPH did notadequately report on study methodology, and theresults should be interpreted with caution.

DEX versus placeboOnly two studies that compared DEX and placeboreported reproducible data on hyperactivity.36,45

When assessing the efficacy of medium-dose DEX,the results for hyperactivity varied depending on thescale used.45 For higher dose DEX, hyperactivityand CGI appeared to be improved with drugtreatment36 (statistical results for hyperactivitywere not reported). Generally, this study rated wellin the quality assessement. An additional study byArnold and colleagues38 evaluated DEX-TRcompared with placebo. They observed significantimprovements in hyperactivity with treatment.However, owing to the poor reporting of somemethodological criteria, the results from this studyshould be interpreted with caution.

No studies comparing DEX plus non-drugtreatment versus placebo or DEX versus non-drugintervention, measured hyperactivity or CGI asoutcome measures. One study compared DEX incombination with a non-drug treatment (academicinstruction and therapeutic recreation) with non-drug treatment alone with hyperactivity as anoutcome.62 Results were significant in favour of thecombined treatment group when assessed usingthe CTRS and the Children’s Psychiatric RatingScale, but not when assessed using the CPRS. CGIwas not examined. In addition to evaluatingimmediate release DEX, James and colleagues62

also examined DEX-SR. In this study, combinationtreatment resulted in improved behaviourcompared with non-drug treatment alone.However, this study did not score very well in the


75


quality assessment, and the results should beinterpreted with caution.

ATX versus placeboGenerally, there was consistent evidence that ATXwas superior to placebo for hyperactivity andCGI.63,73–75,89 No studies compared ATX with non-drug treatment. Two of these studies were deemedto be of good quality (based on the reporting ofthe study methodology) and their results are likelyto be reliable.73,89 [Confidential information fromone study that examined hyperactivity and CGIremoved].

IR-MPH versus ER-MPHOf the studies that evaluated IR-MPH comparedwith ER-MPH, no differences were reported forhyperactivity55,97 or CGI.97 However, these twostudies scored poorly in the quality assessment,and the results should be interpreted with caution.[Confidential information from two studies thatreported on CGI removed].

MPH versus DEXThe results of comparisons between MPH andDEX were variable. One study reported that MPH(medium dose) was superior to DEX (low dose)when assessed using Conners’ teacher-rated scales,but not for Conners’ parent-rated scales.50

Another study observed no differences betweenMPH (medium dose) and DEX (medium dose) forhyperactivity.37 CGI was not assessed in eitherstudy. The study that evaluated medium-doseMPH compared with medium-dose DEX ratedwell in the quality assessment, whereas the otherdid not score very well, and any results should beinterpreted with caution.

MPH versus ATXVery few studies compared MPH and ATX. Onereported no difference between the two drugs forhyperactivity or CGI.70 This study did notadequately report study methodology, and theresults should be interpreted with caution.[Confidential information from one study thatevaluated ER-MPH with ATX on hyperactivityand CGI removed].

ATX versus DEXNo studies compared ATX versus DEX.

Summary of adverse events dataMPH versus placeboPrimary studiesAmongst studies that compared low-dose MPH

with placebo, no differences in adverse events weredetected between the treatment groups.55,60

However, amongst studies that compared medium-or high-dose MPH with placebo, there wasevidence that treatment was associated with higherincidences of headaches, loss of appetite, stomachache and insomnia compared withplacebo.35,39,44,46,57,60,66–68,83,91,95 [Confidentialinformation from one study that examinedadverse events removed].

None of the studies examining MPH (any dose)combined with a non-drug intervention comparedwith placebo presented adverse events datainformative to our analysis.

Of the studies comparing ER-MPH (any dose) withplacebo, treatment was associated with decreasedappetite and increased insomnia amongstparticipants.32,55,59,92,97 [Confidential informationfrom one study that examined adverse eventsremoved].

None of the studies comparing MPH (any dose)alone with a non-drug intervention reportedadequate data on adverse events.

Amongst the trials comparing MPH incombination with a non-drug intervention with anon-drug intervention alone, the majority did notreport differences in adverse events betweentreatment groups.51,77,78,80–82,88

Of the studies comparing ER-MPH plus non-drugintervention with a non-drug intervention alone,the medium-dose study reported that loss ofappetite was significantly greater in the combinedtreatment group.82

Systematic reviewsOnly one SR met the inclusion criteria for ouranalysis of adverse events;114 this reported onstudies assessing MPH treatment. The majority ofthe included studies that examined weightdetected differences between treatment arms orbaseline and active drug conditions. Fewer studiesfound effects on height, and some initialdifferences were no longer significant at long-termfollow-up. Approximately half of the studiesevaluating heart rate or blood pressure detecteddifferences between treatment arms or baselineand active drug conditions. Loss of appetite, sleepdisturbances, dizziness, headaches and stomachaches were the most commonly reported somaticcomplaints. The authors concluded that adverseevents related to growth and cardiovascularoutcomes were predominantly transient in nature,


76

easily rectified with dosage adjustments. They alsohighlighted the need for further research intosomatic complaints, which may be associated withthe disorder itself rather than the initiation oftreatment. Appendix 13 presents the results of thisreview in more detail.

Tertiary sourcesThe BNF29 details the following as possiblephysical side-effects resulting from theadministration of MPH: tremor, rash, pruritus,urticaria, fever, arthralgia, alopecia, exfoliativedermatitis, erythema multiforme,thrombocytopenic purpura, thrombocytopenia,leucopoenia, dizziness, headache, gastrointestinalsymptoms, anorexia, urinary disorders, dry mouth,sweating, convulsions, tachycardia, anginal pain,palpitations, increased blood pressure and visualdisturbances. Rare cases of liver damage, musclecramps and cerebral arteritis have been noted. Inaddition, cardiomyopathy has been reported withchronic use. Choreoathetoid movements, tics andTourette syndrome have also been reported inpredisposed individuals. Growth retardation inchildren has further been noted. Behaviouraladverse effects listed include sleep disturbances,insomnia, depression, confusion, restlessness,irritability and excitability, nervousness, nightterrors and euphoria. Some individuals havepresented with drug dependence and tolerance orpsychosis.

DEX versus placeboPrimary studiesTwo studies comparing DEX with placeboadequately reported adverse events data of interestto our review. One study examining medium-doseDEX reported no difference in adverse eventsbetween treatment groups;56 the other studyexamining high-dose DEX reported a greaterincidence of loss of appetite, but detected nofurther differences.58

Adverse events data were not adequately reportedin the one study comparing a time-release form ofDEX with placebo.

No studies comparing combined DEX andbehavioural treatment with placebo reportedadequate adverse events data.

No adequate data were reported for studiescomparing DEX alone with a non-drugintervention.

One study examining high-dose DEX combinedwith a non-drug intervention compared with a

non-drug intervention alone reported adverseevents data: significantly decreased appetite andincreased insomnia were observed in thecombined treatment group.51

One study comparing an extended-release form ofDEX combined with a non-drug intervention witha non-drug intervention alone reported nodifferences in the incidence of insomnia betweentreatment phases.78 No further adverse events dataof interest were reported.

Systematic reviewsNo SR examining the adverse event profile ofDEX was identified.

Tertiary sourcesThe BNF29 details the following as possiblephysical side-effects resulting from theadministration of DEX: tremor, dizziness,headache, gastrointestinal symptoms, dry mouth,anorexia, sweating, convulsions, tachycardia,anginal pain, palpitations, increased bloodpressure and visual disturbances. In addition,cardiomyopathy has been reported with chronicuse. Choreoathetoid movements, tics and Tourettesyndrome have also been reported in predisposedindividuals. Growth retardation in children hasfurther been noted. Behavioural adverse effectslisted include insomnia, restlessness, irritabilityand excitability, nervousness, night terrors andeuphoria. Some individuals have presented withdrug dependence and tolerance or psychosis.

ATX versus placeboPrimary studiesStudies in this category presented evidence thatATX (all doses) results in significantly reducedappetite compared with placebo, but does notimpact the incidence of headache, stomach acheor insomnia.63,73–75,89,94 In the one trial thatreported weight data, ATX appeared to have anadverse effect on children’s weight gain.73

Systematic reviewsNo SR examining the adverse event profile ofatomoxetine hydrochloride was identified.

Tertiary sourcesThe BNF29 cites the following as physical side-effects associated with the administration of ATX:anorexia, dry mouth, nausea, vomiting, abdominalpain, constipation, dyspepsia, flatulence,palpitations, tachycardia, increased bloodpressure, postural hypotension, hot flushes,dizziness, headache, fatigue, lethargy, tremor,rigors, urinary retention, mydriasis, conjunctivitis,


77



78

dermatitis, pruritus, rash and sweating. Prostatitisin men and menstrual disturbances in womenhave also been observed. Some patients havesuffered from cold extremities, although lesscommonly. Behavioural side-effects include sleepdisturbances, depression, anxiety and sexualdysfunction.

IR-MPH versus ER-MPHPrimary studiesOf the studies that evaluated IR-MPH comparedwith ER-MPH, one study reported a significantlyhigher incidence of headache in the extended-release treatment group.97

MPH versus DEXPrimary studiesAmongst studies comparing IR-MPH with IR-DEX(of any respective dosage, with or without non-drug intervention), no significant differences weredetected between treatment groups with regard toreported adverse events.50,51

No significant differences were detected in theincidence of insomnia in the one study comparingER-DEX combined with a non-drug interventionwith IR-MPH (medium dose) combined with anon-drug intervention. No further data werereported.78

Similarly, no significant differences were found inthe incidence of insomnia between treatmentphases in the one study comparing ER-MPHcombined with a non-drug intervention with ER-DEX combined with a non-drug intervention. Nofurther data were reported.78

MPH versus ATXPrimary studiesThe study comparing MPH (high dose) with ATX(high dose) did not detect significant differencesin adverse events between treatment groups.70

However, the study comparing an extended-release form of MPH (medium dose) with ATX(low/medium dose) found that participantsassigned to ER-MPH suffered from reducedappetite and insomnia to a significantly greaterextent. No differences were detected in theincidence of headache or stomach ache betweentreatment groups.99

ATX versus DEXPrimary studiesNo studies compared ATX versus DEX.


79


AimThe aim of this chapter is to review the publishedliterature and company submissions to the NICEon the QoL and cost-effectiveness of

� oral MPH including Ritalin [Immediate Release(IR-MPH), Equasym (IR-MPH), Equasym® XL(Extended Release ER-MPH8 hour) andConcerta® XL (ER-MPH12 hour)]

� DES (Dexedrine)� ATX (Strattera)

in children and adolescents (under 18 years ofage) diagnosed with ADHD (including HKD). Allstudies in which MPH, DEX or ATX were usedalone or in combination with other drugs or non-drug interventions [e.g. psychological/behaviouraltherapy (BT)] are included.

Literature reviewThe information scientist conducted a search onthe published QoL and cost-effectiveness literature(see Appendix 2 for full details of the searchstrategies and databases used). The QoL searchesidentified 500 records (853 records before de-duplication, that is, before eliminating duplicates)that were assessed by the health economist forrelevance. In addition, the 38 records (47 beforede-duplication) identified from the database-specific searches for economic evaluations and the6535 records (2450 before de-duplication)identified by the generic and adverse eventsearches were also sifted for relevant papers inrelation to the cost-effectiveness of named drugtherapies for ADHD. Five relevant cost-effectiveness and two relevant HRQoL paperswere identified from the searches.

QoL studies were ordered if they contained healthoutcome data for use in economic evaluations ofADHD. The review of the economic evaluationliterature included studies that compared two ormore ADHD interventions in terms of their costsand outcomes and where at least one drugintervention was assessed. Economic evaluations

included cost–consequence, cost–utility, cost-effectiveness, cost-minimisation and cost–benefitanalyses. Economic evaluations reported asconference proceedings or abstracts were excludedsince the data they contain may not be complete.Thirty-one QoL studies and five economicevaluation studies were ordered, of which two QoLstudies and five economic evaluations werereviewed. A data extraction form used in previousTechnology Assessment Reviews was used toabstract data on all economic evaluations reviewed(Appendix 7). Each economic evaluation wasquality assessed independently by two healtheconomists (Appendix 6) based on an economicevaluation checklist.115

Review of quality of life and cost-effectiveness literatureQuality of lifeSearches (see Appendix 2) for studies on the QoLof children and adolescents with ADHD revealed aplethora of instruments, the majority measuringmultiple, disaggregated dimensions of healthdesigned to identify the extent to which healthstatus is affected by the intervention in question.As such, many featured as outcome measures inthe trials that are reviewed in the clinicaleffectiveness chapter (Chapter 4).

Two main types of instruments were used toevaluate ADHD interventions including thosemeasuring (1) symptoms and functionality [e.g.the Peabody Individual Achievement Test(PIAT)]116 and those measuring (2) (HRQoL) inchildren with ADHD. The HRQoL instrumentsused included disease-specific instruments such asthe (SNAP-IV) Rating Scale,117 CTRS andCPRS.118 A few instruments were generic measuresthat can be used to compare effectiveness acrossdiseases, for example, the Index of Health RelatedQuality of Life (IHRQoL)119 or the genericpaediatric measures, for example the Child HealthQuestionnaire (CHQ).120

The choice of outcome measure is a critical designissue. ADHD is a complex neuro-developmental

Chapter 5

Review of economic evaluations of ADHD drug interventions in children and adolescents

constellation of problems rather than a singledisorder, with core symptoms of inattention,hyperactivity and impulsivity and other moregeneral symptoms such as poor schoolperformance and poor social functioning.121 Sincethese behavioural traits can be present inunaffected children, symptoms and functionality-based outcomes need to have discriminant validity,that is, they should be able to discriminatebetween children with and without ADHD. It hasbeen suggested that there might be uniquepatterns of effects created by different treatments(e.g. drug interventions versus BT) and a singleoutcome measure might not be sufficientlysensitive.118 Although measures that are used forreporting outcomes in a disaggregated way may beuseful from a clinical perspective, the use of theseinstruments has limitations from a decision-making perspective. Many of the measures featuresubscales, and the scoring of the instrument maynot be designed to provide an overall summaryscore. Unless the relative importance of eachsubscale or of each different profile measure canbe valued, it is not possible to use these measuresto calculate the net impact on HRQoL.

When assessing the cost-effectiveness of anintervention, the use of such disparate measuresmay lead to conflicting results depending on theinstrument or subscale used. One way to overcomethis problem is to use a preference-based index ofHRQoL. These provide a summary score, typicallybetween 0 (death) and 1 (full health), with therelative importance of changes in differentdimensions of health being weighted according topeople’s preferences. Given the perspective ofNICE, the most relevant values are those of thegeneral population of England and Wales. Thefocus of this review is children and adolescents so,in principle, their preferences may be mostrelevant. When preference values are obtainedusing the standard gamble (SG) or time-trade-off(TTO) techniques, they can be used to representutilities. (The NICE technical guidance requiresthat health states should be measured in patientsusing a generic and validated classification systemfor which reliable UK population preferencevalues, elicited using such a choice-based method,for example.122) Utility values for health states canbe combined with the length of time spent inthose health states to calculate quality-adjustedsurvival.

The quality-adjusted life-year (QALY) is commonlyused to measure health outcomes in healtheconomic evaluations, combining QoL with qualityof life in a single measure. Two studies contained

utility information for use in the construction ofQALYs123 (Gilmore and Milne123 is based on theWessex Institute DEC Report number 78124),125

(authors include Eli Lilly employees, producers ofStrattera/ATX).

Gilmore and Milne123 generated QALYs based ondata from the IHRQoL.119 The populationpreference-based IHRQoL system was used, whichincorporates three dimensions: disability,(physical) discomfort and (emotional) distress.Since the authors could not find accurateestimates of disability directly in the literature,they used their own judgement, in considerationof trial evidence. It was assumed that the QoLimprovements were 0.086 per individual for ayear. [The IHQoL health state was assumed tochange from that of no pain, slight social disability(some role functions slightly impaired by socialdisability) and moderate emotional distress(anxious and depressed most of the time buthappy and relaxed some of the time) to no pain,no physical or social disability and slightemotional distress (happy and relaxed more of thetime, but anxious and depressed some of thetime). Based on IHQoL data, this generates ascore of 0.970 – 0.884 = 0.086]. Estimations weremade for 1 year only, reflecting the better qualityof shorter-term trial data.

Initially, to calculate QALYs it was assumed, basedon the literature, that benefits observed at4–6 months persisted for the year of follow-upprovided that medication continued. In addition,it was assumed that 6% of individuals discontinuedtreatment over the year owing to side-effects andthat the average response rate in those whoremained within the trial was 70%. From this, itwas estimated that 100 children gained 94.06QALYs per year using MPH compared with 88.4QALYs per year for the placebo arm, anincremental difference of 5.66 QALYs or 0.0566QALYs per child.

There are a number of caveats surrounding theusefulness of these findings. The authorsacknowledge that the main limitation of theirwork lies in the generation of the QALY using theIHRQoL. Values for health states for children withADHD were obtained using expert judgement withconsideration of published trials. The process ofsynthesising this information from the literaturewas not explained. The authors state that theIHRQoL is not a sensitive tool for measuring thetypes of disabilities encountered in ADHD. Theymention that typically children with ADHD havemoderate to severe social disability, whereas the


80

only level of social disability that IHRQoLincludes is slight social disability. This is an oft-cited criticism of generic indexes of HRQoL: thattheir content validity may be weak when appliedin specific disease areas. In practice, the IHRQoLis rarely used today and values of health stateswere gained from a small sample of individualswhose preferences are unlikely to be representativeof the population of England and Wales as awhole. The authors undertook multi-waysensitivity analyses in order to explore the effectthat modelling plausible variations in quality oflife improvements pre- and post-treatment had oncost–utility estimates. Overall, the effect ofmodelling different disability/distress levels hadlittle impact on the estimates. The authors did notreport on any uncertainty associated with theQALY estimates.

An alternative approach to employing a genericpreference-based health index of HRQoL is todescribe health states specific to the disease underconsideration, and value them directly usingtechniques such as SG or TTO. The validity ofthese measures depends on the content and styleof the vignette used to describe each health state.Matza and colleagues125 published an abstractincluding utility information based on the SGtechnique. Utility information was elicited from 43parents of children with ADHD. Elevenhypothetical health states were developed basedon the opinion of doctors and informed bypublished literature and unpublished clinical trialdata. Health states that were valued includeduntreated ADHD, stimulant treatment and non-stimulant treatment (e.g. ATX). The actualvignettes used to describe each health state werenot available to us as the study is currentlypublished only as an abstract. The parents ratedeach health state including the current health stateof their child (mean parent SG rating = 0.74), andthe SG utility scores varied from severe untreatedADHD (0.48) to effective, tolerable ATX non-stimulant treatment (0.88). However, the fullrange of utilities for all 11 health states was notreported in the abstract and neither were any dataon the variation around each estimate. Theauthors stated that comparisons between healthstates found expected differences betweenuntreated mild, moderate and severe ADHDstates. In the case where stimulant and non-stimulant medication were both effective andtolerable, parents preferred the latter (ATX)(p < 0.03).

A potential advantage of this approach, therefore,is that small differences in health states can be

estimated in terms of utility values. Also, withreference to this review, the values are based ondirect patient valuations, rather than expertjudgement. We note that NICE prefers a genericand validated classification system for theestimation of health state utilities. However, theestimated utilities available for ADHD usinggeneric instruments are crude, and based onexpert opinion. Hence values obtained directlyfrom patients, using SG methodology, may bemore relevant for this review.

Cost-effectivenessFive relevant economic evaluations were found inthe published literature,4,123,126,128,129 includingLord and Paisley,4 which was part of the originalNICE appraisal. (Reference 126 is reported in theeconomic evaluation section of Miller andcolleagues30 and is also reported in brief in Shuklaand Otten127.) The last two studies128,129 are notformal economic evaluations since costs were notcompared with outcomes for the interventionsbeing assessed. However, they do include thebenefit of response to treatment in terms ofreduced costs and therefore provide useful data.

Gilmore and Milne123 assessed the cost-utility ofIR-MPH compared with placebo for childrendiagnosed using DSM Criteria for PervasiveADHD/ADDH2 or Barkley’s research criteria130

who are otherwise normal. Gilmore and Milne123

argue that they chose to use placebo-controlledtrials as placebo effects are important.

The NHS healthcare perspective was adopted andutility information was determined as described inthe QoL section above. In terms of resource use,the dosage of MPH and the average number ofoutpatient clinic attendances over the year wereestimated based on the opinion of five child andadolescent psychiatrists (personal communication:it is likely that paediatricians’ case loads havedifferent population characteristics and that thiswill have significant implications for the numberof attendances. It should be acknowledged thatthe growing development of nurse-led ADHDservices will also have economic implications). Itwas assumed that all follow-up was hospital-based,that those who terminated treatment or those whodid not respond were treated for 6 weeks, onaverage, and that those who were included in theanalysis for 1 year received five outpatientappointments. Additional information on drugdosages was obtained from the literature, and dataon children’s weights were taken from percentilecharts. The cost of the drugs was obtained fromMIMS (August 1997)131 and the cost of child and


81


adolescent psychiatry and family therapyoutpatient clinics was obtained from fund-holdingtariffs of four Trusts in the South West region.Prices relate to the year 1997.

The incremental cost per QALY gained overplacebo with MPH per child was £9177 (range£5965–14,233) assuming a 70% response rate. A few multi-way sensitivity analyses, that is, varyingmore than one parameter at a time, wereundertaken to test the robustness of findings toplausible variations in QoL improvements, theresponse rate and costs. Under the most optimisticscenario, the cost–utility estimate was £5965 andthe most pessimistic scenario was £29,049.

In addition to the caveats mentioned in the reviewof clinical effectiveness in Chapter 4 relating tothe outcomes measurement, there are a fewconcerns about the economic evaluation data.Little information was provided about the process of obtaining resource use informationfrom the experts. NICE guidance states that thereference case comparator should includealternative therapies routinely used in the NHSrather than a placebo alternative. However, it isrecognised that such head-to-head data may notbe available.

Lord and Paisley4 conducted an economicevaluation based on data from the MTA trial.They compared combined treatment, based on IR-MPH and BT, versus BT alone. The analysis wasconducted from the NHS perspective over 14months, the length of the MTA trial, and only theincremental costs of medication were consideredsince the cost of BT was assumed to be common toboth interventions. A total of 94% of childrenstarted with a 28-day dose titration, taking anaverage of 10 mg of MPH per day, 70% took anaverage of 30 mg per day over the 13-monthmaintenance period, 12% took an average dose of15 mg per day of DEX over the 13-month

maintenance period and 2% received an averagedose of 50 mg per day of imipramine. Half-hourconsultations with a pharmacotherapist were alsoincluded. It was assumed that two visits were madeduring the titration period and then monthly visitsthroughout the maintenance period. Additionally,it was assumed that 6% of children did not starttitration and that 7% of children who did notmake these visits remained persistentlyunmedicated during the maintenance period.

There were 19 different outcome measures used inthe MTA trial and Lord and Paisley4 chose theteacher version of the SNAP-IV index ofhyperactivity/impulsivity since they argued that noHRQoL measure was available and that SNAP-IVwas the most similar to the CTRS used in theCCOHTA economic evaluation.

The best estimate of the incremental cost-effectiveness ratio (ICER) for combined therapyversus BT was £1600 per one SD in the SNAP-IVmeasure (UK 1999 £) as seen in Table 64. A fewone-way sensitivity analyses were conductedincluding varying the incremental cost ofcombination therapy (average of £750 per patientover 14 months) over BT from £500 to £1000.Sensitivity analysis conducted on the ICERsuggested that the ratio could be in the range£700–4500.

The study of Lord and Paisley4 was primarilybased on the MTA that was conducted in the USA,where the diagnosis and treatment of ADHDdiffers in a few respects from the UK, namely thatthe diagnosis criteria are less stringent. Thebehavioural treatment in the MTA trial was moreintensive than typical treatment in the UK. Littleinformation was reported about the two-waysensitivity analyses conducted and a number ofassumptions included in the analysis do notappear to have been tested in the sensitivityanalysis. As the authors mention, the MTA trial


82

TABLE 64 Cost-effectiveness estimates based on the MTA trial

Incremental effect (£) (standardised mean difference in SNAP-IVteacher hyperactive/impulsive dimension at 14 months)

Incremental cost (£) Lower CL Mean Upper CL0.22 0.47 0.72

500 2,273 1,064 694750 3,409 1,596 1,042

1,000 4,545 2,128 1,389

CL, confidence limit.

did not include a ‘global’ HRQoL measure for useas the primary outcome measure in an economicevaluation.

The CCOHTA analysis by Zupancic andcolleagues126 undertook a cost-effectivenessanalysis using a decision analytic model to comparesix interventions including the pharmacologicalintervention magnesium pemoline (PEM). Thebase case analysis was conducted excluding PEMand this is of most relevance in the UK where PEMis no longer prescribed owing to an increased riskof liver failure. Therefore, five interventions werecompared including two pharmacologicalstrategies: IR-MPH and DEX, onepsychological/BT and one combination ofpsychological/BT and IR-MPH for ADHD inchildren and adolescents. The interventions werecompared with a no treatment alternative. Theanalysis was conducted from the Canadian, third-party payer perspective and, like the Gilmore andMilne study123 a 1-year time horizon was adopted.

The effectiveness measure used in the economicevaluation was the Abbreviated CTRS andestimates were derived from the meta-analysis ofpublished clinical trials. The model determinedcost in relation to a one- and a six-point reductionin mean CTRS. A six-point reduction in CTRS wasconsidered as a valid and reliable indicator of aclinical response to treatments for ADHD,54

corresponding to approximately one SD in thedistribution of the CTRS in the studies analysed.The CTRS is widely used in trials and containscore and associated features of children withADHD that the authors believed to be important.The CTRS contains 10 items that have been foundto be sensitive indicators of medication effects.Five items relate to core ADHD symptoms ofinattention/distractibility, hyperactivity andimpulsivity and five to commonly associatedcharacteristics including social and academicadjustment problems that these children

experience (disruptive behaviour, inconsistency,low frustration tolerance, emotional lability). Twokey assumptions are implicit in using the CTRS asa continuous rating scale, that is, that the cost anddesirability of achieving a small gain in CTRSscore for many children are assumed to be thesame as those of achieving a large gain in CTRSscore for few children and that efficacy is constantacross baseline levels of ADHD severity. However,the efficacy of stimulants may depend on thequality and severity of symptoms. An attrition rateof 35% was modelled over 6 months based on aprevious study (Miller and colleagues, unpublishedwork) and at 1 year this was estimated to be 15%.

The costs of care included all interventions(including drugs and/or BT contacts), doctor visitsand hospitalisations. Resource use was based onevidence in the literature and three expert panels.Unit cost data were obtained from the literatureand were expressed in 1997 Canadian dollars.Typically in Canada, the costs for psychologicaltherapies accrue to families except where theservice is obtained from the public sector; however,for the sake of consistency, it was assumed thesecosts were borne by the Ministry. Children on IR-MPH were assumed to have four doctor visits andtwo specialist visits per year and two laboratorytests at baseline and at 1 year. Children on DEXwere assumed to have three doctor visits and twospecialist visits. BT included 16 hours ofcounselling, 8 hours of parent training and 2 hoursof teacher training. Combined therapy combinedthe resource use of BT and IR-MPH above.Children on no treatment were assumed to receivean additional four doctor visits compared with theirunaffected peers. It was assumed that the childrenremained on drug treatment for 1 year and thatany adverse drug reactions, beneficial effects andcosts ceased with discontinuation of the therapy.

Analysis of expected costs and outcomes ofdifferent options indicated that IR-MPH was the


83


TABLE 65 Expected costs and effects of alternative strategies, excluding PEM

Strategy Cost (Can$) Incremental Effectiveness Incremental Incremental cost (Can$) (CTRS points) effectiveness cost-

(CTRS points) effectiveness

Do nothing 128 0IR-MPH 559 431 6.7 6.7 64DEX 566 7 4.7 –2.0 DBT 1946 1380 0.3 –4.4 DBT/IR-MPH combination 2505 559 3.8 3.5 D

D, dominated.

most cost-effective option for the management ofADHD at Can$64 per CTRS point or Can$384 fora six-point, one-SD gain, as shown in Table 65.(For the PEM inclusive programme, high-dosePEM was the next most cost-effective choice;however, its effectiveness data were obtained fromone small study in 28 patients and noconsideration of hepatotoxicity was made.) DEXdid not show a six-point reduction in CTRS butthis estimate of lower efficacy was not statisticallysignificant. BT and combination therapies werenot shown to be effective in producing clinicallysignificant outcomes.

A number of one-way and extreme-case sensitivityanalyses were undertaken for both costs andeffects as shown in Table 66. A number of variableswere tested in the sensitivity analyses and MPHremained the dominant strategy under mostassumptions. The results were not sensitive to theupper CI of effectiveness data for DEX, BT andcombination therapy. Under the worst-casescenario that favoured BT, the combinationtherapy was no longer dominated; however, it wasstill relatively less cost-effective (compared with notreatment) than IR-MPH.

It is worth noting that the cost-effectivenessestimates tested in the sensitivity analyses werebased on average cost-effectiveness (with the effectbeing IR-MPH effect minus the no treatment

effect) per effect rather than incremental cost-effectiveness. There are a number of other caveatsthat might be considered. Perhaps mostimportantly one might question the key modelassumption that improvement in behaviouralrating scales is a good surrogate for clinicallysignificant improvements. It is not clear, from thepresentation of the results, whether thedistribution of change in CTRS is normal and,unless this is so, estimating the number of patientsexperiencing a six-point reduction using thereported overall mean CTRS will not be accurate.It has been shown that a change in the meanCTRS score can give a different outcomecompared with calculating change in numbers ofrespondents, if the distribution is non-normal(Foster N, personal communication, 2004).Therefore, it may be appropriate to estimateresponse on an individual, patient-level basis.

The authors note that there is no proven decreasein drug effectiveness over time and that anychange in length of drug therapy should result inproportionate changes to both costs and effectsover time and therefore that the results may begeneralised to any time horizon. However, thesame may not be true of non-medical therapysince, as Zupancic and colleagues126 hypothesise,BT effectiveness might change over time, arguingthat skills learnt in early counselling sessionsmight be forgotten and therefore effectiveness


84

TABLE 66 Sensitivity analyses (excluding PEM)

Variable tested Cost-effectiveness with reference to non-treatment comparator (Can $)

IR-MPH DEX BT Combination

Base case 83 D D DGeneric IR-MPH 75 D D DSchool days only 119 D D 630120% clinician fee 91 ED D D80% clinician fee 76 D D DFewer counselling hours but same effect 83 D D DConfidence limits 95 D D DIR-MPH low 74 D D DIR-MPH high 83 D D DDEX low 83 D D DDEX high 83 D D DNo treatment low 83 D D DNo treatment high 83 D D DCombination low 83 D D DCombination high 83 D D 311Worst-case scenario (favouring BT) 103 D D 196Weight 16 kg 66 ED D DWeight 40 kg 101 D D D

D, dominated; ED, extended dominated, in weighted average with no-treatment comparator.

might decrease but costs would be expected toremain constant.

Other study limitations include the significantheterogeneity of efficacy studies (includingdecreased power and relatively poor quality ofeffectiveness studies on the psychological andcombination interventions), the sensitivity ofresults to patient drug compliance and the 1-yeartime horizon. The latter assumption may bias theresults against psychological and combinationtherapies since the intervention is received for16 hours yearly rather than daily as for the druginterventions. BT effectiveness was based on asynthesis of two studies. The definition of BT isbroad and may differ considerably across studies.

Marchetti and colleagues128 used a decisionanalytic model to compute total expected directcosts for the treatment and management ofschool-aged children with ADHD, comparing sixpharmacotherapies: ER-MPH8, IR-MPH,Metadate CD (branded ER-MPH8), Concerta(branded ER-MPH12), Ritalin (branded IR-MPH)and Adderall (a combination of DEX andamphetamine salts). It is worth noting thatAdderall is not licensed for use in the UK. Alltreatments required midday dosing apart fromMetadate CD and Concerta. The evaluation wasconducted in the USA from the payer perspective,although school-related costs were included.

A clinical algorithm for ADHD management wasdeveloped, based on expert opinion, for use as theanalytical framework for an economic model tocompute total expected costs. Office visits,laboratory tests and use of therapeuticinterventions were estimated by the clinicians, andunit costs from the literature were applied.Dosages for the treatments were calculated basedon clinical trial data or manufacturers’instructions. To cost the use of midday medicationtaken at school, a telephone survey of schoolnurses or members of staff at eight publicelementary schools in four US States wasundertaken.

Probabilities of clinical success, failure and relatedinformation were calculated based on a meta-analysis of studies in the published literature. Aneffect size was calculated for each outcome in eachstudy based on subtracting the final group meanoutcome for the placebo group from that of theintervention group and dividing by the pooledstandard deviation of the data. The mean of alleffect sizes from all studies was calculated andthen recalculated omitting each study in turn to

give a set of individual effect sizes using Tukey’sjack-knife method. The individual effect sizes werethen weighted by the number of studies andsubtracted from the overall mean to provide aneffect size representing the true overall effect ofthe intervention irrespective of the outcomemeasure used. Finally, the effect sizes werecombined in a random-effects model to estimatethe underlying response rate.

Response rates for Metadate CD, IR-MPH andAdderall were computed based on 10 articlesdeemed acceptable to reviewers for data extractionand analysis. There were no statistically significantdifferences among these response rates. Responserates of the other interventions were not availableto meta-analyse. The response rates for Ritalinand ER-MPH8 were assumed to be the same as IR-MPH and, to estimate the response rate forConcerta XL (ER-MPH12), the mean responsesfor Metadate CD, Adderall and IR-MPH wereused.

The evaluation period for the model was 4 weeks,as the clinical experts estimated that this was theaverage time taken to evaluate response tomedication. The model assumed that if thepatients responded to treatment they wouldcontinue to respond over 1 year. If patients didnot respond within 4 weeks or if they experiencedadverse events related to the medication, it wasassumed that the dose would be adjusted inclinical practice and the patient would be re-evaluated after a further 4 weeks. If at this timepatients did not respond, they would be re-evaluated and the medication would be switched.If the patient did not respond after fourevaluations (16 weeks), the model assumed thatthe patient would require non-drug interventionsin addition to primary ADHD care. Metadate CDhad the lowest total annual per patient expectedcost ($1487), followed by Concerta XL ($1631),ER-MPH8 ($1792), IR-MPH ($1845), Ritalin($2080) and Adderall ($2232).

Rank order (one-way) sensitivity analyses wereconducted to test the robustness of the results tochanges in the drug acquisition cost (per tablet)and the cost of in-school dosing. Thresholdanalyses for drug acquisition costs and responserates were also undertaken. The rank ordering ofresults remained fairly robust to these tests.

Compliance was not considered in the analysis.However, the authors state that once-daily dosingtends to have a compliance benefit over twice-dailyand three times daily dosing. Lack of compliance


85


associated with multiple daily dosing is expected toresult in higher total costs. The impact of co-morbid conditions on the costs of care was notassessed, nor were patient preferences or QoLvariables. The model relies on assumptions aboutthe relative efficacy of the drugs evaluated anddoes not have a strong evidence base. MetadateCD was assumed to have a higher response ratethan its generic counterpart, ER-MPH8, and theauthors failed to justify this.

Vanoverbeke and colleagues129 also used a decision-analytic framework to model costs for managementof ADHD in 6–16-year-old children in the UK.Medications compared starting treatment with IR-MPH (once, twice or three times daily), ConcertaXL or BT over a 1-year time horizon.

An incidence-based decision tree was constructedand the probabilities of success or failure werebased on average probabilities derived from theliterature. The probabilities for second-linetreatment were obtained from an expert panel ofeight UK psychiatrists and paediatricians, as weredata on treatment choices in response to adverseevents, co-morbidities and/or insufficient responseto treatment. Six out of the eight experts involvedalso estimated resource use data for a typicalADHD patient requiring treatment. To obtain thedata from an expert panel a two-stage approachwas followed, including a questionnaire completedindependently. Group average responses andrange of responses for each item were presentedand the experts were asked to provide newestimates and a ‘certainty score’ to indicate theexpected variability associated with a given value(of between one and four).

Costs were based on published estimates andhospital prices and relate to 2001. Costs ofmedications, laboratory tests, clinical personneland school staff personnel involved in BT wereincluded. Clinical outcomes for BT and IR-MPHwere obtained from the MTA trial and for ConcertaXL were obtained from Pelham and colleagues.82

The cost of starting treatment with IR-MPH wasmarginally lower than with Concerta XL (£1332and £1362, respectively) and BT was the mostcostly initial treatment (£2147). The probability oftreatment success was highest for Concerta XL(77.8%),82 then IR-MPH (55.6%),133 followed byBT (33.8%).133

Data from different trials were used in the modeland this breaks the randomisation achieved in theindividual trials. As the authors state, Pelham and

colleagues’ study82 on which the Concerta XLclinical outcomes are based was a short-term, small-scale study. A probabilistic sensitivity analysis wasundertaken which showed that results were sensitiveto treatment success and the proportion of patientswith co-morbidities. Although the sensitivityanalysis did not alter the results, the response ratesused in the model may be questioned.

In summary, across the five studies reviewedabove, all were based on a 1-year time horizon,with the exception of the Lord and Paisley study4

that covered a period of 14 months. However,ADHD and treatment are known to continue formuch longer. Therefore, no consideration of long-term adverse events or outcomes is incorporatedwithin the analyses. None of the full economicevaluations compared all treatment strategiesrelevant to this review. Zupancic and colleagues126

did compare a number of treatments, but noassessment of ATX drug therapy, which isnecessary for this review, was provided. A commonfeature across all studies is the lack of data, withexpert/author opinion being used to fill in gaps.

In addition to the existing economic evaluations,three submissions were received from Janssen-Cilag, Celltech and Eli Lilly.

Review of the Janssen-CilagsubmissionOverviewThe aim of the Janssen-Cilag submission was tocompare Concerta XL (ER-MPH12) with IR-MPH,ATX, Equasym XL (ER-MPH8) and behaviouraltherapy (BT) using new evidence made availablesince the previous NICE guidance was issued.134

The previous guidance recommended the use ofIR-MPH as part of a comprehensive treatmentprogramme, including advice and support toparents and/or teachers and potentially BT, forchildren diagnosed with severe ADHD. ‘SevereADHD’ was defined as broadly similar to HKD,although it also includes some patients with severeproblems with inattention and/or hyperactivitywho do not meet the diagnostic criteria for HKD.Treatment initiation was restricted to child andadolescent psychiatrists or paediatricians, butprescriptions could then be maintained by GPs.

A cost–utility analysis (classified CIC) wasconducted based on the results of two recentrandomised, open-label studies comparingConcerta XL with IR-MPH90 and ATX99 and theMTA trial.133 The model took the form of a


86

decision tree with a time horizon of 1 year, andwas conducted from the perspective of the UKNHS. A ‘no treatment’ arm was not included onthe basis of the previous NICE guidance.134

Model structureIn the model, patients begin therapy on BT, IR-MPH, Concerta XL, ATX or Equasym XL. If theyresponded to this first-line therapy andexperienced no side-effects, then they wereassumed to continue on the same treatment forthe rest of the year. If patients failed to respond tofirst-line treatment, or experienced intolerableside-effects within 1 month, they could then switchto second-line treatment with BT (where BT wasnot the first-line therapy), combination treatment(BT plus the first-line pharmacotherapy) or otherdrug treatment. The BT components were notdefined explicitly in the model, but as theeffectiveness data are sourced from the MTAtrial133 we may assume that they correspond tothat trial protocol. The other drug treatment wasassumed to be DEX for patients receiving first-linepharmacotherapy and IR-MPH for patientsreceiving first-line BT. Patients responding tosecond-line treatment remained on that therapy

for the remainder of the year, and patients notresponding within 1 month of beginning second-line therapy were assumed to discontinue alltreatment. Figure 21 presents a flow-chart of themodel. Owing to the 1-year time horizon,discounting was not necessary.

Summary of effectiveness dataThe measure of clinical effectiveness used in themodel was response rate to treatment. For thecomparison of Concerta XL with IR-MPH, theeffectiveness evidence was taken from the CON-CAN-1 study.90 The definition of a response wastaken from Swanson and colleagues133 as a meanscore of ≤1 on the parent-rated SNAP-IV scale.The same definition of response was used toextract effectiveness data for BT from the BT-onlyarm of the MTA trial. For the comparison ofConcerta XL with ATX, the effectiveness evidencewas taken from the FOCUS study.99


The model also allowed patients to discontinuetherapy owing to inefficacy or adverse eventswithin 1 month. The data on discontinuation rates


87


Continue

Continue

No treatment

Response

Noresponse

Continue

No treatment

Response

Noresponse

Continue

No treatment

Response

Noresponse

BT

Combinationtreatment

2nd-linepharmacotherapy

Noresponse orside-effects

Response, no side-effects

1st linepharmacotherapy

FIGURE 21 Flow-chart of the cost–utility model submitted by Janssen-Cilag

were based on CON-CAN-190 for Concerta XLand IR-MPH.


The response rates and discontinuation rates usedin the submission are shown in Table 67.

The proportions of patients switching to BT,combination treatment or DEX as second-linetherapy were based on a previously publishedstudy that used expert opinion.129 Theseproportions are shown in Table 68.

Summary of resource utilisation andcost dataThe resource use associated with managing ADHDpatients on BT, pharmacotherapy and

combination treatment was based on an update ofa previously published study that used expertopinion.129 Resource use quantities were estimatedseparately for responders and non-responders andfor the presence of co-morbidities. The panelestimated the annual costs of each programme,including the cost of the BT, the cost of follow-upconsultations and the cost of monitoring tests, andthese were divided into monthly figures for thepurpose of the model. In order to do this, annualcosts were divided by 12, with the exception of thetreatment cost of BT. This was assumed to takeplace over a period of 8.63 weeks, and so the costof this was divided by 8.63 and multiplied by4 weeks (assuming 4 weeks = 1 month), and wasonly applied in the first 2 months in the model.Table 69 shows the relevant annual costs for eachstrategy.


88

TABLE 67 Response rates and discontinuation rates used in the cost–utility model submitted by Janssen-Cilag

Treatment Response rate Discontinuation rate

BTIR-MPHConcerta XLATX [Confidential information removed]Equasym XLCombination treatmentDEX

TABLE 68 Proportions of patients switching to alternative second-line therapies in cost–utility model submitted by Janssen-Cilag

Switch from

Switch to 1st line BT (%) 1st line pharmacotherapy (%)

IR-MPH 57.92 NACombination treatment 42.08 37.42BT NA 17.35DEX NA 45.24

NA, not applicable.

TABLE 69 Annual treatment, follow-up and monitoring costs used in cost–utility model submitted by Janssen-Cilag

Item Behavioural therapy (£) Drug treatment (£) Combination treatment (£)

Programme cost 1033 Varies 1033 + drug cost

Consultations:Responders 333 737 737Non-responders 808 1012 1012Co-morbidities 652 349 349

Tests:Responders 5.3 0.5 0.5Non-responders 43 79.6 80Co-morbidities 6.0 4.9 4.9

The follow-up consultations and monitoring testsinclude those resulting from adverse events. Thecosts were inflated to 2003 values using theinflation rate for hospital and community healthservices in the UK.135 For the purpose of themodel, it was assumed that 50% of patients hadco-morbidities. The cost of a patient discontinuingall therapy was assumed to be equal to that of anon-responder to behavioural therapy.

The unit costs of the pharmacotherapy were basedon published sources for those treatmentsavailable in the UK.29 The cost per milligram ofeach of the drugs is £0.019 for DEX and IR-MPH,£0.050 for 18-mg tablets of Concerta XL and£0.034 for 36-mg tablets of Concerta XL.


The dosing schedules considered are shown inTable 70.


The price of Equasym XL was not available for theUK, so it was assumed to be £1 per day, which was

entered into the model as 0.8 times the weightedaverage cost of Concerta XL. The annual costs ofeach drug are shown in Table 71.

Summary of utility dataThe utility data were based on a previouslypublished poster136 that obtained UK-based EQ-5D scores elicited from parents acting as proxiesfor their children with ADHD. The utility valueswere 0.837 for responders, and 0.773 for non-responders, regardless of treatment type.


Summary of cost-effectivenessThe results from the deterministic model areshown in Table 72. First-line Equasym XL, BT andATX were all found to be dominated by ConcertaXL. A treatment dominates an alternative when itis less costly and more effective. The cost perQALY gained with Concerta XL compared withIR-MPH is well within the commonly used valuesof willingness to pay for a QALY.


The data used for the probabilistic analysis areshown in Tables 73 and 74.

The CIs around the non-drug resource use arerelatively wide. It is not clear, either from thesubmission or the previous publication, whetherthese estimates represent first- or second-orderuncertainty. One would expect first-orderuncertainty to be greater than second-orderuncertainty, which may explain the width of theCIs, but this would mean that the uncertainty inthe model results is slightly overestimated.


89


TABLE 70 Data for weighted average daily dose in the cost–utility model submitted by Janssen-Cilag

Dosing IR-MPH or DEX Nearest equivalent dose Concerta XL Patients continue on dose (%)


TABLE 72 Cost-effectiveness results from the model submitted by Janssen-Cilag

First-line treatment Cost per patient £) QALYs per patient ICER

IR-MPH –Concerta XL 4992Equasym XL [Confidential information removed] DATX DBT D

TABLE 71 Annual costs of pharmacotherapy used in thecost–utility model submitted by Janssen-Cilag

Drug Annual cost (£)

DEX 112IR-MPH 110Concerta XL 381Equasym XL 317ATX 914


90

TABLE 73 Effectiveness data for probabilistic analysis in the cost–utility model submitted by Janssen-Cilag

Parameter Mean SE

Response rate (%)BTIR-MPHConcerta XLATXEquasym XLCombination treatmentDEX [Confidential information removed]

Discontinuation rate (%)BTIR-MPHConcerta XLATXEquasym XLDEX

Switching probabilities (%)From 1st-line BTTo IR-MPH 57.92To combination treatment 42.08From 1st-line pharmacotherapyTo BT 17.35To combination treatment 37.42To DEX 45.24

SE, standard error.

TABLE 74 Annual cost data for probabilistic analysis in the cost-utility model submitted by Janssen–Cilag

Parameter Mean (£) Minimum (£) Maximum (£) Median (£)

BTProgramme cost 1033Consultations – responders 333Consultations – non-responders 808Consultations – co-morbidities 652Tests – responders 5Tests – non-responders 43Tests – co-morbidities 6

PharmacotherapyConsultations – responders 737Consultations – non-responders 1012 [Confidential information removed]Consultations – co-morbidities 349Tests – responders 0Tests – non-responders 80Tests – comorbidities 5

Drug costsIR-MPH 110Concerta XL 381DEX 112ATX 941Equasym XL 317

The probabilistic results were used to constructcost-effectiveness acceptability curves (CEACs)comparing all five treatment strategies, accordingto the method described by Fenwick andcolleagues.137


Comments on methodologyTime horizonAlthough ADHD is a chronic condition, thesubmission assesses the cost-effectiveness of thealternative treatment strategies over the course of 1 year. [Confidential information removed].It is important to note that although the treatment decision may be reviewed annually, thedecision problem at the point of review is differentto the initial treatment decision. At the point ofreview, the GP will know which treatments thepatient has failed to respond to, which causedadverse events and which produced a favourableresponse.


The consequence of the short time horizon is thatlong-term implications of treatment for ADHD arenot considered. As a consequence, it is difficult toknow how cost-effective Concerta XL wouldappear over a time horizon longer than 1 year.

Use of trial dataThe cost–utility model in the submission madevery selective use of the body of evidenceconcerning the efficacy of the alternativetreatment strategies. The effectiveness evidencewas based on only three trials,90,99,133 out of 65potentially relevant studies identified in theclinical effectiveness review in Chapter 4. Thejustification for ignoring the majority of the RCTdata was that double-blind, double-dummy trialscannot assess the effectiveness of a once-dailyextended release formulation of IR-MPH incomparison with twice- or three times dailyadministration of IR-MPH as both arms receivethe same number of pills. Janssen-Cilag argue thatcompliance to a product given once in themorning may be higher than one needing to betaken during the school day, and hence they makeuse of two open-label pragmatic randomisedtrials.90,99


The response rates to BT, [Confidentialinformation removed] were taken directly fromthe MTA trial.133 This extraction of data from

single arms of a trial breaks the randomisation,and implicitly assumes that the baselinecharacteristics of the patients in the CON-CAN-1,FOCUS and MTA trials are identical. This isunlikely to be the case and this methodology isinappropriate. A more appropriate approachwould have been to take the relative treatmenteffects from each study and apply them to acommon baseline. Finally, the definition ofresponse for ATX [Confidential informationremoved] differs from that used for the othercomparators considered in the model (£1 meanparent-rated SNAP-IV score). This is most likelydue to a lack of outcome data employing theSNAP-IV scale for ATX. The submission does notprovide evidence as to the comparability of thesetwo definitions of response.

The discontinuation rates were also based only onCON-CAN-1 [Confidential information removed].This limited use of data will fail to incorporatedifferences in side-effect profiles between thesedrugs. The costs associated with adverse eventswere also not differentiated by drug. In this respect,it is important to consider that the minimum doseof Concerta XL is 18 mg per day, as it is notpossible to divide the extended-release capsules.This compares with a minimum dose of 5 mg withthe other drugs, and so patients requiring <18 mgper day, or >18 mg and <36 mg per day, wouldreceive a higher dose of IR-MPH on Concerta XLcompared with, for example, IR-MPH (see Table70). The submission by Celltech states thatEquasym XL will be available in 10, 20 and 30 mg.It is also important to consider that Concerta XL,owing to its 12-hour action, is suitable for replacingthree times daily dosing of IR-MPH, and so maynot be suitable for patients requiring once- or twice-daily IR-MPH. Likewise, Equasym XL, with its 8-hour action, would not be suitable for a patientrequiring IR-MPH once daily, but could becombined with an evening dose of IR-MPH forchildren requiring doses three times daily. A furtherconcern is that the definition of discontinuation isgiven as ‘discontinuation due to inefficacy oradverse events’, which may overlap with non-response. Those patients who withdraw fromtreatment owing to inefficacy would be included inthe calculation of response rate in an ITT analysis,and thus be represented in both the discontinuationrate and the rate of non-response. In the model,these patients would be counted twice, andtherefore the model may overestimate the numberof patients discontinuing treatment owing toinefficacy or adverse events. The model results wereshown to be sensitive to the choice ofdiscontinuation rate.


91


Model structureThe model assumes the same response tocombination treatment (i.e. drug treatment plus BT), regardless of the accompanying drugtherapy. The model allows the combinationtreatment to include the drug to which the patient has previously failed to respond, orexperienced adverse events with as first-linetherapy. A more realistic assumption would havebeen to have combination treatment with DEX,according to the assumption used for third-linepharmacotherapy in the model. This amendmentforms option 1 for the re-analyses shown in Table 75.

Errors in modelAs mentioned earlier in the review, the annual costof each drug excludes the cost during the averagetitration period [Confidential informationremoved]. This appears to have been omitted inerror. Correcting for this error forms option 2 forthe re-analyses shown in Table 75. An alternativemethod would be to calculate more precisely aweighted average cost using the actual number ofdays spent on each dosage, rather than theaverage over all doses. The monthly cost of apatient discontinuing all treatments is assumed tobe equal to that for a non-responder to BT, but inthe model the cost also includes the total cost of aBT treatment programme divided by 12. Thereasons for this are unclear; such a patient wouldnot receive the BT programme, only the follow-upconsultations and tests. This extra monthly cost isfairly high (£86), and the effect of its removal canbe seen in option 3 in Table 75. Another errorappears to be present in the way that the cost ofco-morbidity is included in the model. As Table 69shows, the annual medical cost associated with co-morbidities for patients receiving medicaltreatment is lower than that for responders andnon-responders. Hence the 50% of non-responders who are assumed to have co-morbid

conditions cost less than those without co-morbidities. It appears that the quoted cost of co-morbidities actually refers to the additional cost ofco-morbidities, on top of medical costs forresponders and non-responders. This formsoption 4 in Table 75.

The assumption that the utility for the first monthof any treatment is equal to [Confidentialinformation removed] seems rather arbitrary. Analternative approach would have been to observethe number of patients responding at 1 month foreach treatment, and assume that on average thesepatients responded half way through the firstmonth.

When re-run, the results of the probabilisticanalysis proved to be fairly variable. This can beovercome by increasing the number of simulationsfrom 1000 to, for example, 10,000.

In the probabilistic sensitivity analysis, theproportions switching to each second-line therapywere modelled independently. This allowed thesum of those proportions to fall below or exceedone in nearly all simulations. Clearly, it is notappropriate in a decision-modelling context tohave proportions or probabilities not summing to one when all possible paths have beenidentified. Therefore, for the two transitionspossible following first-line BT, this error iscorrected by making one of those probabilities (forexample, the probability of switching to DEX)equal to one minus the other. For the threetransitions possible following first-linepharmacotherapy, this error is corrected bydichotomising the three-way transition andadjusting the estimated probabilities appropriatelyusing Bayes’ revision theorem. These adjustmentsensure that the number of patients in the modelremains constant and does not increase ordecrease.


92

TABLE 75 Results of amendments to the cost–utility model submitted by Janssen-Cilag

Amendment Comparison Incremental cost per QALY (£)

None – base case Concerta XL vs IR-MPH 4,992

Second-line combination therapy with DEX (1) Concerta XL vs IR-MPH 4,903

Including drug costs during titration (2) Concerta XL vs IR-MPH 7,954

Taking BT component out of cost of discontinuing treatment (3) Concerta XL vs IR-MPH 19,303

Including co-morbidity costs as additional to responder/non-responder costs (4) Concerta XL vs IR-MPH 61

All of the above (1-4) Concerta XL vs IR-MPH 9,253

The review also uncovered a mis-reference incalculating the costs of behavioural therapy.Correcting this error did not change the results asbehavioural therapy was consistently dominated. Were-analysed the model by Janssen-Cilag, correctingfor each of the detailed errors in turn. The resultsof these re-analyses are shown in Table 75.

Review of the Celltech submissionOverviewThe aim of the Celltech submission was tocompare Equasym XL (ER-MPH8) with notreatment in patients unable to comply with twice-daily IR-MPH, using new evidence made availablesince the previous NICE guidance was issued.134

The submission therefore concentrates on dataconcerning the effectiveness of Equasym XL,which was not included in the previous review. Asecondary analysis also compared Equasym XLwith no treatment and twice-daily IR-MPH.

A partially probabilistic cost–utility analysis wasconducted based primarily on reviews of treatmentfor ADHD. The model took the form of a decisiontree with a time horizon of 1 year, and wasconducted from the perspective of the UK NHS.Concerta XL (ER-MPH12) and ATX were notincluded as comparators.

Model structurePatients enter the model on MPH, in the form ofEquasym XL, in the base case. IR-MPH is added asa second comparator in the secondary analysis.After a 42-day titration period, patients may complyor not with treatment. Non-compliers are assumedto continue on treatment, but experience no healthbenefits. Among those complying with treatment,the proportion that experience a response and nointolerable side-effects are assumed to remain ontreatment for the rest of the year; the proportionwho do not respond, or who experience intolerableside-effects, progress to second-line treatment withDEX. Second-line therapy follows the same patternas first-line therapy, with a 42-day titration period,compliers/non-compliers and discontinuations dueto inefficacy or adverse events. Those patients whodiscontinue DEX may progress to BT or notreatment. In the base case it is assumed that 50%of these patients will progress to BT and experiencehealth benefits, whereas the remaining 50% willprogress to no treatment and receive no healthbenefits. The BT was described as intensivepsychosocial treatment involving eight visits tomembers of child/adolescent psychiatry orpsychology teams.

Patients progressing through the model werecompared with ‘no treatment’ in the base case,which was assumed to incur no health costs orhealth benefits. Owing to the 1-year time horizon,discounting was not necessary.

Summary of effectiveness dataThe model assumed that compliance withmorning doses of each drug would be 85% (95%CI 50 to 100%) and compliance with lunchtimedoses of twice-daily drugs would be 55% (95% CI40 to 85%). The uncertainty around theseestimates was characterised using a normaldistribution. These values were chosen tocorrespond to reported figures of overallcompliance to IR-MPH of 65–75%.41,138 Hencecompliance to Equasym is calculated to be 85%and compliance to IR-MPH or DEX is, onaverage, 70%.

The measure of effectiveness used in the modelwas response rate to treatment. The modelassumes that response rates to IR-MPH, EquasymXL and DEX will be equal, based on previouslypublished evidence that they are similar.59,139 Theresponse rate is set at 70% (Scottish IntercollegiateGuidelines Network, 2001).4,123 A definition ofresponse is not provided. The model also assumedthat around 6% of patients beginning any of thedrug therapies would discontinue owing toadverse events,123 which gives a continuation rateof 66% (70% of 94%). The uncertainty around thisestimate was characterised using a normaldistribution where the upper and lower bounds ofthe 95% CI were assumed to be 60 and 82%,respectively.

Summary of resource utilisation andcost dataThe unit costs of medication were based onpublished pricing lists for the UK (BNF 47).Equasym XL is not currently priced in the UK,and so the submission includes costs specified bythe manufacturer. The prices used are shown inTable 76.

During the titration period, it was assumed thatone-third of patients would be on the equivalentof 5, 10 and 15 mg of IR-MPH twice daily,respectively. The average dose after titration wasassumed to be 30 mg per day, based on previouslypublished data123 and data from the IMS HealthDisease Analyser Mediplus dataset (reference notprovided in submission). Patients progressing toDEX were assumed to receive 5 mg once daily atthe start of the titration period, and assumed toreach an average of 10 mg per day subsequently,


93


according to a published UK protocol.140 It wasassumed that 50% of the titration period would beat an average dose of 5 mg and 50% at an averageof 10 mg once daily. Non-compliers were assumedto incur the same drug costs as those complyingwith therapy.

The resource use associated with ADHD was basedon Wessex DEC evaluation,123 which used expertopinion to determine treatment patterns. Allpatients receiving drugs were assumed to receivesix outpatient visits with a child psychiatrist orpaediatrician at a cost of £111 per visit, and sixGP visits per year at a cost of £20.141 Patientsdiscontinuing treatment were assumed to receivetwo outpatient visits per year. Patients receivingBT were assumed to receive eight 100-minuteconsultations, 50% with members of achild/adolescent psychiatry team and 50% withmembers of a clinical psychology team. The costof these was obtained from published UKsources,141 and was £64 per person-hour for thepsychiatry team and £39 per hour for thepsychology team. Non-compliers were againassumed to receive the same cost as compliers.

The model did not include any costs associatedwith side-effects of treatment.

Summary of utility dataThe base case analysis used utility values from apreviously published evaluation of treatments forADHD.4 These were calculated using EQ-5D, byassuming that an untreated ADHD patientcorresponded to the EQ-5D health state 11211,which gives a utility of 0.883. A successfully treatedpatient was assumed to return to full health,health state 11111, giving a utility of 1.000.

Summary of cost-effectivenessThe base case analysis compared Equasym XLwith no treatment in patients unable to complywith twice-daily IR-MPH. The estimated total costsand QALYs for no treatment were £0 and 0.883

(no probabilistic parameters were involved in thiscalculation). The total costs and QALYs forEquasym XL were estimated to be £1073 (95% CI£1035 to 1116) and 0.9562 (95% CI 0.9425 to0.9667). The ICER for Equasym XL comparedwith no treatment was therefore estimated to be£14,657 per QALY (95% CI £12,564 to 18,538).The 95% CI for the ICER was calculated by takingthe 5th and 95th percentile of the 10,000probabilistic simulations. In this instance thismethod is valid because all of the simulatedincremental costs and benefits of Equasym XL layin the northeast quadrant of the cost-effectivenessplane (i.e. Equasym XL was more effective andmore expensive than no treatment in allsimulations), and so the problem of negativeICERs did not arise. The probabilistic simulationswere used to plot a CEAC, which showed that at awillingness-to-pay threshold of £30,000 per QALY,Equasym XL was the most cost-effective strategy in100% of simulations.

A number of one-way sensitivity analyses were alsoconducted in which the ICER for Equasym XLcompared with no treatment was shown to remainunder £30,000 per QALY for a range of valuesinput for compliance to the morning dose ofmedication, the utility estimates, response rate,length of titration, cost of paediatric outpatientappointments and drug costs for non-compliantpatients. The results of these are shown in Table 77.

A second scenario considered three times dailydosing, where the dose of Equasym XL must besupplemented with an evening dose of IR-MPH.This increased the costs of Equasym XL to £1.45per day. Compliance to this evening dose of IR-MPH was assumed to be 85%. The reported resultfrom this analysis is an ICER of £15,536 forEquasym XL relative to no treatment.

The secondary analysis compared Equasym XLwith IR-MPH and no treatment. The estimatedcosts and QALYs for the no treatment and


94

TABLE 76 Unit costs of IR-MPH, Equasym XL and DEX used in the cost–utility model submitted by Celltech

Drug Pack Price (£)

DEX 5 mg × 28 2.61IR-MPH 5 mg × 30 2.78

10 mg × 30 5.5720 mg × 30 9.98

Equasym XL 10 mg × 30 25.0020 mg × 30 30.0030 mg × 30 35.00

Equasym XL arms remained the same as in thebase case analysis. The estimated total costs andQALYs for IR-MPH were £930 and 0.9433. Whenmore than two programmes are being compared,the ICERs are calculated using the followingprocess:

1. The strategies are ranked in terms of cost (fromthe least expensive to the most costly).

2. If a strategy is more expensive and less effectivethan the previous strategy, then this strategy issaid to be dominated and is excluded from thecalculation of the ICERs.

3. The ICERs are calculated for each successivealternative, from the cheapest to the mostcostly. If the ICER for a given strategy is higherthan that of the next more effective strategy,then this strategy is ruled out on the basis ofextended dominance.

Finally, the ICERs are recalculated excluding anystrategies that are ruled out using the notions ofdominance and extended dominance. The ICERfor IR-MPH compared with no treatment was£15,432 per QALY and the ICER for Equasym XLcompared with IR-MPH was £11,043 per QALY.Hence in the secondary analysis, IR-MPH wasextended dominated by Equasym XL. The resultsof the secondary analysis were very sensitive to thechoice of compliance rates, and the results of a

two-way sensitivity analysis of these are shown inTable 78.

Although the submission does not clarify whetherthe negative ICERs are the result of a positivedifference in cost and a negative difference ineffect, or of a negative difference in cost and apositive difference in effect, one can assume thatthey were a result of the former as they fall intothose analyses where the compliance to thelunchtime dose is higher than compliance to themorning dose.

Introducing costs for the in-school administrationof a lunchtime dose into this secondary analysisimproved (reduced) the ICER for Equasym XLcompared with IR-MPH.

Comments on methodologyChoice of comparatorsThe model did not include the full range ofcomparators. This was justified in the model byspecifying the potential population to be thosepatients who were intended to receive IR-MPHtwice daily, and so longer-acting alternatives suchas Concerta XL and ATX would not be relevant. Asecondary analysis that considered patientsrequiring IR-MPH three times daily also did notthen introduce those further relevant comparatorssuch as Concerta XL and ATX.


95


TABLE 77 Results of sensitivity analyses conducted in the cost–utility model submitted by Celltech

Parameter value ICER for Equasym vs no treatment (£)

Compliance to morning dose (£)50 26,748

100 12,066

Utility (treated, non-treated)1.000, 0.692 5,5680.970, 0.884 15,238

Daily cost Equasym XL (baseline £1.17) (£)0.58 12,8271.76 16,487

Response rate (%)60 14,83682 14,237

Length of titration period (days)21 13,63470 16,256

Cost of outpatient appointment (£)89 12,991

138 16,702

Reduction in drug cost for non-compliant (%)50 14,383

100 14,108

Time horizonThe model employed a time horizon of 1 year,which excludes consideration of the longer termoutcomes associated with ADHD and its treatment.

ComplianceThe main difference between treatments is theassumed levels of compliance to morning andafternoon doses of medication. Unfortunately, thisinformation was based on expert opinion using anestimate of overall compliance and a two-waysensitivity analysis showed that the results werevery sensitive to the assumption made. Thesubmission also fails to justify why compliant non-responders would be identified and moved to analternative treatment, but non-compliers, who bydefinition fail to respond, would not be identifiedas non-responders.

Review of the Eli Lilly submissionThe electronic model to accompany the Eli Lillysubmission was not examined owing to its late

submission. Therefore, this review reports themodel structure, inputs and results as reported byEli Lilly, but these values could not be validated byexamination of the model.

OverviewThe submission states that a Markov model wasused to examine the cost-effectiveness of addingATX to an existing strategy of medicalmanagement for ADHD. The underlying generalstrategy was that a new patient would be given anMPH formulation first-line, followed by DEX as asecond-line alternative in the event of treatmentdiscontinuation, followed by no treatment. ATXwas added as an additional treatment option priorto MPH. The model considered five subgroupswith varying ineligibility for one or more of thetreatment options. Table 79 details the subgroupsconsidered in the model.

For groups 1 and 5, the treatment strategiescompared are IR-MPH or ER-MPH, followed byDEX, followed by no treatment, with or withoutATX prior to MPH. For groups 2 and 4, the


96

TABLE 79 Subgroups considered in the cost–utility model submitted by Eli Lilly

Subgroup Description

1 Stimulant-naïve patients with no contraindications to any treatment option2 Stimulant-naïve patients with contraindications to stimulants (MPH and DEX)3 Stimulant-exposed patients who have previously failed on MPH therapy owing to adverse events or

inefficacy4 Stimulant-exposed patients with contraindications to stimulants (MPH and DEX)5 Stimulant-exposed patients who have no contraindications to any treatment option, and who have not

failed on previous therapy for either adverse events or inefficacy

TABLE 78 Two-way sensitivity analysis of morning and lunchtime compliance rates in secondary analysis of the cost–utility modelsubmitted by Celltech

ICER for Equasym XL compared to IR-MPH (£)

Lunchtime compliance (%)

Morning compliance (%) 50 55 60 65 70 75 80 85 90 95

50 –90,012 –45,925 –31,229 –23,882 –19,473 –16,534 –14,435 –12,860 –11,63555 84,780 –88,457 –45,148 –30,711 –23,493 –19,162 –16,275 –14,212 –12,66660 40,694 83,226 –86,902 –44,370 –30,193 –23,104 –18,851 –16,016 –13,99065 25,998 39,916 81,671 –85,348 –43,593 –29,675 –22,716 –18,540 –15,75770 18,650 25,480 39,139 80,116 –83,793 –42,816 –29,156 –22,327 –18,22975 14,242 18,262 24,962 38,362 78,561 –82,238 –42,038 –28,638 –21,93880 11,302 13,931 17,873 24,443 37,584 77,007 –80,683 –41,261 –28,12085 9,203 11,043 13,620 17,484 23,925 36,807 75,452 –79,128 –40,48390 7,629 8,981 10,784 13,309 17,096 23,407 36,029 73,897 –77,57495 6,404 7,434 8,759 10,525 12,998 16,707 22,889 35,252 72,342

treatment strategies are no treatment, with orwithout ATX as first-line therapy. For group 3, thetreatment strategies are DEX, followed by notreatment, with or without ATX prior to DEX.

Model structurePatients beginning the model on active treatmentcould experience a response to that treatment.Patients responding to treatment could relapse insubsequent cycles to become non-responders.Patients on active treatment could also experienceadverse events, which may resolve in subsequentcycles, or discontinue with treatment. Patientscould discontinue treatment owing to lack ofresponse, in reaction to an adverse event or forother reasons. Discontinuation of treatment isfollowed by the next treatment in the prespecifiedstrategy, until the patient reaches no treatment atthe end of the strategy.

The model was estimated using patient-levelsimulation, and 20,000 simulations were executedfor each run of the model. The submission statesthat the model was used to determine which of thefive subgroups each simulated patient would fallinto, although further details of this process arenot given. Further details of the execution of thepatient-level simulation are not provided. Thetime horizon was 1 year, and so discounting wasnot relevant.

Summary of effectiveness dataThe response rates were calculated from a meta-regression using patient-level data from fiveclinical trials70,89,142 (two of which are currentlyunpublished) comparing ATX with MPH, and insome cases also to placebo. One of the trialsincluded ER-MPH (unpublished) and theremaining four included IR-MPH. Four wererandomised double-blind studies and the fifth wasa randomised open-label study.70 The modelassumes equivalence of IR-MPH and ER-MPH,and does not differentiate between study types.Response was defined as ≥ 25% reduction inparent-rated ADHD-RS score, and was estimatedusing a fixed-effects logistic regression withtreatment, stimulant exposure, age, sex and

duration of treatment as covariates. Duration oftreatment refers to the duration of the acute phaseof each trial. An additional study-level covariatewas included to allow differences in baselinebetween studies. The results of the logisticregression are shown in Table 80.

Data on relapse was obtained from two trials,56,75

both specifically designed to look at relapse toATX and amphetamine sulphate, respectively. Therelapse rates over 9 months were approximately20% for ATX and 30% for amphetamine sulphate,compared with placebo rates of 40 and 70%,respectively. The submission states that owing todifferences in the definition of response andrelapse used in the two trials, and the absence ofdata regarding MPH, they deem the evidenceinsufficient for differentiating between thealternative treatments. Hence the probability ofrelapse is the same for all active treatments.

The probability of response to treatment inpatients who have failed MPH was taken from acrossover trial of IR-MPH and DEX143 where67.74% of a subgroup of patients who failed torespond to first-line therapy with MPHsubsequently responded to DEX. The definition ofresponse in the study was a ≥10-point reduction inthe hyperactivity subscale of the revised CPRS. Themodel assumes that the response to ATX inpatients who have failed MPH will be equal to that with DEX. The submission does not report theresponse rate to DEX in patients who have not failed on MPH. The relative risk of responsefor placebo compared with ATX calculated in themeta-regression was applied to the rate of responsein MPH-failed patients receiving DEX to calculatethe response rate for no treatment. In patientscontraindicated for stimulants, the response ratefor ATX was taken from a clinical trial whoseinclusion criteria included the presence of ticsdisorder or Tourette syndrome (unpublished – nofurther information provided in submission). Theresponse rate to ATX is reported to be 66.67% inthis co-diagnosed population. Table 81 shows theresponse and relapse rates used for each subgroupexamined in the model.


97


TABLE 80 Responsea rates (%) estimated in the meta-regression submitted by Eli Lilly

Treatment Stimulant-naïve patients (%) Stimulant-exposed patients (%) Overall (%)

ATX 70.51 62.17 65.08Methylphenidate 77.27 70.03 73.35Placebo 41.46 32.75 35.85

a Response is defined as ≥ 25% reduction in parent-rated ADHD-RS.

It is not clear why the response and relapse ratesfor DEX and no treatment are not applicable forsubgroups 1 and 5, as these patients are eligiblefor these treatments, which both also feature inthe treatment strategy assigned to that subgroup.It is also not made clear why the probability ofrelapse differs for sub-groups 1 and 2, ascompared with 3, 4 and 5. These and otherunexplained variations in the reported figurescannot be verified without examining theelectronic model. It was not possible to ascertainwhether the response rates extracted andestimated were monthly response rates, or whetheran overall response rate from the trials wasapplied as a monthly transition probability in themodel, which would overestimate response in themodel given that the average trial length was> 1 month.

The probabilities of adverse events,discontinuation due to adverse events,discontinuations in non-responders due toinefficacy, discontinuations due to other reasonsand the persistence of insomnia to the next cycleare all assumed equal between active treatmentoptions. The probability of experiencing anyadverse event and the probability ofdiscontinuation are taken from a post hoc analysisof data pooled from six trials comparing ATX withplacebo.63,73,74,89,142,144 One of the trials142 was asubset analysis of another89 and so to pool thesetwo trials was not appropriate. The pooled datashowed an adverse event rate of 49.2% on ATXand 36.3% on placebo, and so the submission

employs the difference of 12.9% as the monthlytransition probability. Again, it is unclear whetherthis rate was the monthly rate in the trials or, as ismore likely, the overall rate from the trials. Theaverage length of the six trials was just over9 weeks, and so applying an overall rate as amonthly transition probability would overestimatethe number of adverse events in the model. Theprobability of discontinuation due to other reasons(i.e. not inefficacy or adverse events) was estimatedto be 9.46% and may suffer from the sameproblems as just detailed.

The probability of discontinuation due toinefficacy does appear to have been calculatedcorrectly as a monthly probability of 9.89%;although the assumptions used in this calculationare not specified, we may infer that it wascalculated by assuming a constant transition rateand hence an exponential distribution. The samemay be true of the monthly probability ofdiscontinuation due to adverse events, estimatedto be 12.09%.

An indirect meta-analysis of safety data (notreferenced) estimated an RR of insomnia of 0.428for ATX compared with IR-MPH. This was appliedto an estimate of the risk of insomnia on ATX of4.7% from the six pooled placebo-controlledtrials.63,73,74,89,142,144 The subsequent rate ofinsomnia on IR-MPH of 11.0% was then net of therate of insomnia on placebo from the same studies(5.1%) as the model assumed that insomnia wasonly experienced on medication. This net rate was


98

TABLE 81 Monthly response and relapse transition probabilities by subgroup used in the cost–utility model submitted by Eli Lilly

Parameter ATX MPH (instant or extended release) DEX No treatment

Subgroup 1Response 0.7051 0.7727 NA NARelapse 0.0206 0.0206 NA NA

Subgroup 2Response 0.6667 NA NA 0.4231Relapse 0.0206 NA NA 0.0387

Subgroup 3Response 0.6774 NA 0.6774 0.3983Relapse 0.0257 NA 0.0257 0.0447

Subgroup 4Response 0.5273 NA NA 0.3478Relapse 0.0257 NA NA 0.0447

Subgroup 5Response 0.6217 0.7003 NA NARelapse 0.0257 0.0257 NA NA

NA, not applicable.

then used to calculate the probability that anadverse event experienced by a patient on IR-MPH was insomnia, and this was estimated to be46%. It is likely that the proportion of adverseevents that are insomnia would be availabledirectly from clinical trials of IR-MPH, and so theneed for this indirect calculation is not clear.Adverse events other than insomnia are given a47.3% chance of persisting to the next cycle forthe first four cycles in the model and a 100%chance of persisting thereafter. Insomnia is given a95.3% chance of persisting to the next cycle forthe first four cycles in the model and a 100%chance thereafter. These estimates are a modellingassumption made with consideration of expertopinion.

Summary of resource utilisation andcost dataThe model includes only the costs of the activemedication and excludes all other costs. Theestimated daily dose of each medication was takenfrom published sources (IMS BPI/HPAI database2003, reference not provided in submission). Thesource of the unit costs of each medicine isunclear. As ATX is flat-priced regardless of dose,the submission assumes that 90% of patients willtake one tablet per day and 10% will take two. Thedrug costs used in the model are shown inTable 82.

Summary of utility dataThe utility data were based on a study previouslypublished as a poster.145 This study obtainedutility values for 14 hypothetical health states from83 parents as proxies for their children withADHD using SG. The 18 health states weredifferentiated according to treatment received,response and side-effects, and the vignettesdescribing each state were designed to maximisethe differences between the treatment options.The results of this utility study are shown inTable 83.

DEX was not included in the utility study, and sothe model assumes that the values for treatmentwith IR-MPH are applicable. Some of these utilityvalues appear inconsistent, for example, the utilityfor a non-responder to ATX who is experiencingside effects with treatment is higher than that of aperson receiving no medication. The health statedescriptions shown to parents in the elicitationstudy are shown in Appendix 10. The maindifference between ATX and the stimulanttherapies is related to treatment coverage in theearly morning and late evening. This translatesinto a difference in utility of approximately 0.04between responders to ATX and IR-MPH. This isa relatively large difference in utility in thispopulation. As the results of this study are onlyavailable in poster format, it is not possible to


99


TABLE 82 Drug costs used in the cost–utility model submitted by Eli Lilly

Value ATX IR-MPH ER-MPH DEX

Average daily dose 1.1 pills 25.46 mg 32.75 mg 13.11 mgDaily cost (£) 2.15 0.47 1.34 0.18Monthly cost (£) 64.35 14.19 40.04 5.40

TABLE 83 Utility values used in the cost–utility model submitted by Eli Lilly

Health state N Mean SD

Treatment with ATX; responder; no side-effects 83 0.959 0.077Treatment with ATX; responder; side-effects 83 0.937 0.096Treatment with ATX; non-responder; no side-effects 83 0.902 0.133Treatment with ATX; non-responder; side-effects 83 0.886 0.148Treatment with IR-MPH; responder; no side-effects 83 0.913 0.128Treatment with IR-MPH; responder; side-effects 83 0.904 0.137Treatment with IR-MPH; non-responder; no side-effects 83 0.889 0.154Treatment with IR-MPH; non-responder; side-effects 83 0.875 0.164Treatment with ER-MPH; responder; no side-effects 83 0.930 0.107Treatment with ER-MPH; responder; side-effects 83 0.912 0.124Treatment with ER-MPH; non-responder; no side-effects 83 0.898 0.130Treatment with ER-MPH; non-responder; side-effects 83 0.884 0.143No medication; responder 23 0.880 0.133No medication; non-responder 23 0.880 0.133

assess the quality of the methodology or thesuitability of the sample of parents from whom thevaluations were elicited. The sample consisted ofparents of children with ADHD, and so thecurrent treatment of their children couldpotentially introduce bias into the results. As such,the results of this utility elicitation study should beinterpreted with caution.

Summary of cost-effectivenessThe submission states that owing to the use ofpatient-level simulation, a probabilistic sensitivityanalysis was not practical, and so the model is rundeterministically. As a result, the model cannotprovide an estimate of the uncertainty around theestimated costs and effects. The cost-effectivenessresults are shown in Table 84. Subgroups 1 and 5are associated with two sets of results because thetreatment strategies include MPH, of which thereare two formulations, extended and instantrelease.

No detail is given of the pseudo-standard errorsone would expect from a patient-level simulation,so it is not possible to judge whether enoughpatients were simulated to ensure stable estimates.The two pair-wise comparisons in subgroups 1 and5 can be reduced to one four-way comparison bycomputing the ICERs using the rules of

dominance and extended dominance (see p. 95).Using this method, the use of ER-MPH withoutATX is ruled out by extended dominance in bothsubgroups. The ICER for IR-MPH with ATXcompared with IR-MPH without ATX is £15,236 insubgroup 1 and £15,357 in subgroup 2 and theICER for ER-MPH with ATX compared with IR-MPH with ATX is £19,906 in subgroup 1 and£20,581 in subgroup 2.

Several one-way sensitivity analyses wereconducted. As expected, the results of the modelwere affected predominately by the utility valuesused.

Comments on methodologyAs the electronic model was not submitted withinthe time frame for this review, it was not possibleto verify or clarify further the informationprovided in the submission. It is also not possibleto re-analyse the model for any further sensitivityanalyses. The use of ATX is shown to be cost-effective, this result being driven by the utilityvalues employed and the assumptions regardingthe persistence of side-effects. Like the others, thesubmission did not consider long-term effects ofmedication for ADHD. Without examining theelectronic model, the need for a patient-levelsimulation rather than cohort structure is unclear.


100

TABLE 84 Cost-effectiveness results from cost–utility model submitted by Eli Lilly

Subgroup MPH QALYs: strategy Cost: strategy QALYs: strategy Cost: strategy ICER (£)with ATX with ATX (£) without ATX without ATX (£)

1 ER 0.9341 599.78 0.9140 334.07 13,2411 IR 0.9308 534.09 0.9040 125.76 15,2242 NA 0.9217 480.94 0.8800 0.00 11,5233 NA 0.9268 488.26 0.8967 39.48 14,9454 NA 0.9120 395.98 0.8800 0.00 12,3705 ER 0.9331 595.32 0.9126 316.32 13,6095 IR 0.93 531.52 0.9033 121.49 15,355

NA, not applicable.

The review of the economic evidence from theliterature and the manufacturers’ submissions

highlighted a number of potential limitations inexisting studies assessing the cost-effectiveness ofMPH, DEX and ATX in children and adolescentsdiagnosed with ADHD, including HKD. Inparticular, the review highlighted limitations inestimating treatment effectiveness and associatedutility values. In an attempt to overcome theselimitations, a new economic model was developedfor this report. The scope of this review is toidentify the most cost-effective treatment strategyfor children and adolescents with ADHD, onceone has assessed there to be a need for medicalmanagement. The scope specifically excludes thechoice between medical management and non-drug interventions, for example MPH comparedwith BT without concurrent medication.

MethodsA new model was developed to assess the cost-effectiveness of IR-MPH (3–4 hour action), ER-MPH (8 and 12 hour action), DEX and ATX forthe treatment of ADHD and HKD in children andadolescents. The model assesses the use of thesedrugs alone and in combination with abehavioural therapy element (combinationtherapy). The model is probabilistic, meaning thatrelevant input parameters are entered asprobabilistic distributions rather than pointestimates in order to represent the uncertaintyaround each point estimate.146 The followingsections of the report outline the structure of themodel, the key assumptions made and the datasources used to populate the model.

OverviewThe model has been developed to estimate costsfrom the perspective of the UK NHS and PersonalSocial Services (PSS), and health outcomes interms of QALYs. The model was developed inExcel and the evidence synthesis used to calculateclinical effectiveness parameters was conducted inWinBUGS.147 The model considers a hypotheticalcohort of children aged 6 years. For the base caseanalysis, a 1-year time horizon was selected. Asnoted in the review of the literature and themanufacturers’ submissions, this time span

excludes the long-term outcomes associated withADHD. It is known that patients can remain ontreatment for more than 1 year, but there is a lackof data about the mechanism that determines thelength of treatment. In particular, there is a lack ofdata for discriminating between treatments interms of length of treatment for responders.Similarly, there is a lack of data regarding therelative effect of the alternative treatments onother long-term outcomes, including long-termadverse events and cost offsets. These issues willbe discussed further in following sections of thereport.

In a secondary analysis, we explore a limitedextrapolation of the model beyond 1 year using anestimate of the age-dependent decline ofsymptoms with ADHD.148

Treatment strategiesThe model considers alternative sequences oftreatments. Patients who withdraw or fail on eachtreatment are assumed to move to the next in line,until they reach no treatment at the end of thesequence. It is assumed that medication is receivedas part of a comprehensive treatment plan thatincludes visits to child psychiatrists andpaediatricians. The model separately assessesthree formulations of MPH, and it is assumed thatindividuals who withdraw from one formulationwould not be offered one of the alternativeformulations, as they contain the same activeingredient.

Clearly, treatment sequences could be composedof strategies featuring one, two or three activetreatments. Conducting such an analysis allows usto estimate the incremental cost-effectiveness ofadding a second or third active treatment option.Such an analysis was conducted (see Appendix 8),and the results show that strategies with threeactive treatments are cost-effective. Subsequently,the results presented for the base case analysis andthe sensitivity analyses concern only thosetreatment strategies featuring three activetreatments. The benefit of omitting these extrastrategies from the base case analysis and othersensitivity analyses is to simplify greatly thepresentation of results (19 strategies instead of 38,without considering combination therapy). The


101


Chapter 6

Economic model

Economic model

102

reduction in the number of strategies aidsinterpretation of the model results. Theconsequence is that by removing a number ofstrategies, the decision uncertainty isunderestimated.

Each treatment sequence is composed of aformulation of MPH (IR-MPH or ER-MPH8 orER-MPH12), DEX and ATX, followed by notreatment last in sequence. This results in 18treatment strategies to be considered in themodel, all of which can be compared on the basisof pharmacotherapy only, or as part ofcombination therapy. In addition, there is a no(drug) treatment option, so there are 37 sequencesfor comparison. Table 85 provides moreinformation on the sequences for comparison.

Model structureUpon entering the model, patients begin titrationon the first-line treatment. This titration period isassumed to last for 1 month. The purpose of thistitration period is to determine the optimal dose,in terms of adverse events and response totreatment. During the titration period, patients

can withdraw from treatment if they experienceintolerable side-effects and they are then assumedto enter the titration period for the next drug insequence. The model assumes that patientstolerating treatment will continue with thattreatment if they experience a response. Thosepatients who do not respond to treatment by theend of the titration period are assumed to move to titration for the next treatment in sequence. Inthe base case analysis, it is assumed thatresponders remain on therapy and continue to beresponsive for the rest of the year. Anotherassumption of the model is that, althoughresponders may experience side-effects withtreatment, these will be relatively minor andtolerable; hence they will not develop intolerableside-effects beyond the titration period. Thisassumption was considered reasonable followingconsultation with a clinical expert. In summary, inthe base case analysis, patients only withdrawowing to side-effects during the titration period,and patients are moved to the next treatment insequence if they fail to respond by the end of thetitration period. Figure 22 provides a simplesummary of the model structure.

TABLE 85 Treatment sequences compared in economic model

Treatment sequences 1 2 3 4 5 6

1st line MPH MPH ATX ATX DEX DEX2nd line ATX DEX MPH DEX MPH ATX3rd line DEX ATX DEX MPH ATX MPH4th line No treatment No treatment No treatment No treatment No treatment No treatment

×3 to represent each formulation of MPH = 18×2 to include combination therapy = 36+ no treatment = 37

Continue on treatment

Drug 1

Tolerate

Intolerable side-effects

Response

No response

Drug 2

Drug 2

FIGURE 22 Representation of economic model structure. Once all treatment options are exhausted, patients are assumed to remainnon-responders on no treatment.


103


ExtrapolationA secondary analysis extends the time horizonfrom 1 year to when the cohort reaches 18 years ofage. In this secondary analysis, it is assumed thatpatients on active medication are given an annualdrug-free period of 2 weeks to assess the ongoingneed for medication. A proportion of patients,whose symptoms are found to be in remission, donot resume medication following this 2-week drug-free period. The model also incorporates theproportion of patients on no treatment whoexperience remission from symptoms, and henceare no longer defined as non-responders. Themodel makes the simple assumption that patientsnot in remission will return to their pre-drug-freeperiod medication, that this medication will betolerated and that the patient will continue torespond on medication as before. Patients inremission are assumed to be identical toresponders in terms of utility and non-drug healthexpenditure.

In this extended analysis, costs are discounted atan annual rate of 6%, and health benefits arediscounted at an annual rate of 1.5%, inaccordance with NICE guidance.

Clinical effectivenessAs identified in the clinical effectiveness review inChapter 4, there is huge variation in theinstruments used and the outcomes reportedregarding the clinical effectiveness of alternativetreatment options. This reflects the lack ofconsensus on what constitutes successful treatmentof ADHD in the absence of a biological markerand hampers between-study comparisons. Theclinical effectiveness review addressed this issue ofdisparity by identifying the most valid and reliablemeasure in consultation with a clinical expert andadopting a standard method of presentation. ThusChapter 4 presents an overview of the change inmean score on the Conners’ TeachersHyperactivity Subscale (CTRS-H) and Conners’Parent Hyperactivity Subscale (CPRS-H). Althoughthis measure facilitates comparison across a largenumber of studies, there are limitations in its usein a decision-analytic model. As noted in theliterature review in Chapter 4, if this measure isused, a gain of one point on the scale is valued thesame, regardless of where one begins on that scale,so the relative value of different effect sizes is notreadily interpretable. In addition, no publishedstudies were available to provide a link betweenmean CTRS-H or CPRS-H score and utility data.In order to identify the most optimal treatmentstrategy in a decision-analytic model, one must beable to value the differences in outcome, and this

is not currently possible with mean CTRS-H orCPRS-H score.

Alongside mean scores on various rating scales, aproportion of studies also report response rates totreatment. These indicate the percentage ofchildren who reach a specified level ofimprovement. Response rates would be preferredfrom an economic evaluation perspective for tworeasons. First, they explicitly identify a clinicallymeaningful change on the rating scale used todetermine effectiveness. Second, the utility datacurrently available for patients with ADHD pertainto the states of ‘responder’ and ‘non-responder’.Therefore, an economic model based on responsecan estimate a cost per QALY, which is moreinterpretable and generalisable than a cost perpoint gain in CTRS-H or CPRS-H. On the otherhand, there are limitations with the use ofresponse rate as a measure of clinical effectiveness,primarily owing to the lack of an agreed definitionof response to treatment for ADHD. This makescomparisons between studies difficult as they mayuse different definitions.

Hence the measure of clinical effectiveness in theeconomic model is response rate to treatment.This reflects the approach used in the existingeconomic literature and the manufacturers’submissions, but must be interpreted according tothe caveats used in calculating response rates. Themodel considers a number of different sets ofresponse rates, each estimated from the clinicalstudies with different inclusion criteria. The firstset of response rates was estimated using the mostcommon definition of response in the includedstudies, namely a score of 1 or 2 (much improvedor improved) on the clinician-rated Clinical GlobalImpression improvement subscale (CGI-I). Theserates are used in the base case analysis. This allowsa comparison of all relevant treatment strategieswithout making assumptions about thecomparability of different definitions of response.The second definition of response is a score of 1or 2 on the clinician-rated Clinical GlobalImpression severity subscale (CGI-S). There wasno estimate available for the response rate to DEXusing this definition. A third group of trialsdefined response as a reduction of ≥ 25% on theparent-rated ADHD-RS. Measures of responseaccording to this criterion were only available forplacebo, ATX and ER-MPH12. Hence for thisthird definition of response, further modelling wasrequired in order to compare all relevanttreatment strategies [details follow in the section‘Sensitivity to structural assumption regardingMPH’ (p. 117)]. Finally, estimates of response

defined as a score of ≤1 on the SNAP-IV scalewere used in the MTA trial. The problems withinterpreting the results of this trial have beendescribed in more detail in Chapter 4, butnevertheless it is an important trial in this diseasearea. As such, it was felt appropriate to include ascenario using response rates defined according tothe definition used in the MTA trial.

The base case analysis uses a consistent definitionof response to compare all relevant options.Because this excluded a number of trials, andhence reduced the amount of available data,sensitivity analyses were conducted by relaxing thedefinition of response to include more trials andto assess the impact of different definitions ofresponse on the estimates of cost-effectiveness.

Table 86 displays the source trials used to estimateresponse rate in the base case analysis. Furtherdetail about each trial has been provided inChapter 3, where the trial concerned was includedin the effectiveness review. A number of studiesexcluded from the effectiveness review, for reasonsof data presentation, were nevertheless found toprovide information on response rate. Thesestudies were therefore included in the calculationof response rate for the cost-effectiveness analysis.Further details of these excluded studies are givenin Appendix 3. All of the trials were set in NorthAmerica (five in the USA and one in Canada90),and most recruited children aged between 6 and12 years (one study recruited from age 6 to16 years59). Four used the DSM-IV diagnostic

criteria, and the remaining two used otherdiagnostic interviews.65,83 The average daily doseof IR-MPH and ER-MPH12 varied between thetrials, and this is not reflected in the calculation ofresponse rates. It is important to note that in theclinical trials, patients were titrated to the ‘best’dose, which reflects our model structure, but doesallow average dose to differ between trials. Threeof the trials excluded subjects who were knownnon-responders to stimulant therapy,59,90,99 andthis is also not reflected in the calculation ofresponse rates. This heterogeneity between trialsmust be borne in mind when interpreting theresults of the model.

Table 86 excludes one trial84 [Confidentialinformation removed]. An important assumptionin the base case model is that the treatment effectsare independent of treatments previously received.In other words, the response rate to IR-MPH isthe same if it is received as first-line therapy aswhen it is received following failure on DEX orATX. This assumption was necessary as data werenot available to calculate response ratesconditional on specified previous treatments.

Ideally, the relative treatment effects of notreatment, IR-MPH, ER-MPH8, ER-MPH12, ATXand DEX would be estimated in a single, directhead-to-head RCT. However, such a trial does notexist, and instead we have a number of trialsassessing the treatment effects of different subsetsof the full set of relevant comparators. Clearly, theabsolute response rates differ by trial, but the

Economic model

104

TABLE 86 Response rates used in the base case analysis: response defined as score of 1 or 2 on CGI-I

Trial Treatment Responders (%) No. in group

Sharp, 1999149a IR-MPH 26 (81) 32DEX 27 (84) 32Placebo 5 (16) 32

Greenhill, 200259 ER-MPH8 125 (81) 154Placebo 78 (50) 156

Kemner, 200499 ER-MPH12 583 (69) 850ATX 250 (53) 473

Steele, 200490 ER-MPH12 58 (83) 70IR-MPH 45 (62) 73

Pliszka, 200083 IR-MPH 13 (65) 20Adderall 18 (90) 20Placebo 5 (28) 18

Klein, 199765 IR-MPH + BT 28 (97) 29IR-MPH 23 (79) 29Placebo + BT 14 (50) 28

a Not currently reviewed in Chapter 4.

economic model rests on the assumption that therelative treatment effects will be the same acrosstrials. This is a common assumption used in anyroutine meta-analysis. As Table 86 shows, there isnot a common comparator between all the trials.Therefore, in order to pool the data, a mixedtreatment comparison (MTC) model wasused.150,151 An MTC provides an explicit analyticalframework to combine all the evidencesimultaneously in order to estimate a set ofresponse rates for the economic model. Theframework requires few additional assumptionsover those routinely made in simple meta-analyses.

Mixed treatment comparisonThis section provides a brief overview of theprinciples underlying an MTC. Suppose there arethree clinical trials comparing three treatments ofinterest, A, B and C. Each clinical trial assesses adifferent pair-wise comparison, AB, AC and BC. Asimplistic method might be to compare directtreatment effects against a common baseline, forexample A, and merely discard the informationprovided by the BC comparison. This is inaccordance with the view that indirect comparisonsof A and C based on comparisons of AB and BCrepresent a lower level of evidence. However, it isevident that, based on the principle of transitivity,if the true differences between AB, AC and BC are(on the appropriate scale) �AB, �AC and �BC, thenwe expect

�AC = �AB + �BC

Hence the information provided by the BCcomparison need not be discarded and can beused to update the direct comparisons of AB andAC. Higgins and Whitehead150 have shown howthe use of ‘external’ AB and BC evidence cansubstantially reduce uncertainty about an ACcomparison of primary interest. For example, withreference to this report, the estimate of the effectof ER-MPH12 compared with IR-MPH from Steeleand colleagues90 can be combined with anestimate of the effect of IR-MPH relative toplacebo in order to obtain an estimated relativetreatment effect for ER-MPH12 compared withplacebo. This estimate would be otherwiseunavailable in this dataset.

Based on these general principles, a Bayesianmeta-analysis of the proportion of respondersassuming random treatment effects was conductedusing Markov Chain Monte Carlo (MCMC)implemented in WinBUGS.147 The WinBUGSmodel used to estimate the proportions ofresponders assumes a regression-like structure,

with the logit of the proportion of responders forany treatment k, depending on a ‘baseline’ term �iin trial i, i = 1, 2, …, 6, and a treatment effect �i

k.The trial-specific baselines are drawn from acommon random normal distribution, whoseparameters must be estimated from the data, givenvague priors. Formally, this can be expressed as

logit(�ik) = �i + �i

k

�i ~ N(�,�a) � ~ N(0,0.0001), sa2 ~ uniform(0,10)

�a = 1/sa2

The trial-specific treatment effects �ik are assumed

to be drawn from a common random normaldistribution around the ‘true’ treatment effect �k.A binomial likelihood is assumed from theavailable data points:

�ik ~ N(�k,�b) �k ~ N(0,0.0001),

sb2 ~ uniform(0,10)�b = 1/sb2

rik ~ Bin(pi

k, nik),

where k denotes all treatment indices in study i, ridenotes the observed number of responses and nidenotes the total number in the group.

The WinBUGS code for the model is reported inAppendix 9. The code represents an extension ofthe Higgins and Whitehead 1996 model150 tomore general MTC structures. The output fromthe model incorporates the uncertainty around theestimated response rates and also any correlationbetween treatments. However, for simplicity, three-arm trials were treated as two two-arm trials with acommon comparator.

Combination therapyAs noted in the clinical effectiveness review inChapter 4, where behavioural therapies are usedin combination with pharmacotherapy, they varybetween trials. As such, there is no one ADHDcombination therapy that we could assess incomparison with drug monotherapy. However,some of the trials did show a (non-statisticallysignificant) favourable effect of combinationtherapy compared with drug monotherapy. Thedata do not allow us to compare a single consistentbehavioural therapy component in combinationwith all of the relevant treatment comparators.Instead, we can only estimate the relative increasein response rate associated with IR-MPH and BTcompared with IR-MPH alone.65,133 Hencecombination therapy was considered in asecondary analysis, where the relative increase inresponse rate for IR-MPH with BT compared with


105


IR-MPH was assumed to apply to all drugtreatments. However, owing to the variation inprogrammes, this analysis can tell us little aboutcombination therapy in practice in the UK.

Adverse eventsThe model assumes that patients who fail totolerate treatment during the titration period willdiscontinue that treatment if dose modificationdoes not address the problem, and that any side-effects will dissipate upon treatment cessation.Hence the main consequence of adverse events inthe model is discontinuation of treatment, andintolerable side-effects are not associated with anadditional utility decrement. The cost and utilityestimates for responders (i.e. patients continuingon treatment) take account of the minor side-effects that commonly accompany treatment forADHD. However, it must be noted that they donot discriminate between treatments in thisrespect. As noted in the clinical effectivenessreview in Chapter 4, there is little trial evidence todiscriminate between the active therapies in termsof tolerable side-effects. Of the events assessed,insomnia may be more common with stimulanttherapy in comparison with ATX, so this must beconsidered when interpreting the model results.Another important factor is the lack of dataregarding long-term adverse events beyond theperiod observed in the clinical trials.

In the clinical trials, patients withdrew fromtreatment for many reasons, including lack ofefficacy and intolerable adverse events. Thereasons given differed between trials, and in somecases the reason for withdrawal was not stated. Asa result, the withdrawal rates for the model werecalculated to include all reported withdrawals,regardless of the reason given. This approach waschosen to maintain consistency between trials withdiffering definitions of withdrawal, and it is also aconservative approach from the point of view ofNICE. This is because a higher rate of withdrawalwill increase the cost-effectiveness ratios associatedwith active treatment, and so a treatment thatappears cost-effective using this broad definitionof withdrawal would appear even more cost-effective under a more precise definition.However, it must be noted that in four of the 10trials shown in Table 87,63,83,90,94 a proportion ofwithdrawals were attributed to non-response. Thisdemonstrates that although the approach takenmay be conservative and consistent in calculatingwithdrawal rates, it may include some doublecounting of non-responders. The degree of doublecounting is difficult to quantify because, as notedin the clinical effectiveness review in Chapter 4,

none of the four trials calculated response in anITT analysis.

The probability of withdrawal was calculated in thesame way as the response rates, hence the samemodel applies, as shown in Appendix 9. Owing tothe lack of reported data on withdrawal, one set ofrates was calculated including all studies thatprovided an estimate of response rate, regardlessof definition of response. Table 87 details the dataused to calculate withdrawal rates for theeconomic model. These data apply to all analyses,regardless of definition of response.

In the secondary analysis including combinationtherapy, it was assumed that withdrawal wouldonly be induced by pharmacotherapy, and so thesame withdrawal rates are applied as to drugmonotherapy.

As noted earlier in this chapter, few data areavailable on adverse events associated with long-term use of pharmacotherapy for ADHD.Therefore, the adverse events reflected in themodel are limited to those observed during thetreatment phase of the included clinical trials.

ComplianceCompliance can be thought of as the ability ofpatients to take the required number of doses ofmedication, or of the ability of patients to takepills within the correct time frame; these are dose-taking and dose-timing compliance, respectively.This section of the report refers to dose-takingcompliance. If patients take fewer pills thanprescribed by their doctor, they will be receiving alower dose of medication. In their submissions,Janssen-Cilag and Celltech both put forward theargument that compliance to a midday dose oftreatment will be lower than compliance to anearly morning dose. They argue that double-blind,double-dummy trials do not capture the effect ofimproved compliance to the once-daily extended-release formulations of MPH as both comparatorgroups must take more than one pill per day (theextra pills being placebo in the ER-MPH groups).In explanation, in a double-blind, double-dummytrial, patients are blinded to treatment and receivethe same number of doses of medication per dayin each arm. Dummy pills (placebo) are used toprevent patients identifying the treatment towhich they have been allocated by counting thenumber of doses they receive. This means that in adouble-blind, double-dummy trial of IR-MPH TIDversus ER-MPH12, patients in both arms wouldreceive three pills each day. In the IR-MPH arm,all the pills would contain an active dose of IR-

Economic model

106

MPH. In the ER-MPH12 arm, only the first pillwould contain active medication and the secondand third doses each day would consist of placebo.In our base case analysis, it is assumed that the trialdata adequately capture the effect of compliance onresponse to treatment. In other words, the dailydose received by patients taking ER-MPH wouldnot be reduced if patients comply poorly with alunchtime dose (as this is placebo in this group).Therefore, the improved effectiveness, due to bettercompliance to the morning dose, should beadequately captured. The clinical trials used toestimate response rate include open-label trials, andthese should capture any effect of compliance onclinical outcomes. In an open-label trial, patientsare not blind to treatment and may not receive thesame number of doses of medication per day. In anopen-label trial of IR-MPH TID versus ER-MPH12,patients in the IR-MPH arm would receive threepills per day, whereas patients in the ER-MPH12arm would receive one pill per day. The open-labelnature of the trial removes the need to includedummy, or placebo, pills.

For the argument that double-blind, double-dummy trials do not capture the effects of

compliance to hold, the assumption must be thattaking three pills per day has a deleterious effecton compliance to the morning dose, rather thanthat compliance to a midday dose is lower than toa morning dose. A systematic review of complianceto different dosing regimens in a range of diseaseareas has shown that compliance does appear tofall as the number of doses per day increases.152

However, these data on average compliance perday cannot resolve the issue of whether patients’reduced compliance to twice- or three-times dailyschedules is the result of taking fewer pills at eachdose timing or as a result of taking fewer pills atthe later dose timings (i.e. lunchtime andevening). Also, none of the studies in the SR ofcompliance looked specifically at ADHD, andinstead the emphasis is on medication for adults.The exploration of the effects of non-compliancewould involve a number of assumptions: theassumption that RCT data capture none of theeffects of compliance; the application of a selectedestimate of compliance from a source outside ofthe clinical trials; and an assumption regardingthe distribution of reduced compliance betweenmorning, lunchtime and evening doses ofmedication. It was felt that these modelling


107


TABLE 87 Data used in calculating withdrawal rates for the economic model

Trial Treatment Withdrawals (%) No. in group

Sharp, 1999149a IR-MPH 1 (3) 32DEX 0 (0) 32Placebo 0 (0) 32


Kemner, 200499 ER-MPH12 41 (5) 850ATX 26 (5) 473

Steele, 200490 ER-MPH12 12 (16) 73IR-MPH 12 (16) 74

Pliszka, 200083 IR-MPH 1 (5) 20Adderall 2 (10) 20Placebo 2 (11) 18

Klein, 199765 IR-MPH + BT 0 (0) 29IR-MPH 1 (3) 31Placebo + BT 2 (7) 29

Kelsey, 200463 ATX 26 (20) 133Placebo 17 (27) 64

Michelson, 200274 ATX 12 (14) 85Placebo 11 (13) 86

Weiss, 200494 ATX 17 (17) 101Placebo 4 (8) 52

Spencer, 200289 ATX 8 (6) 129Placebo 7 (6) 124

a Not currently reviewed in Chapter 4.

assumptions would not be reasonable given thelack of available data, which would render theresults of any sensitivity analysis aroundcompliance uninformative to decision-makers.

Resource utilisation and cost dataAs identified in the review of existing economicevaluations, there are few observed data on theresource use associated with ADHD. In theabsence of readily available data, it was necessaryto base the resource use in the model on estimatesobtained from expert opinion. Hence the resourceuse in the model is based on that used in thesubmission by Janssen-Cilag, reviewed inChapter 5. The study from which these estimateswere obtained129 has been reviewed in more detailin Chapter 5. The estimates of resource use wereobtained from a Delphi panel, in which the UK-based experts were asked to specify their drugtreatment programmes according to treatmentused, response status and presence of adverseevents. These data are updated with current, UK-specific price data (NHS reference costs 2003).153

The resource use includes visits to psychiatristsand paediatricians to reflect a more comprehensivetreatment programme than drug therapy alone.Current guidance1 recommends IR-MPH as partof a comprehensive treatment programme. Thedetails of such a programme are not defined, butit does not need to include specific psychologicaltreatment, that is, what we refer to as BT.

The uncertainty around the estimated resource usewas characterised using a gamma distribution.Table 88 shows the resource use and unit costsemployed in the economic model.

The impact of using alternative estimates ofresource use for children and adolescents withADHD was assessed in a sensitivity analysis. Asnoted earlier in this chapter, the review found nodata on resource use associated with long-term useof ADHD. In order to extrapolate the modelbeyond 1 year, it was assumed that patients whocome off therapy due to remission of symptomswould incur the same non-drug resource use asresponders. This acknowledges the fact thatpatients in remission of symptoms are not ‘cured’of ADHD, but likely overestimates the costs in thelong run.

The average dose for each active medication wastaken from the trials used in calculating responserates. Although these doses may not reflect exactlycurrent UK practice, they are the doses at whichthe effectiveness data were obtained. As we onlyhave clinical trial data for treatment effectiveness,it is not possible to determine the effectiveness atthe current average UK dose. Hence the drugcosts are consistent with the effectiveness dataused in the model. The drug prices were obtainedfrom published UK pricing lists,29 where available.ER-MPH8 is not currently priced in the UK, sothe model employs the prices reported in themanufacturer’s submission.154 Table 89 displays thedose and unit cost data employed in the economicmodel. The unit cost data for IR-MPH are basedon the generic formulation (note that the cost of10-mg Ritalin is the same as that for 10-mggeneric IR-MPH).

The data regarding average drug dose are entereddeterministically, which means that the model is

Economic model

108

TABLE 88 Resource use and unit cost data used to populate the economic model

Item Average per year Lower CI limit Upper CI limit Unit cost (£)

Responders:ConsultationsPsychiatrist 3.5 2.3 4.7 109.5Paediatrician 2.25 0.6 3.7 188GP 3 1.3 4.7 24TestsBlood test 0.05 0 0.1 7

Non-responders:ConsultationsPsychiatrist 5.75 4.3 7.4 109.5Paediatrician 2.5 0.9 3.7 188GP 2.75 0.6 4.7 24TestsBlood test 0.35 0.06 0.76 7ECG 0.33 0 0.79 29.48EEG 0.43 0.06 0.85 111.93Allergy test 0.5 0.09 0.91 67

not fully probabilistic. The average daily drugdoses correspond to the moderate/high dose trialsin the clinical effectiveness review in Chapter 4.

Utility dataThe review of the literature in Chapter 5 illustratedthat there are few sources of data on utility forchildren and adolescents with ADHD. Two of themanufacturers’ submissions used utility valuesobtained using an SG technique from parents ofchildren with ADHD, providing proxy valuationsfor their children. These utility estimates wereavailable in the public domain as posterpresentations. The first set of utility values136,155

provided an estimate of the utility of a responderto treatment (mean 0.837, standard error 0.039)and a non-responder to treatment for ADHD(mean 0.773, standard error 0.039), independentof the treatment being received. These values areused in the base case analysis. The second set ofavailable utility values145,156 provided estimates ofthe utility of responders and non-responders, withand without side-effects, that were treatment-specific. There were some concerns over thevalidity of these estimates, as detailed in Chapter 5,so these are employed in a sensitivity analysis.None of the available utility values were obtainedwith the use of a generic health valuation measurevalued with public preferences, as recommended inguidance from NICE.157 This is a commonproblem when assessing QoL in children.

The uncertainty around these estimated utilityvalues was characterised using a beta distribution,as it was felt reasonable to assume that the valueswould not drop below zero.

ResultsResponse and withdrawal rates in thebase case analysisIn the base case analysis, the definition ofresponse was defined as a score of 1 or 2 on theCGI-I. This restricted the number of included

trials to six of the 65 identified in the clinicaleffectiveness review in Chapter 4. The estimatedresponse rates were subject to large uncertainty,which is unsurprising given the small size of someof the included studies, the restricted amount ofdata available and the heterogeneity betweentrials. The output of the WinBUGS model isshown in Table 90.

The estimated response rates are all in theexpected direction given the trial evidence.However, it must be noted that the volume of trialdata is small and varies between treatments. Dataon the effectiveness of IR-MPH were available infour trials (154 patients), ER-MPH8 in one trial(154 patients), ER-MPH12 in two trials (920patients), ATX in one trial (473 patients) and DEXin one trial (32 patients). The uncertainty in thecalculated response rates incorporates the size ofthe evidence base for each treatment.

Cost-effectiveness results for the basecase analysisThe base case analysis assessed the cost-effectiveness of alternative treatment strategies,comprising drug monotherapies followed by notreatment. Nineteen relevant strategies werecompared, including a no treatment option (seeTable 85), over a time horizon of 1 year. Theresults from this analysis identified a dominanttreatment strategy, number 13, which wasassociated with the lowest costs and the highestQALY gains relative to other comparators.However, the difference in QALY gains betweenthe active treatment strategies was very small. Thisis unsurprising given the uncertainty surroundingthe relative clinical effectiveness of the activetreatments. Also, the loss of QoL by trying anineffective treatment before a relatively moreeffective one will endure for only 1 month beforenon-responders move to the next treatment insequence. Table 91 shows the results from the basecase economic analysis, employing the responseand withdrawal rates shown in Table 90. Theresults should therefore be interpreted with


109


TABLE 89 Average dose and unit cost used in economic model: IR-MPH, ER-MPH8, ER-MPH12, ATX and DEX

Treatment Average dose per day Average dose per day Cost per day titration Cost per day after during titration (mg) following titration (mg) period (£) titration period (£)

IR-MPH 22 39 0.36 0.64ER-MPH8 25 41 1.27 1.58ER-MPH12 27 35 1.33 1.76ATX 28 45 2.15 2.19DEX 14 22 0.19 0.42

reference to the caveats used in calculating thesetreatment effects.

In order to display the decision uncertainty,Figure 23 shows the CEACs137 for all 19 strategiescompared. However, the CEACs cannot be used todetermine the optimal treatment strategy. Theoptimal treatment strategy is the one with thehighest expected net benefit at a given value ofwillingness to pay per QALY. If the distribution ofincremental net benefits is skewed, the optimaltreatment strategy may not have the highestprobability of being cost-effective.137 The results ofthis analysis showed that strategy 13 (shown inbold in Table 91) had the highest expected netbenefit over the full range of values of willingnessto pay per QALY considered (£0 to £60,000).

At any value of willingness to pay per QALY withinthe range explored, strategy 13 is the mostoptimal treatment strategy. If the societalwillingness to pay were £30,000 per additionalQALY, strategy 13 has a 31% probability of being

the most cost-effective (see Appendix 8). However,the CEAC shown here includes only a subset ofrelevant treatment strategies: those featuring threeactive treatments. Hence the probability thatstrategy 13 is optimal shown in Figure 23(60%) ishigher than the true probability because half asmany treatment strategies are compared. Theresult reflects the fact that the trial data providelittle evidence for discriminating between thealternative medications for ADHD in terms ofeffectiveness, but that DEX, followed by IR-MPH,are much cheaper than the other comparators.

The manufacturer’s submissions consistentlyconsidered DEX as second-line therapy and neveras a first-line treatment option. The licence forDEX specifies its use for refractory hyperkineticstates in children. Hence for patients who havepreviously failed BT or remedial measures, DEXcould be considered as first-line drug therapy.However, if the term refractory refers to previousmedical management, then DEX may only besuitable as second-line therapy. If this assumption

Economic model

110

TABLE 90 Response and withdrawal rates estimated in MTC model in WinBUGS: response defined as score of 1 or 2 on CGI-I

Treatment Response rate (SD) Withdrawal rate (SD)

Placebo 0.28 (0.04) 0.11 (0.02)IR-MPH 0.68 (0.30) 0.09 (0.05)ER-MPH8 0.57 (0.33) 0.08 (0.06)ER-MPH12 0.75 (0.32) 0.12 (0.04)ATX 0.67 (0.37) 0.11 (0.06)DEX 0.75 (0.32) 0.02 (0.05)

TABLE 91 Results of the base case analysis of the economic model

Strategy Order of treatments received Cost QALYs

1 IR-MPH – ATX – DEX – No treatment 1,233 0.82792 ER-MPH8 – ATX – DEX – No treatment 1,470 0.82733 ER-MPH12 – ATX – DEX – No treatment 1,479 0.82784 ATX – IR-MPH – DEX – No treatment 1,480 0.82785 ATX – ER-MPH8 – DEX – No treatment 1,550 0.82776 ATX – ER-MPH12 – DEX – No treatment 1,563 0.82747 IR-MPH – DEX – ATX – No treatment 1,140 0.82838 ER-MPH8 – DEX – ATX – No treatment 1,336 0.82779 ER-MPH12 – DEX – ATX – No treatment 1,410 0.8284

10 ATX – DEX – IR-MPH – No treatment 1,466 0.828111 ATX – DEX – ER-MPH8 – No treatment 1,485 0.828112 ATX – DEX – ER-MPH12 – No treatment 1,488 0.827813 DEX – IR-MPH – ATX – No treatment 1,098 0.828914 DEX – ER-MPH8 – ATX – No treatment 1,157 0.828715 DEX – ER-MPH12 – ATX – No treatment 1,159 0.828716 DEX – ATX – IR-MPH – No treatment 1,158 0.828817 DEX– ATX – ER-MPH8 – No treatment 1,177 0.828818 DEX– ATX – ER-MPH12 – No treatment 1,180 0.828519 No treatment 1,223 0.7727

is reasonable, then strategies 13–18 are no longerrelevant, and the optimal treatment strategy isthen number 7 (first-line IR-MPH, second-lineDEX, third-line ATX). Strategy 7 does notdominate all the remaining alternatives. Strategy 9(first-line ER-MPH12, second-line DEX, third-lineATX) is more costly and more effective, with a costper QALY gained of £5,595,829 compared withstrategy 7. If society were willing to pay £30,000per additional QALY, strategy 7 would have an84% probability of being most cost-effective.Again, this probability only includes uncertaintybetween treatment strategies featuring three activetreatments. There is less uncertainty than whenDEX is considered suitable as first-line therapybecause there are six fewer treatment strategiesbeing compared (13 compared with 19).

Medication as part of combinationtherapyThe trials used to estimate response rates for thebase case included a trial comparing IR-MPHalone with IR-MPH combined with BT. This trialwas used to estimate the relative change inresponse rate of adding BT to medication. Thisrelative effect was then applied to all of the drugtreatments, yielding the same relative change inresponse rate across all treatments. Thissimplification was necessary as trials were notavailable assessing combination therapy with eachof the relevant drug comparators. Also, had suchtrials been available, it is likely that the

behavioural component would differ betweentrials, as highlighted in the section ‘Combinationtherapy’ (p. 105). Hence this sensitivity analysis represents a simplistic analysis of the cost-effectiveness of drug therapy in combinationwith BT.

The results of this sensitivity analysis indicate thatstrategy 13 remains the optimal treatmentstrategy. However, strategy 13 does not dominatethe strategies that include combination therapy.This is because the drugs in combination with BTare marginally more effective than when givenalone. By calculating the ICERs, according to therules of dominance and extended dominance, theonly alternative not ruled out is strategy 36(combination therapy with first-line DEX, second-line ATX, third-line ER-MPH8). The cost perQALY gained with strategy 36 compared withstrategy 13 is £1,241,570, hence combinationtherapy does not appear cost-effective in thissensitivity analysis. Increasing the number ofstrategies from 19 to 37 increases the decisionuncertainty. If society were willing to pay £30,000per additional QALY, strategy 13 has a 40%probability of being the optimal strategy(compared with 60% in the base case excludingcombination therapy).

If DEX is not suitable as a first-line therapy, thenstrategy 7 (first-line IR-MPH, second-line DEX,third-line ATX) is the optimal treatment strategy.


111


0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0 10,000 20,000 30,000 40,000 50,000 60,000Willingness to pay per QALY (£)

Prob

abili

ty c

ost-

effe

ctiv

e

12345678910111213141516171819

Strategy 13

FIGURE 23 CEACs for 19 strategies compared in the base case analysis

Sensitivity to estimated utility valuesThe base case analysis employs estimates of theutility associated with response to treatment andnon-response to treatment that are independent ofthe treatment received. A responder to ATXtherefore receives the same utility value as aresponder to MPH or DEX. The submission by EliLilly utilised a different set of utility estimatesderived using SG methodology. These estimatesvalued response (and non-response) to treatmentdependent on the medication received. Henceseparate values were available for response (andnon-response) to ATX, IR-MPH and ER-MPH,also separated by the presence or absence oftreatment side-effects. The review of the companysubmissions highlighted some concerns about thevalidity of these estimates, particularly the factthat the utility of a non-responder without side-effects differs between treatments. For example,the utility associated with non-response to ATX,without side-effects, is estimated to be 0.902,which compares with an estimated utility of 0.880associated with non-response and no medication.A difference in utility of 0.022 is relatively large inthis population, particularly between health stateswith identical characteristics.

A sensitivity analysis was conducted using thesealternative estimates of utility. The reason for thedifferences in utility of non-response by treatment(including no treatment) is unclear, so thesensitivity analysis uses the utility of non-responseassociated with no medication. Our model doesnot separate responders into those with side-effects and those without, so we conducted thesensitivity analysis including the utility of responsewithout side-effects. Table 92 shows the utilityvalues used in this sensitivity analysis.

The health state descriptions used to obtain thesevaluations are shown in Appendix 10. Thesevignettes were designed to maximise thedifferences between treatments. The results of thissensitivity analysis rely on the validity of thesehealth state descriptions. No estimate was

available for DEX, so the utility associated with IR-MPH was applied to patients responding to DEX,in accordance with the assumption made in thesubmission by Eli Lilly.

The results of the sensitivity analysis are shown inTable 93. Strategy 13 remained the cheapeststrategy, but it no longer dominated the otherstrategies. By calculating the ICERs, according tothe rules of dominance and extended dominance(p. 95), we see that strategies 5, 10, 11 and 16 arenot ruled out by dominance or extendeddominance. Strategies 5, 10 and 11 all featureATX as first-line therapy. This is unsurprisinggiven that a response to ATX is associated with autility gain of 0.046 over a response to IR-MPH(and DEX). Response to ATX is associated with autility gain of 0.079 compared with non-responsewith no treatment, whereas response to IR-MPHentails a gain of only 0.033 over non-responsewith no treatment. For comparison, in the basecase analysis, response is associated with a utilitygain of 0.064 compared with non-response.

Figure 24 shows the cost-effectiveness acceptabilityfrontier for the optimal strategies in this sensitivityanalysis. If society were willing to pay £30,000 peradditional QALY, strategy 11 is the optimalstrategy with a 3% probability of being the optimalstrategy.

The discontinuities in the frontier in Figure 24illustrate that the distribution of incremental netbenefit is skewed. Strategies 13 and 6 have ahigher probability of being cost-effective thanstrategies 5, 10, 11 and 16 for values of willingnessto pay per QALY >£11,000, but they do not havethe highest expected net benefit.

Co-morbid conditionsThe base case analysis does not include theadditional costs of the common co-morbidconditions of CD and ODD. Estimates of theadditional cost of common co-morbid conditionswere available from the same source that provided

Economic model

112

TABLE 92 Alternative utility values used in sensitivity analysis of the economic model

Health state Utility value (SE)

Responder to ATX, no side-effects 0.959 (0.077)Responder to IR-MPH, no side-effects 0.913 (0.128)Responder to ER-MPH, no side-effects 0.930 (0.107)Non-responder, no medication 0.880 (0.133)

SE, standard error.


113


TABLE 93 Results of sensitivity analysis employing treatment-specific utility values

Strategy Order of treatments received Cost (£) QALYs Cost per QALY (£)(compared with)

1 IR-MPH – ATX – DEX – no treatment 1,237 0.9141 ED2 ER-MPH8 – ATX – DEX – no treatment 1,474 0.9264 D3 ER-MPH12 – ATX – DEX – no treatment 1,481 0.9228 D4 ATX – IR-MPH – DEX – no treatment 1,484 0.9329 D5 ATX – ER-MPH8 – DEX – no treatment 1,554 0.9361 31,107 (vs 11)6 ATX – ER-MPH12 – DEX – no treatment 1,566 0.9359 D7 IR-MPH – DEX – ATX – no treatment 1,144 0.9085 D8 ER-MPH8 – DEX – ATX – no treatment 1,340 0.9176 ED9 ER-MPH12 – DEX – ATX – no treatment 1,413 0.9190 ED

10 ATX – DEX – IR-MPH – no treatment 1,470 0.9330 15,448 (vs 16)11 ATX – DEX – ER-MPH8 – no treatment 1,489 0.9340 20,173 (vs 10)12 ATX – DEX – ER-MPH12 – no treatment 1,491 0.9287 D13 DEX – IR-MPH – ATX – no treatment 1,103 0.9088 –14 DEX – ER-MPH8 – ATX – no treatment 1,161 0.9117 D15 DEX – ER-MPH12 – ATX – no treatment 1,163 0.9107 D16 DEX – ATX – IR-MPH – no treatment 1,162 0.9131 13,539 (vs 13)17 DEX– ATX – ER-MPH8 – no treatment 1,181 0.9140 ED18 DEX– ATX – ER-MPH12 – no treatment 1,183 0.9138 D19 No treatment 1,228 0.8780 D

D, ruled out by dominance; ED, ruled out by extended dominance.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0


Prob

ablii

ty c

ost-

effe

ctiv

e

131610115

Strategy 13

Strategy 16

Strategy 10Strategy 11

Strategy 5

FIGURE 24 Cost-effectiveness acceptability frontier showing the optimal strategies in a sensitivity analysis employing treatment-specific utility values

the non-drug costs used in the base case analysis.Table 94 presents the additional cost of these co-morbid conditions.

When the model is re-analysed to include theseadditional costs, strategy 13 remains the dominantstrategy, with a cost of £1491 and 0.8281 QALYsper patient over a time horizon of 1 year. Thiscompares with a cost of £1098 and 0.8289 in thebase case (difference in QALYs due to randomvariation). This sensitivity analysis relies on theassumption that the relative treatment effects onADHD are independent of the presence of co-morbid conditions.

If DEX is not suitable as first-line therapy, strategy7 (first-line IR-MPH, second-line DEX, third-lineATX) is optimal. In this analysis strategy 9 (first-line ER-MPH12, second-line DEX, third-line ATX)is more costly and more effective compared withstrategy 7, but the cost per QALY gained is£5,697,763.

The base case analysis also does not considerpatients with co-morbid conditions that makethem unsuitable for treatment with stimulants,such as severe tics or Tourette’s syndrome. In thesepatients, MPH and DEX may be unsuitable,leaving ATX as the only availablepharmacotherapy. The submission by Eli Lillyused data from an unpublished trial comparingATX with placebo in children with Tourette’ssyndrome and severe tics. The response rate toATX was found to be 66.67%, defined as areduction of ≥25% on the parent-rated ADHD-RS,which compares with a response rate of 65.08% inchildren and adolescents without thesecontraindications to treatment with stimulants.

No published trials were available comparing ATXwith placebo in patients with severe tics orTourette’s syndrome. As such, a sensitivity analysiswas conducted based on the assumption that therelative treatment effect of ATX on ADHD is

independent of the presence of tics or Tourette’ssyndrome. Using the base case estimates ofresponse rates (estimated from trials that excludedpatients with severe tics and Tourette’s syndrome),the cost per QALY gained with ATX comparedwith no treatment (strategy 19) is £7951. Hencetreatment with ATX appears cost-effective inpatients who are contraindicated to treatment withstimulants. If society were willing to pay £30,000per additional QALY, treatment with ATX wouldhave an 86% probability of being cost-effective.

Sensitivity to time horizonThe base case model considers a time horizon of1 year. At the end of this year, the cohort isdivided into responders on various medicationsand non-responders on no medication. It isunlikely that the proportion of patients in each ofthe health states at the end of 1 year will remainthe same indefinitely. Unfortunately, there is a lackof long-term data in this area that might informthe model in terms of long-term adverse eventswith treatment, length of treatment and long-termbenefits of treatment. As such, no extrapolationwas considered in the base case analysis.

Two studies were identified that explored the age-dependent decline of symptoms of ADHD.148,158

The data provided by Hill and Schoener148 weretransformed into a yearly probability of remissionof 13% (50% over 5 years). This estimate of 50%remission over 5 years was calculated using a non-linear regression analysis on cohorts of childrenwho received a mixture of treatments for ADHD.The yearly rate was applied to patients in eachhealth state, including non-responders. Patients inremission were assumed to be identical withmedication responders, without the cost ofmedication itself. The long-term model has beendescribed in the section ‘Extrapolation’ (p. 103).

Figure 25 illustrates the simple structure, whichemploys a 1-year cycle length. Table 95 shows theresults from the extrapolation. Strategy 13 remains

Economic model

114

TABLE 94 Additional resource use data attributed to co-morbid conditions in patients with ADHD

Item Average per year Lower CI limit Upper CI limit

ConsultationsPsychiatrist 2.25 1.30 3.00Paediatrician 0.25 0.00 0.70GP 2.25 1.30 3.40TestsBlood test 0.53 0.00 1.43ECG 0.03 0.00 0.07

the cheapest alternative, and strategies 7 and 9 arenot dominated. For values of willingness to payper QALY of more than £7128, strategy 7 (first-line IR-MPH, second-line DEX, third-line ATX)appears cost-effective. The cost per QALY gainedwith strategy 9 (first-line ER-MPH12, second-lineDEX, third-line ATX) compared with strategy 7 is

£37,802,566. If the societal value of willingness topay per additional QALY were £30,000, strategy 7would have a 19% probability of being the optimalstrategy. Again, the distribution of incremental netbenefit is skew, and so at £30,000 per QALY,strategy 13 has a 51% probability of being theoptimal strategy, but it does not have the highest


115


Responder to MPH

Responder to ATX

Responder to DEX

Non-responder

Base casemodel Remission

FIGURE 25 Representation of the long-term extrapolation model

TABLE 95 Results of the sensitivity analysis extrapolating the model to when the cohort reaches 18 years of age


1 IR-MPH – ATX – DEX – no treatment 9,514 9.2403 D2 ER-MPH8 – ATX – DEX – no treatment 10,649 9.2395 D3 ER-MPH12 – ATX – DEX – no treatment 10,696 9.2386 D4 ATX – IR-MPH – DEX – no treatment 10,646 9.2402 D5 ATX – ER-MPH8 – DEX – no treatment 11,002 9.2398 D6 ATX – ER-MPH12 – DEX – no treatment 11,009 9.2382 D7 IR-MPH – DEX – ATX – no treatment 9,016 9.2597 7,1288 ER-MPH8 – DEX – ATX – no treatment 10,553 9.2590 D9 ER-MPH12 – DEX - ATX – no treatment 10,882 9.2597 37,802,566

10 ATX – DEX – IR-MPH – no treatment 11,554 9.2594 D11 ATX – DEX – ER-MPH8 – no treatment 11,574 9.2594 D12 ATX – DEX – ER-MPH12 – no treatment 9,036 9.2592 D13 DEX – IR-MPH – ATX – no treatment 8,885 9.2413 –14 DEX – ER-MPH8 – ATX – no treatment 9,187 9.2408 D15 DEX – ER-MPH12 – ATX – no treatment 9,196 9.2394 D16 DEX – ATX – IR-MPH – no treatment 9,172 9.2412 D17 DEX– ATX – ER-MPH8 – no treatment 9,277 9.2409 D18 DEX– ATX – ER-MPH12 – no treatment 9,291 9.2393 D19 No treatment 9,580 8.8896 D

D, dominated.

expected net benefit. The cost-effectivenessfrontier for this sensitivity analysis is shown inFigure 26. Strategy 7 appears cost-effective forvalues of willingness to pay per QALY above£7128, and strategy 9 does not appear to be cost-effective.

Clearly, this model does not incorporate long-termadverse effects of pharmacotherapy, as the datawere not available to include this, and this is animportant omission from the model. The modelalso does not include long-term benefits oftreatment, which could perhaps be avoidance ofprison, lower numbers of exclusions from schooland improved peer relations. However, thisomission is less critical, because the inclusion ofthese benefits would only improve the cost-effectiveness of active treatment.

Sensitivity to estimated resource useIn the base case model, the no treatment option(strategy 19) is dominated by an active treatmentoption. This is due in part to the estimates ofresource use employed in that model, in particularthe fact that a non-responder is more costly interms of non-drug resource use compared with aresponder. In the submission by Celltech, analternative assumption was used regardingresource use. They followed the Wessex DECevaluation124 in assuming that responders to

treatment had six visits to a psychiatrist per yearand six visits to a GP per year. Non-responderswere assumed to have only two visits to a GP peryear, and were therefore less costly thanresponders in terms of non-drug resource use.

In this sensitivity analysis, we evaluate the impacton the model results of employing this alternativesource of resource use. The uncertaintysurrounding the resource use was characterisedusing a gamma distribution. The 95% CI for thenumbers of psychiatrist and GP visits forresponders was assumed to be 3 to 9 per year. The95% CI around the number of GP visits for non-responders was assumed to be 0 to 4 per year. Theresults of this analysis are shown in Table 96.Strategy 19 (no treatment) is no longerdominated.

The cost per QALY gained with strategy 13 (first-line DEX, second-line IR-MPH, third-line ATX)compared with strategy 19 is £14,939. Strategy 13remains the optimal treatment strategy in thissensitivity analysis with radically differentestimates of resource use.

If DEX is not suitable as first-line therapy, strategy7 (first-line IR-MPH, second-line DEX, third-lineATX) remains the optimal treatment strategy, witha cost per QALY gained of £15,662 compared with

Economic model

116

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0


Prob

ablii

ty c

ost-

effe

ctiv

e 13

7

FIGURE 26 Cost-effectiveness acceptability frontier showing the optimal strategies in a sensitivity analysis of long-term extrapolation

no treatment. Again, strategy 9 is not ruled out bydominance or extended dominance, but the costper QALY gained compared with strategy 7 isoutside the range normally considered cost-effective (£5,808,184).

Sensitivity to structural assumptionregarding MPHA major structural assumption in the model is thata patient who has failed treatment with oneformulation of MPH would not then receive adifferent formulation of the same drug. However,if failure on IR-MPH was due to non-compliance,or because a midday dose was simply unworkablefor the patient, ER-MPH may be a relevanttreatment option. The effect of compliance onresponse rates to IR-MPH and ER-MPH isreflected in the model. However, the model, andthe trial data, do not identify the proportion of

non-responders that fail owing to lack ofcompliance. As such, this sensitivity analysisconsiders a hypothetical cohort of patients forwhom a midday dose of medication is consideredunsuitable. This includes patients who have failedon IR-MPH or DEX where the clinician canidentify that the reason for failure is non-response.Also included are patients for whom the clinicianjudges a midday dose of medication to beunworkable.

The model is a simple comparison of ER-MPH8,ER-MPH12 and ATX and no treatment. Theresults of the analysis are shown in Table 97. Theyshow that in this selected patient population, ATXis dominated by ER-MPH12. Hence in thosepatients for whom a midday dose of medication isunsuitable, through non-compliance or for otherreasons, ER-MPH12 would precede ATX in any


117


TABLE 96 Results of the sensitivity analysis employing alternative estimates of resource use


1 IR-MPH – ATX – DEX – no treatment 1006 0.8279 D2 ER-MPH8 – ATX – DEX – no treatment 1233 0.8273 D3 ER-MPH12 – ATX – DEX – no treatment 1250 0.8278 D4 ATX – IR-MPH – DEX – no treatment 1250 0.8278 D5 ATX – ER-MPH8 – DEX – no treatment 1320 0.8277 D6 ATX – ER-MPH12 – DEX – no treatment 1327 0.8274 D7 IR-MPH – DEX – ATX – no treatment 920 0.8283 D8 ER-MPH8 – DEX – ATX – no treatment 1105 0.8277 D9 ER-MPH12 – DEX – ATX – no treatment 1190 0.8284 D

10 ATX – DEX – IR-MPH – no treatment 1242 0.8281 D11 ATX – DEX – ER-MPH8 – no treatment 1261 0.8281 D12 ATX – DEX – ER-MPH12 – no treatment 1259 0.8278 D13 DEX – IR-MPH – ATX – no treatment 887 0.8288 14,93914 DEX – ER-MPH8 – ATX – no treatment 944 0.8287 D15 DEX – ER-MPH12 – ATX – no treatment 945 0.8286 D16 DEX – ATX – IR-MPH – no treatment 946 0.8288 D17 DEX– ATX – ER-MPH8 – no treatment 965 0.8288 D18 DEX– ATX – ER-MPH12 – no treatment 963 0.8285 D19 No treatment 48 0.7727 –

D, dominated.

TABLE 97 Cost-effectiveness of ER-MPH8, ER-MPH12, ATX and no treatment in patients for whom a midday dose of medication isunsuitable


19 No treatment 1223 0.7731 –20 ATX only 1517 0.8093 D22 ER-MPH8 only 1360 0.8053 4251 (vs 19)23 ER-MPH12 only 1427 0.8140 7670 (vs 22)

D, dominated.

treatment strategy. Whether it is possible toidentify such patients in practice is a challenge forthe clinician in charge.

Sensitivity to estimated response ratesThe section ‘Clinical effectiveness’ (p. 103)illustrated the numerous definitions of responseavailable in the trial data. In order to utilisedifferent definitions of response, it was necessaryto model a relationship between the alternativedefinitions. This was achieved by extending thebase case MTC model. As a reminder, the basecase model assumes a binomial likelihood fromthe available data points, r and n, and theproportion of responders, p, is estimated in aregression-like structure, with the logit of theproportion of responders dependent on a study-specific baseline, �, and a treatment effect, �:

rik ~ Bin(pi

k, nik)

logit(pik) = �i + �i

k

To extend this model, let us assume that responsedefined as a score of 1 or 2 on the CGI-I isresponse 1, which is modelled as follows:

r1ik ~ Bin(p1i

k, n1ik)

logit(p1ik) = �1i + �1i

k

Response defined as a score of 1 or 2 on the CGI-S is response 2, and response on this scale ismodelled to be conditional on the estimatedresponse rate on scale 1:

r2ik ~ Bin(p2i

k, n2ik)

logit(p2ik) = �2i + �2i

k

If both measures capture the same effect, �2 maybe random error. If the measures capture differenteffects, the relationship between the measures isestimated according to this model, and the

correlation between the two is reflected. Becausewe have trials that report response on more thanone measure, the relationship between thedifferent measures is estimated from the data.Through this relationship, trials that reportresponse only on scale 2 can inform the estimateof response on scale 1. By selecting whichdefinition of response is to be the baseline in themodel (scale 1), we can infer response rates onthat scale, strengthened by information aboutresponse on different scales. The code for thisextended model is given in Appendix 11.

Using the framework described above, we couldbring in trials reporting response on the CGI-S, inorder to synthesise all clinician-rated responsedata. Two of the trials reporting response on CGI-I for the base case analysis also reported responsedefined as a score of 1 or 2 on the CGI-S. Threemore trials reported response on CGI-S, but noton CGI-I. The additional information provided bythese trials is shown in Table 98.

The three additional trials were all set in the USAand used DSM-IV diagnostic criteria. Theyrecruited patients of varying age ranges (6–12,6–16 and 8–12 years), and this again is notreflected in the model. The output from 2WinBUGS models, using CGI-I and CGI-S asbaseline, respectively, is shown in Table 99. Where CGI-I is used as the baseline responsedefinition, the results are slightly higher than thebase case analysis and the uncertainty around the estimated treatment effects is reduced. Usingthe same information, but specifying CGI-S as the baseline scale, the order of the treatmenteffects remains stable, but the absolute effects arelower. This reflects the lower absolute responserates reported using CGI-S in comparison with CGI-I.

Economic model

118

TABLE 98 Response rates defined as score of 1 or 2 on CGI-S



Steele, 200490 ER-MPH12IR-MPH [Confidential information removed]



Weiss, 200494 ATXPlacebo [Confidential information removed]

The same method for synthesising the clinician-rated response rates can be used to estimate therelationship between response defined on aclinician-rated scale and response defined on aparent-rated scale. This allows us to overcome thelack of parent-rated response rate data for ER-MPH8 and DEX. As highlighted in the section‘Clinical effectiveness’ (p. 103), a group of sixtrials reported response defined as a reduction of≥ 25% on the parent-rated ADHD-RS. Three ofthe six trials also reported response defined as ascore of 1 or 2 on CGI-S, and one reportedresponse defined as a score of 1 or 2 on CGI-I.The additional data provided on this newdefinition of response is shown in Table 100.

The trials reported by Spencer and colleagues89

recruited patients aged between 7 and 13 years,using DSM-IV diagnostic criteria in a US setting.At this stage, we can also incorporate the results ofthe MTA trial,133 but only by assuming that themedical management group in that trialrepresents treatment with IR-MPH. The review inChapter 4 highlighted that although the majorityof the medication in the MTA trial was IR-MPH, aproportion consisted of other medications. Thedefinition of response in the MTA trial was a score

of ≤1 on the SNAP-IV scale. In order to calculateresponse rates, investigators averaged over theteacher and parent ratings for each item on thescale. The trial by Steele and colleagues90 alsoreported response defined as a score of ≤1 on theSNAP-IV scale. Table 101 shows the response rateinformation available on the SNAP-IV scale(assuming that medical management in the MTAtrial is equal to IR-MPH). The nature of thetreatment received in the community comparisonarm of the MTA trial is still unclear, and as a resultthese data are omitted from the analysis.

Hence in the final estimation of response rates, weinclude all of the data from Tables 86, 98, 100 and101. In other words, we synthesise responsedefined on the CGI-I scale, CGI-S scale, ADHD-RS and SNAP-IV scale, by estimating therelationships between response defined on thedifferent scales. This final analysis alsoincorporates data reported in Quinn84 and theresults from Elia and colleagues.51 [Confidentialinformation removed]. The study by Elia andcolleagues51 used DSM-III diagnostic criteria anddefined response as a score of 1, 2 or 3 on CGI-I.This increases the heterogeneity between studiessynthesised in this estimate of treatment effects.


119


TABLE 99 Response rates estimated in extended MTC model in WinBUGS: response defined as score of 1 or 2 on CGI-I or CGI-S

Treatment Response rate, CGI-I baseline (SD) Response rate, CGI-S baseline (SD)


TABLE 100 Response rates defined as reduction of ≥ 25% on the ADHD-RS




Kemner, 200499 ER-MPH12ATX [Confidential information removed]

Weiss, 200494 ATXPlacebo

Spencer, 2002 (reported results of 2 trials)89 ATX 42 (65) 65Placebo 15 (24) 62

Spencer, 200289 ATX 38 (59) 64Placebo 25 (40) 62

The ADHD-RS is chosen as the baseline responserate, in order to obtain parent-rated estimates oftreatment effect. The results of this analysis arecompared to the same model using a CGI-Ibaseline, to aid comparison with the base caseresults. These data are shown in Table 102.

Again, the new information has altered theresponse rates somewhat, but the order in terms oftreatment effect remains the same. Responsedefined on the ADHD-RS produces a higherabsolute number of responders than responsedefined on the CGI-I scale.

Results using all clinician-rated responseTable 103 shows the results of the model using theresponse rates reported in Table 98. The new setsof response rates alter the results slightly becausestrategy 13 no longer dominates all the otherstrategies. Instead, strategy 15 (first-line DEX,second-line ER-MPH12, third-line ATX) is moreeffective, but at a higher cost. The point estimateof response to ER-MPH12 is higher than the pointestimate of response to IR-MPH in the base caseand in the extended MTC, but in the latter modelthe relative difference in treatment effects is morefavourable towards ER-MPH12.

The cost per QALY gained with strategy 15compared with strategy 13 falls when responsesare estimated on the CGI-S scale in comparison tothe CGI-I scale. This is because overall theresponse rates are lowered when measured on

CGI-S, and there is greater potential to increasethe number of responders by switching to a moreeffective treatment. In the analysis using responsemeasured on CGI-I, a much larger proportion ofpatients will have responded to first-line DEX, sothe potential for increasing the number ofresponders by altering the second- or third-linetherapies is reduced. In both cases, the cost perQALY gained with strategy 15 compared withstrategy 13 is probably outside the range of valuesnormally considered to be cost-effective, sostrategy 13 remains optimal.

For both of these analyses, if DEX is notconsidered suitable as first-line therapy, theoptimal strategy is number 7 (first-line IR-MPH,second-line DEX, third-line ATX). Strategy 9 isnot ruled out by dominance or extendeddominance, but the cost per QALY gainedcompared with strategy 7 is outside the rangenormally considered cost-effective (>£150,000 perQALY).

Results using parent-rated response: synthesisingall response ratesTable 104 shows the results of the model using theresponse rates reported in Table 102. The resultsare very similar to those using only clinician-ratedresponse data (Table 101). Strategy 13 is no longerthe dominant treatment strategy, but the cost perQALY gained to move to the next most effectivestrategy, number 15, is outside the range normallyconsidered cost-effective.

Economic model

120

TABLE 102 Response rates estimated in extended MTC model in WinBUGS: response defined on CGI-I, CGI-S, ADHD-RS or SNAP-IV

Treatment Response rate, CGI-I baseline (SD) Response rate, ADHD-RS baseline (SD)


TABLE 101 Response rates defined as score of ≤ 1 on the SNAP-IV scale


Steele, 2004 ER-MPH12 [Confidential information removed]IR-MPH

Swanson, 2001a IR-MPH 81 (56) 144IR-MPH + BT 99 (68) 145Community comparison 37 (25) 146

a Results for BT alone omitted as not relevant to this review.


121


TABLE 103 Results of the economic model using all response rates estimated on clinician-rated scales

Strategy Order of treatments received CGI-I scale baseline CGI-S scale baseline

Cost (£) QALY ICER (£) Cost (£) QALY ICER (£)

1 IR-MPH – ATX – DEX – no treatment 1,223 0.8322 D 1,239 0.8244 D2 ER-MPH8 – ATX – DEX – no treatment 1,468 0.8315 D 1,413 0.8230 D3 ER-MPH12 – ATX – DEX – no treatment 1,507 0.8328 D 1,456 0.8261 D4 ATX – IR-MPH – DEX – no treatment 1,496 0.8319 D 1,378 0.8238 D5 ATX – ER-MPH8 – DEX – no treatment 1,569 0.8315 D 1,460 0.8229 D6 ATX – ER-MPH12 – DEX – no treatment 1,598 0.8321 D 1,510 0.8250 D7 IR-MPH – DEX – ATX – no treatment 1,119 0.8326 D 1,140 0.8254 D8 ER-MPH8 – DEX – ATX – no treatment 1,341 0.8320 D 1,293 0.8242 D9 ER-MPH12 – DEX – ATX – no treatment 1,426 0.8331 D 1,378 0.8269 D10 ATX – DEX – IR-MPH – no treatment 1,478 0.8322 D 1,356 0.8247 D11 ATX – DEX – ER-MPH8 – no treatment 1,488 0.8320 D 1,378 0.8241 D12 ATX – DEX – ER-MPH12 – no treatment 1,491 0.8324 D 1,387 0.8256 D13 DEX – IR-MPH – ATX – no treatment 1,065 0.8334 – 1,102 0.8267 –14 DEX – ER-MPH8 – ATX – no treatment 1,098 0.8332 D 1,142 0.8260 D15 DEX – ER-MPH12 – ATX – no treatment 1,102 0.8337 177,000 1,150 0.8278 46,00016 DEX – ATX – IR-MPH – no treatment 1,101 0.8334 D 1,131 0.8266 D17 DEX – ATX – ER-MPH8 – no treatment 1,111 0.8332 D 1,153 0.8260 D18 DEX – ATX – ER-MPH12 – no treatment 1,114 0.8336 D 1,162 0.8275 D19 No treatment 1,225 0.7741 D 1,224 0.7727 D

D, dominated.

TABLE 104 Results of the economic model using all response rates defined on CGI-I, CGI-S, ADHD-RS or SNAP-IV

Strategy Order of treatments received CGI-I scale baseline ADHD-RS scale baseline

Cost (£) QALY ICER (£) Cost (£) QALY ICER (£)

1 IR-MPH – ATX – DEX – no treatment 1,240 0.8304 D 1,225 0.8316 D2 ER-MPH8 – ATX – DEX – no treatment 1,449 0.8298 D 1,464 0.8312 D3 ER-MPH12 – ATX – DEX – no treatment 1,492 0.8317 D 1,507 0.8324 D4 ATX – IR-MPH – DEX – no treatment 1,444 0.8301 D 1,487 0.8313 D5 ATX – ER-MPH8 – DEX – no treatment 1,527 0.8298 D 1,565 0.8311 D6 ATX – ER-MPH12 – DEX – no treatment 1,568 0.8308 D 1,596 0.8316 D7 IR-MPH – DEX – ATX – no treatment 1,117 0.8311 D 1,115 0.8321 D8 ER-MPH8 – DEX – ATX – no treatment 1,312 0.8306 D 1,339 0.8318 D9 ER-MPH12 – DEX – ATX – no treatment 1,405 0.8322 D 1,424 0.8328 D10 ATX – DEX – IR-MPH – no treatment 1,423 0.8308 D 1,467 0.8317 D11 ATX – DEX – ER-MPH8 – no treatment 1,435 0.8306 D 1,475 0.8316 D12 ATX – DEX – ER-MPH12 – no treatment 1,439 0.8312 D 1,478 0.8319 D13 DEX – IR-MPH – ATX – no treatment 1,068 0.8325 – 1,059 0.8333 –14 DEX – ER-MPH8 – ATX – no treatment 1,096 0.8323 D 1,082 0.8331 D15 DEX – ER-MPH12 – ATX – no treatment 1,101 0.8330 63,690 1,085 0.8335 108,74716 DEX – ATX – IR-MPH – no treatment 1,095 0.8325 D 1,083 0.8332 D17 DEX – ATX – ER-MPH8 – no treatment 1,106 0.8323 D 1,091 0.8331 D18 DEX – ATX – ER-MPH12 – no treatment 1,111 0.8329 D 1,094 0.8334 D19 No treatment 1,225 0.7739 D 1,227 0.7735 D

D, dominated.

The CEACs from the synthesis of all responserates are very similar whether they are based onCGI-I or ADHD-RS. Figure 27 shows the CEACswhen the response rates are synthesised against anADHD-RS baseline.

If society were willing to pay £30,000 peradditional QALY, strategy 13 has a 93%probability of being most cost-effective (amongstrategies including three active treatments). Thisgain in certainty compared with the base casereflects the incorporation of more trial data.

For both of these analyses, if DEX is not suitableas first-line therapy, the optimal treatment strategyis number 7 (first-line IR-MPH, second-line DEX,third-line ATX). In the analysis with all responserates synthesised against the parent-rated ADHD-RS baseline, if society were willing to pay £30,000per additional QALY, strategy 7 has a 100%probability of being the optimal treatmentstrategy. Again, the uncertainty is decreased by theincorporation of more trial data. Also, theuncertainty is reduced compared to the case whereDEX is considered suitable as first-line therapybecause there are six fewer strategies beingcompared. Strategy 9 is not ruled out bydominance or extended dominance, but the costper QALY gained compared with strategy 7 isoutside the range normally considered cost-effective (>£250,000 per QALY).

[Following this appraisal in 2004, the price ofdexamfetamine rose, in 2005, from £1.92 to £3.00per 28 5-mg tablets. Despite this price increase, itremained the cheapest of the alternative drugtreatments, and the results of the economicanalysis were robust to this price increase.]

DiscussionThis study sought to answer the question of whichtreatment or treatment strategy is most cost-effective once one has assessed there to be a needfor medical management in children andadolescents with ADHD. A very simple economicmodel was constructed to assess the cost-effectiveness of alternative treatment options. Themodel was necessarily simple owing to a lack ofdata relating to, in particular, clinically meaningfulresponse rates, long-term effects of treatment andresponse to treatment conditional on factors suchas previous treatment, ADHD subtype, gender andage. Some of these data deficiencies were exploredin scenario analyses by specifying modellingassumptions.

In the base case analysis, response rates werecalculated by using only few additionalassumptions above those found in a typical meta-analysis. However, this severely reduced thenumber of trials informing the model to only six

Economic model

122

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0 10,000 20,000 30,000 40,000 50,000

Strategy 15

Strategy 13

60,000Willingness to pay per QALY (£)

Prob

abili

ty c

ost-

effe

ctiv

e

12345678910111213141516171819

FIGURE 27 CEACs for 19 strategies compared: synthesising all response rates against an ADHD-RS baseline

of the 64 identified in the clinical effectivenessreview. This analysis showed that a treatmentstrategy of first-line DEX, followed by second-lineIR-MPH for treatment failures, followed by third-line ATX for repeat treatment failures wasoptimal. This strategy remained optimal whenincluding the additional costs of co-morbidconditions, extrapolating the model to age18 years, and using alternative estimates ofresource use in ADHD. A sensitivity analysisemploying alternative estimates of the utilityvalues associated with ADHD did alter the results,but problems in interpreting these alternativeutility values mean that the result may not bevalid. It is interesting that, despite using adifferent measure of efficacy to that used in theclinical effectiveness review in Chapter 4, theconclusions are broadly similar. Chapters 4 and 6indicate the superiority of drugs over no drugtherapy, but show no significant difference interms of efficacy or side-effects between the activetherapies. This lack of significant difference couldbe due simply to the lack of available evidence inboth cases. The main additional feature of thischapter in comparison with Chapter 4 is theconsideration of costs. Given the lack of evidencefor differences in efficacy between active therapies,the results of the economic model are largelydriven by drug cost, in which there are markeddifferences between the relevant comparators.

It is possible that DEX may not be consideredsuitable as a first-line therapy; if the interpretationof its licence is that it can only be used followingfailure on another medication for ADHD, or ifconcerns about its abuse potential (not reflected inthe model) are so much greater than concernsabout MPH. In this case, the model identified anoptimal treatment strategy of first-line IR-MPH,followed by second-line DEX for treatmentfailures, followed by third-line ATX for repeattreatment failures. However, data on the relativeabuse potential of MPH and DEX are scarce, andthe costs and dis-benefits associated with suchabuse would be hard to quantify in this simplemodel. As such, we presented results for the casewhere DEX is considered suitable as first-linetherapy, and the case where it is not, leaving thereader to decide whether strategies with DEX first-line therapy are relevant comparators.

It is feasible that factors in the school environmentor characteristics of particular patients withrespect to compliance make a midday dose ofmedication unworkable. If this is the case, theanalysis in the section ‘Sensitivity to structuralassumption regarding MPH’ (p. 117) indicates

that ER-MPH12 would be preferred to ATX, and so would precede ATX in any treatmentstrategy.

In order to increase the number of trials used incalculating response rate, the MTC model used tosynthesise the data in the base case analysis wasextended to estimate a relationship betweenresponses defined by different criteria. Bringing inadditional trials increased the certainty around themodel results, and a three-treatment strategy(first-line DEX, second-line IR-MPH, third-lineATX) remained cost-effective. However, even thisextended analysis only employs data from 14 outof 64 trials. The model also relies on non-drugresource use data estimated using an expert panel.This increases the uncertainty in the modelresults, but the model was robust to using verydifferent estimated resource use.

For a decision taken now, with current availabledata, the results of the economic evaluation clearlyidentify an optimal treatment strategy. However,the model is not without limitations, and new dataon long-term outcomes associated with medicalmanagement of ADHD could change the analysissignificantly. The model considers a broadlydefined set of patients with ADHD, as the data didnot allow us to discriminate between patients interms of ADHD subtype, gender, age or previoustreatment. The resource use data are based onexpert opinion, which may not reflect resource usein practice in the UK. Similarly, resource use andmonitoring in UK practice may be lower than thatobserved in clinical trials, which could translate tolower effectiveness in practice compared with thatobserved in clinical trials. The effectiveness dataused in the model are based on trials set in NorthAmerica, owing to the lack of data specific to theUK. Although there are many trials in this area,the outcome measures used, and the way in whichresults are reported, vary widely. As such, thetreatment effects used to populate the economicmodel are based on a subset of the clinicalevidence, predominantly recent trials. An analysisof treatment effects that included all the availabletrial data would involve many assumptionsregarding the clinical value of mean difference indifferent rating scales, the role of baseline meanscore in relation to mean difference, thedistribution of mean score and mean difference oneach rating scale and the utility associated withmean score on different rating scales. Suchassumptions would go beyond what is justified bythe data, so it was felt that a more robust synthesis,based on a subset of the available evidence, wouldbe more appropriate.


123


Effectiveness evidenceThis review presents a comprehensive overview ofstudies (that meet our inclusion criteria) on MPH,DEX and ATX published up to July/August 2004.Overall, the systematic review of effectivenessstudies represents an archive of data for reference.The results are presented separately for low,medium and high doses. Although weacknowledge that this may be less relevant forclinical practice, because the dose will beestablished by titration for individual patients, thisapproach was chosen for the SR because most ofthe evidence was presented by dose.

The majority of included studies were of poorquality. Only five out of the 60 non-confidentialstudies properly reported the method used toassign participants to patient groups. Nine out ofthe 60 non-confidential studies reported that thesequence of allocation was truly random. Moststudies were blinded, but none of the includedstudies reported whether blinding was successful.Eleven out of the 60 non-confidential studiesreported to have used an ITT analysis and 18 outof the 60 non-confidential studies provided acomplete description of withdrawals. For most ofthe crossover trials the statistical analysis waseither not clearly reported or not appropriate; forparallel trials the quality of the analysis was betterreported.

It should be noted that some of the quality itemsreflect the quality of reporting and not necessarilythe quality of the actual trial. However, given thepoor scores on these quality items, the reliabilityof the study results is unknown.

As reported in the previous NICE report4 and theAHRQ28 and CCOHTA30 reviews, the plethora ofMPH studies suggest that MPH (both immediate-and extended-release) is effective at reducinghyperactivity, and improving QoL (as determinedby CGI – either the Improvement or Severitysubscale) in children. It was noted, however, thatthe majority of studies that evaluated theeffectiveness of MPH did not adequately reporttheir study methodology. There appears to belittle evidence of a statistically significant differencein the effectiveness of IR-MPH and ER-MPH.

Similarly, DEX also appears to be effective atreducing hyperactivity and improving QoL.However, there were generally very few high-qualitystudies that evaluated the effectiveness of DEX. Aswith MPH, some of the results were variable.

There was consistent evidence that ATX wassuperior to placebo for hyperactivity and CGI.These more recent studies on ATX had well-reported study methodologies and the results arelikely to be reliable.

Very few studies made head-to-head comparisonsbetween the drugs. The previous NICE report4

stated that there appeared to be little difference inthe effectiveness of MPH and DEX. No recentstudies were found in our updated search. Whilethe studies reported variable results, the one studythat reported no statistically significant differencesbetween drugs was deemed to be of good quality,whereas the quality of the others was uncertain,given the poor reporting of study methodologies.

One study that compared MPH and ATX reportedno statistically significant difference between thedrugs for hyperactivity or CGI. However, thisstudy did not adequately report on their studymethodology. Hence there is insufficient evidenceto judge the relative effectiveness of these drugs.

Few studies were included in the review thatexamined a non-drug intervention in combinationwith MPH, DEX or ATX. Generally, the resultswere variable. The studies were, however,heterogeneous regarding the type of non-druginterventions examined and the scales used tomeasure outcomes.

Adverse eventsUpon examination of the evidence, it is clear thatadequate and informative data regarding thepotential adverse effects of MPH, DEX and ATXare lacking. Of the 64 studies included in theclinical effectiveness section of this review, 38contributed some data to the analysis of adverseevents. However, this contribution was minimal inthe majority of cases: only eight studies, forexample, provided usable data regarding the


125


Chapter 7

Discussion

impact of active drug treatment on participantweight loss/gain.

The analysis of adverse events was focused on themost commonly reported complaints of individualsreceiving treatment for ADHD, which were identifiedby examining previous literature. Several studiespresented mean severity ratings for adverse eventswithout accompanying data indicating the numbersof participants actually suffering from such effects,thus making adverse impact difficult to quantify.Furthermore, a distinction between numbers ofadverse events and numbers of participants sufferingfrom adverse events was often not clearly stated.Total numbers of participants included in safetyanalyses were often unclear, requiring assumptions tobe made which could have underestimated effectsizes. Measures of variance were not always reported,making calculation of relative risks impossible.Although standardised reporting was employed bymany authors, some studies appeared to rely uponspontaneous reporting of somatic complaints, whichmay also have introduced considerable bias into theresults. Some trials reported the mean weights ofparticipants across treatment groups at follow-uprather than mean weight change; for crossover trials,this information was not informative owing to thelikelihood of carryover effects.

Overall, higher dosages of IR-MPH appear to beassociated with the occurrence of headache, loss ofappetite, stomach ache and insomnia comparedwith placebo. ER-MPH appears to be associatedwith decreased appetite and increased insomnia.However, a systematic review114 highlighted theneed for further research into somatic complaints,which may be associated with the disorder itselfrather than MPH treatment. Similarly, high dosesof DEX appear to be associated with decreasedappetite and increased sleeping problems. ATX ofany dose may impair appetite.

As previously highlighted, head-to-headcomparisons have not often been examined and,together with poor reporting of adverse eventsoutcomes, data are very sparse. One studycomparing immediate- and extended-releaseformulations of MPH reported a higheroccurrence of headache in the latter group. Nostatistically significant differences in adverseevents were detected in the studies comparingMPH with DEX. However, participants assigned toER-MPH were found to suffer from decreasedappetite and increased insomnia compared withthose assigned to ATX.

No studies compared ATX with DEX.

Economic evidenceThis review presents a comprehensive overview ofexisting economic evaluations of MPH, ATX andDEX for children and adolescents with ADHD,including three submissions from manufacturers ofthese medications. The review highlighted anumber of potential limitations in the existingliterature. In particular, the review highlightedlimitations in estimating treatment effectivenessand associated utility values. These limitationsmay stem from a lack of available data.

A new economic model was developed for thisreport. Pooling was limited in the clinicaleffectiveness review, owing to heterogeneitybetween trials. However, some degree of pooling isnecessary to proceed with an economic model.The issue of heterogeneity was overcome by basingthe base case on trials that are more similar interms of how they measure the outcome ofinterest. In a series of sensitivity analyses, moretrials were included by relaxing the criterion ofsimilarity in outcome measurement. Data onresource use associated with ADHD in the UKwere lacking, so the model relies on estimatesfrom experts.

Given the lack of available evidence for statisticallysignificant differences in efficacy between thealternative drugs, the results of the economicmodel were largely driven by drug cost, in whichthere are marked differences. For a decision takennow, with current available data, the results of theeconomic evaluation clearly identified an optimaltreatment strategy, that is, first-line DEX, followedby second-line IR-MPH for treatment failures,followed by third-line ATX for repeat treatmentfailures. If DEX is considered not suitable as afirst-line therapy, the optimal strategy is first-lineIR-MPH, followed by second-line DEX, and third-line ATX. For patients contraindicated tostimulants, ATX is preferred to no treatment. Forpatients in whom a midday dose of medication isunworkable, ER-MPH is preferred to ATX, andER-MPH12 appears more cost-effective than ER-MPH8.

The model is not without limitations. As identifiedin the clinical effectiveness review, the reporting ofstudies was poor, there are few data todiscriminate between the drugs in efficacy oradverse events and there are few data on long-term efficacy and adverse events associated withmedical management of ADHD. The data do notallow discrimination between patients with ADHDin terms of ADHD subtype, age, gender or

Discussion

126

previous treatment. These caveats must be bornein mind when interpreting the model results.

Limitations of the clinicaleffectiveness studies and need forfurther researchThe limitations of the research base remain similarto those reported in the previous NICE review andin other SRs.

� Generally, the methodology used in the studieswas poorly reported. Inadequate informationmakes it difficult to assess validity and reliabilityof the studies.

� The literature is dominated by studies of MPH,although the quality of these studies is generallypoor. For DEX, both the quantity and quality ofstudies are poor. For ATX, there are few studies,but most are of better quality. The quantity andquality of studies evaluating MPH, DEX andATX should be a consideration for futureresearch. Researchers should use theCONSORT approach to study design.159

� Very few studies compared MPH and DEXdirectly. No studies compared ATX and DEXdirectly. More head-to-head studies may bebeneficial.

� Most of the studies report short-term outcomes.More long-term studies are needed to establishclinical effectiveness and adverse events.

� There is an almost total absence of researchinvestigating the effectiveness of MPH andother interventions for children with a diagnosisof hyperactivity disorder, rather than thebroader diagnosis of ADHD. This may haveimplications for the generalisability of thesestudies in the UK where HKD makes up most ofthe cases.

� The majority of studies have been conducted onchildren between the ages of 5 and 13 years.More studies in adolescents may be useful.

� It appears that, in some cases, parents are lesslikely than teachers to rate children as beingimproved.

� It may be useful when possible to collect datafrom the children to obtain a betterunderstanding of treatment from a patient’sperspective and to estimate patient-basedmeasures of clinical outcome.

� The studies used a number of outcomes andscales to measure behaviour, includinghyperactivity, making comparisons of resultsdifficult. It would be beneficial if researchgroups selected a core set of validated andclinically relevant outcomes to be measured inall studies, and to be consistent in analysing andreporting the results.

� A number of studies used a crossover design;however, the data were not always analysedusing a method specific to paired data.Presenting within-patient differences (andassociated SDs) would be useful.

� A number of the included studies have smallsample sizes. Presenting sample size calculationswould be useful in order to make sure thestudies are not underpowered, so that clinicallysignificant differences between interventions aredetected. As part of sample size calculations,authors should clearly specify a clinicallysignificant effect on a particular outcomemeasure.

� Very few studies included in the reviewpresented usable data for analysing adverseevents. Standardised reporting of data (e.g.presenting the numbers of adverse events,numbers of participants suffering from adverseevents and total numbers of participantsincluded in the safety analysis) is required toestimate prevalence.


127


MPH, DEX and ATX improve hyperactivityand QoL (as assessed using CGI as a

proxy) compared with placebo. The validity of the results is not known for studies of MPH and DEX, given the poor reporting of studymethodology in the majority of these trials. The recent trials on ATX are likely to be morereliable.

There were very few head-to-head studiescomparing MPH, DEX and ATX. No significantdifferences between the various drugs in terms ofefficacy or side-effects were found, mainly owingto a lack of evidence.

The main conclusions from this report are:

� Drug therapy seems to be superior to no drugtherapy.

� No significant differences between the variousdrugs in terms of efficacy or side-effects werefound, mainly owing to a lack of evidence.

� The additional benefits from BT (incombination with drug therapy) are uncertain.

The main additional feature of the economicmodel is the consideration of costs. Given the lackof evidence for any differences in effectivenessbetween the drugs, the model tends to be drivenby drug costs, which differ considerably.

Research recommendations� Future trials examining MPH, DEX and ATX

should include the assessment of tolerabilityand safety as a priority. Reporting should bestandardised and transparent. Researchersshould refer to the CONSORT approach tostudy design.159

� Longer term follow-up of individualsparticipating in trials could further informpolicy makers and health professionals. Suchdata could potentially distinguish between thesedrugs in a clinically useful way.

� In addition, research examining whethersomatic complaints are actually related to drugtreatment or to the disorder itself would beinformative.


129


Chapter 8

Conclusion

We thank the expert advisory panel for theiruseful advice and constructive comments on

the report. We also wish to thank ProfessorStephen Senn, University of Glasgow, whoprovided statistical advice on crossover trials.Thanks are also due to Alison Brown, AngeDavison, and Sally Kingsley, University of York,who provided technical assistance with the report,and to Carol Forbes who produced some of thetables. Ailsa Snaith and Linda McIntyre wereinvolved in data extraction.

Contribution of authorsSarah King (Research Fellow) was the leadreviewer responsible for writing the protocol andfinal report and was involved in study selection,data extraction, validity assessment and synthesisof data. Susan Griffin (Research Fellow) wasinvolved in the analysis of cost-effectiveness,writing the protocol, study selection, dataextraction, development of the economic modeland report writing. Zoé Hodges (Research Fellow)was the second reviewer involved in designing theAccess database, study selection, data extraction,validity assessment and writing the final report.Helen Weatherly (Research Fellow) was involved inthe analysis of cost-effectiveness, writing theprotocol, study selection, data extraction,development of the economic model and reportwriting. Christian Asseburg (Research Fellow) wasinvolved in the analysis of cost-effectiveness. GerryRichardson (Research Fellow) was involved in theanalysis of cost-effectiveness, data extraction,development of the economic model and

commenting on drafts of the economic sections ofthe report. Su Golder (Information Officer)devised the search strategy and carried out theliterature searches, wrote the search methodologysections of the protocol and final report andmanaged the Endnote library. Eric Taylor(Professor of Child and Adolescent Psychiatry)provided input and advice at all stages,commented on drafts of the protocol and finalreport and contributed to the discussion section ofthe report. Mike Drummond (Professor of HealthEconomics) provided input at all stages,commented on various drafts of the report andhad overall responsibility for the cost-effectivenesssection of the report. Rob Riemsma (SeniorResearch Fellow) was review manager responsiblefor overall management of the project and wasalso involved in study selection, data extraction,synthesis of data and writing the report.

NoteA number of sponsors submitted information tothe National Institute for Health and ClinicalExcellence in confidence and references to thisinformation have been removed from the report.However, it should be noted that the Institute’sAppraisal Committee has access to the full reportwhen drawing up their guidance.

This report was commissioned by the NHS R&DHTA Programme. The views expressed in thisreport are those of the authors and not necessarilythose of the NHS R&D Programme. Any errorsare the responsibility of the authors.


131


Acknowledgements

1. National Institute for Clinical Excellence (NICE).Guidance on the use of methyphenidate (Ritalin,Equasym) for attention deficit/hyperactivity disorder(ADHD) in childhood. Report No. 13. London:NICE; 2000.

2. American Psychiatric Association. Diagnostic andstatistical manual for mental disorders, fourth edition(DSM-IV). Washington, DC: American PsychiatricAssociation; 1994.

3. Agency for Health Care Policy and Research.Diagnosis of attention-deficit/hyperactivity disorder.Rockville, MD: Agency for Health Care Policy andResearch; 1999.

4. Lord J, Paisley S. The clinical effectiveness and costeffectiveness of methylphenidate for hyperactivity inchildhood. London: NICE; 2000.

5. Ford T, Goodman R, Meltzer H. The British childand adolescent mental health survey 1999: theprevalence of DSM-IV disorders. J Am Acad ChildAdolesc Psychiatry 2003;42:1203–11.

6. Mautner VF, Kluwe L, Thakker SD, Leark RA.Treatment of ADHD in neurofibromatosis type 1.Dev Med Child Neurol 2002;44:164–70.

7. Dunn DW, Austin JK, Harezlak J, Ambrosius WT.ADHD and epilepsy in childhood. Dev Med ChildNeurol 2003;45:50–4.

8. Elgen I, Sommerfelt K, Markestad T. Populationbased, controlled study of behavioural problemsand psychiatric disorders in low birthweightchildren at 11 years of age. Arch Dis Child FetalNeonatal Ed 2002;87:F128–32.

9. Indredavik MS, Vik T, Heyerdahl S, Kulseng S,Fayers P, Brubakk AM. Psychiatric symptoms anddisorders in adolescents with low birth weight. ArchDis Child Fetal Neonatal Ed 2004;89:F445–50.

10. Rasmussen P, Gillberg C. Natural outcome ofADHD with developmental coordination disorderat age 22 years: a controlled, longitudinal,community-based study. J Am Acad Child AdolescPsychiatry 2000;39:1424–31.

11. Sturm H, Fernell E, Gillberg C. Autism spectrumdisorders in children with normal intellectuallevels: associated impairments and subgroups. DevMed Child Neurol 2004;46:444–7.

12. Biederman J, Faraone SV. Current concepts on theneurobiology of attention-deficit/hyperactivitydisorder. J Atten Disord 2002;6 (Suppl 1):s7–16.

13. Bradley JD, Golden CJ. Biological contributions tothe presentation and understanding of attention-

deficit/hyperactivity disorder: a review. Clin PsycholRev 2001;21:907–29.

14. World Health Organization. Behavioural andemotional disorders with onset usually occurringin childhood and adolescence. In The ICD-10classification of mental and behavioural disorders.Clinical descriptions and diagnostic guidelines. Geneva:World Health Organization; 1992. Chapter 5,pp. 260–6.

15. Clarke AR, Barry RJ, McCarthy R, Selikowitz M.Hyperkinetic disorder in the ICD-10: EEGevidence for a definitional widening? Eur ChildAdolesc Psychiatry 2003;12:92–9.

16. Foreman DM, Foreman D, Prendergast M, Minty B.Is clinic prevalence of ICD-10 hyperkinesisunderestimated? Impact of increasing awarenessby a questionnaire screen in an UK clinic. EurChild Adolesc Psychiatry 2000;10:130–4.

17. Mota VL, Schachar RJ. Reformulating attention-deficit/hyperactivity disorder according to signaldetection theory. J Am Acad Child Adolesc Psychiatry2000;39:1144–51.

18. Hartman CA, Hox J, Mellenbergh GJ, Boyle MH,Offord DR, Racine Y, et al. DSM-IV internalconstruct validity: when a taxonomy meets data.J Child Psychol Psychiatry 2001;42:817–36.

19. Guardiola A, Fuchs FD, Rotta NT. Prevalence ofattention-deficit hyperactivity disorders instudents. Comparison between DSM-IV andneuropsychological criteria. Arq Neuropsiquiatr2000;58(2-B):401–7.

20. Burns GL, Boe B, Walsh JA, Sommers-Flanagan R,Teegarden LA. A confirmatory factor analysis onthe DSM-IV ADHD and ODD symptoms: what isthe best model for the organization of thesesymptoms? J Abnorm Child Psychol 2001;29:339–49.

21. Byrne JM, Bawden HN, Beattie TL, DeWolfe NA.Preschoolers classified as having attention-deficithyperactivity disorder (ADHD): DSM-IV symptomendorsement pattern. J Child Neurol1999;15:533–8.

22. Rohde LA. ADHD in Brazil: the DSM-IV criteriain a culturally different population. J Am AcadChild Adolesc Psychiatry 2002;41:1131–3.

23. Graetz BW, Sawyer MG, Hazell PL, Arney F,Baghurst P. Validity of DSM-IVADHD subtypes in anationally representative sample of Australianchildren and adolescents. J Am Acad Child AdolescPsychiatry 2001;40:1410–17.


133


References

24. Crystal DS, Ostrander R, Chen RS, August GJ.Multimethod assessment of psychopathologyamong DSM-IV subtypes of children withattention-deficit/hyperactivity disorder: self-,parent, and teacher reports. J Abnorm Child Psychol2001;29:189–205.

25. Datapharm Communications. ElectronicMedicines Compendium (eMC): summary ofproduct characteristics. URL:http://emc.medicines.org.uk/. Accessed 18November 2004.

26. Janssen-Cilag. Putting patients at the heart ofeverything we do. High Wycombe: Janssen-CilagLimited: Bucks; 2004. URL: http://www.janssen-cilag.co.uk/index.asp. Accessed 9 December 2004.

27. Healthyplace. Atomoxetine HCl. Brand name:Strattera. 2002. URL:http://www.healthyplace.com/medications/strattera.htm. Accessed 9 December 2004.

28. Jadad AR, Boyle M, Cunningham C, Kim M,Schachar R. Treatment of attention-deficit/hyperactivitydisorder. Evidence Report: Technology Assessment No. 11. Rockville, MD: Agency for Health CarePolicy and Research; 1999.

29. Joint Formulary Committee. British NationalFormulary. 47th ed. London: British MedicalAssociation and the Royal Pharmaceutical Societyof Great Britain; 2004.

30. Canadian Coordinating Office for HealthTechnology Assessment (CCOHTA). A review oftherapies for attention-deficit/hyperactivity disorder –systematic review. Ottawa: Canadian CoordinatingOffice for Health Technology, Assessment(CCOHTA), 1998. URL: http://www.ccohta.ca

31. Centre for Reviews and Dissemination. CRD ReportNo. 4. Undertaking systematic reviews of research oneffectiveness: CRD’s guidance for those carrying out orcommissioning reviews. 2nd ed. York: Centre forReviews and Dissemination; 2001.

32. Swanson JM, Wigal SB, Wigal T, Sonuga-Barke E,Greenhill LL, Biederman J, et al. A comparison ofonce-daily extended-release methylphenidateformulations in children with attention-deficit/hyperactivity disorder in the laboratoryschool (the Comacs Study). Pediatrics2004;113:e206–16.

33. Elbourne DR, Altman DG, Higgins JP, Curtin F,Worthington HV, Vail A. Meta-analyses involvingcross-over trials: methodological issues. Int JEpidemiol 2002;31:140–9.

34. Drummond M, Jefferson T. Guidelines for authorsand peer reviewers of economic submissions toBMJ. BMJ 1996;313:275–83.

35. Ahmann PA, Waltonen SJ, Olson KA, Theye FW,van Erem AJ, LaPlant RJ. Placebo-controlledevaluation of ritalin side effects. Pediatrics1993;91:1101-6.

36. Arnold LE, Huestis RD, Smeltzer DJ, Scheib J,Wemmer D, Colner G. Levoamphetamine vsdextroamphetamine in minimal brain dysfunction.Replication, time response, and differential effectby diagnostic group and family rating. Arch GenPsychiatry 1976;33:292–301.

37. Arnold LE, Christopher J, Huestis R, Smeltzer DJ.Methylphenidate vs dextroamphetamine vscaffeine in minimal brain dysfunction: controlledcomparison by placebo washout design with Bayes’analysis. Arch Gen Psychiatry 1978;35:463–73.

38. Arnold LE, Kleykamp D, Votolato NA, Taylor WA,Kontras SB, Tobin K. Gamma-linolenic acid forattention-deficit hyperactivity disorder: placebo-controlled comparison to D-amphetamine. BiolPsychiatry 1989;25:222–8.

39. Barkley RA, McMurray MB, Edelbrock CS,Robbins K. Side effects of methylphenidate inchildren with attention deficit hyperactivitydisorder: a systematic, placebo-controlledevaluation. Pediatrics 1990;86:184–92.

40. Barkley RA, Connor DF, Kwasnik D. Challenges todetermining adolescent medication response in anoutpatient clinical setting: comparing Adderalland methylphenidate for ADHD. J Atten Disord2000;4:102–13.

41. Brown RT, Wynne ME, Medensis R.Methylphenidate and cognitive therapy: acomparison of treatment approaches withhyperactive boys. J Abnorm Child Psychol1985;13:69–87.

42. Brown RT, Wynne ME, Boren KA, Clingerman SR,Geniesse R, Spunt AL. Methylphenidate andcognitive therapy in children with attention-deficitdisorder: a double-blind trial. J Dev Behav Pediatr1986;7:163–74.

43. Brown RT, Sexton SB. A controlled trial ofmethylphenidate in black adolescents. Attentional,behavioural, and physiological effects. Clin Pediatr1988;27:74–81.

44. Buitelaar JK, van der Gaag R, Swaab-Barneveld H,Kuiper M. Pindolol and methylphenidate inchildren with attention-deficit hyperactivitydisorder: clinical efficacy and side-effects. J ChildPsychol Psychiatry 1996;37:587–95.

45. Conners CK, Taylor E, Meo G, Kurtz MA, FournierM. Magnesium pemoline anddextroamphetamine: a controlled study inchildren with minimal brain dysfunction.Psychopharmacologia 1972;26:321–36.

46. Conners CK, Taylor E. Pemoline,methylphenidate, and placebo in children withminimal brain dysfunction. Arch Gen Psychiatry1980;37:922–30.

47. Conrad WG, Dworkin ES, Shai A, Tobiessen JE.Effects of amphetamine therapy and prescriptive

References

134

tutoring on the behavior and achievement of lowerclass hyperactive children. J Learn Disabil1971;4:509–17.

48. Dopfner M, Bonaschewski T, Schmidt J, Uebel H,Schmeck K, Gerber WD, et al. Long-actingmethylphenidate preparation in children withADHD – a multicenter study. Nervenheilkd2003;22:85–92.

49. DuPaul GJ, Rapport MD. Does methylphenidatenormalize the classroom performance of childrenwith attention deficit disorder? J Am Acad ChildAdolesc Psychiatry 1993;32:190–9.

50. Efron D, Jarman F, Barker M. Methylphenidateversus dexamphetamine in children with attentiondeficit disorder: a double-blind, crossover trial.Pediatrics 1997;100:e6.

51. Elia J, Borcherding BG, Rapoport JL, Keysor CS.Methylphenidate and dextroamphetaminetreatments of hyperactivity: are there truenonresponders? Psychiatry Res 1991;36:141–55.

52. Fine S, Johnston C. Drug and placebo side effectsin methylphenidate–placebo trial for attentiondeficit hyperactivity disorder. Child Psychiatry HumDev 1993;4:25–30.

53. Firestone P, Crowe D, Goodman JT, McGrath P.Vicissitudes of follow-up studies: differential effectsof parent training and stimulant medication withhyperactives. Am J Orthopsychiatry 1986;56:184–94.

54. Fischer M, Newby RF. Assessment of stimulantresponse in ADHD children using a refinedmultimethod clinical protocol. J Clin Child Psychol1991;20:232–44.

55. Fitzpatrick PA, Klorman R, Brumaghim JT,Borgstedt AD. Effects of sustained-release andstandard preparations of methylphenidate onattention deficit disorder. J Am Acad Child AdolescPsychiatry 1992;31:226–34.

56. Gillberg C, Melander H, von Knorring AL, Janols LO, Thernlund G, Hagglof B, et al. Long-term stimulant treatment of children withattention-deficit hyperactivity disorder symptoms.A randomized, double-blind, placebo-controlledtrial. Arch Gen Psychiatry 1997;54:857–64.

57. Gittelman-Klein R, Klein DF, Katz S, Saraf K,Pollack E. Comparative effects of methylphenidateand thoridazine in hyperkinetic children. I.Clinical results. Arch Gen Psychiatry1976;33:1217–31.

58. Greenberg LM, Deem MA, McMahon S. Effects ofdextroamphetamine, chlorpromazine, andhydroxyzine on behavior and performance inhyperactive children. Am J Psychiatry1972;129:532–9.

59. Greenhill LL, Findling RL, Swanson JM, GroupAS. A double-blind, placebo-controlled study ofmodified-release methylphenidate in children with

attention-deficit/hyperactivity disorder. Pediatrics2002;109:E39.

60. Handen B, Feldman HM, Lurier A, Murray PJ.Efficacy of methylphenidate among preschoolchildren with developmental disabilities andADHD. J Am Acad Child Adolesc Psychiatry1999;38:805–12.

61. Hoeppner JA, Hale JB, Bradley AM, Byrnes M,Coury DL, Lennie L, et al. A clinical protocol fordetermining methylphenidate dosage levels inADHD. J Attention Disord 1997;2:19–30.

62. James RS, Sharp WS, Bastain TM, Lee PP, Walter JM, Czarnolewski M, et al. Double-blind,placebo-controlled study of single-doseamphetamine formulations in ADHD. J Am AcadChild Adolesc Psychiatry 2001;40:1268–76.

63. Kelsey D, Sumner C, Casat CD, Coury DL,Quintana H, Saylor KE, et al. Once-dailyatomoxetine treatment for children with attention-deficit/hyperactivity disorder, including anassessment of evening and morning behaviour: adouble-blind, placebo controlled trial. Pediatrics2004;114:e1–8.

64. Kemner JE, Starr HL, Ciccone PE, Hooper-Wood CG, Crockett S. Treatment outcomes ofOROS(R) methylphenidate compared with atomoxetinein children with ADHD: a multicenter, randomized,prospective study. Washington, DC: McNeilConsumer and Specialty Pharmaceuticals; 2004.

65. Klein RG, Abikoff H. Behavior therapy andmethylphenidate in the treatment of children withADHD. J Atten Disord 1997;2:89–114.

66. Klorman R, Coons HW, Borgstedt AD. Effects ofmethylphenidate on adolescents with a childhoodhistory of attention deficit disorder: I. Clinicalfindings [published erratum appears in J Am AcadChild Adolesc Psychiatry 1987;26:820 (Ref. 304)].J Am Acad Child Adolesc Psychiatry 1987;26:363–7.

67. Klorman R, Brumaghim JT, Fitzpatrick PA,Borgstedt AD. Clinical effects of a controlled trialof methylphenidate on adolescents with attentiondeficit disorder. J Am Acad Child Adolesc Psychiatry1990;29:702–9.

68. Klorman R, Brumaghim JT, Fitzpatrick PA,Borgstedt AD, Strauss J. Clinical and cognitiveeffects of methylphenidate on children withattention deficit disorder as a function ofaggression/oppositionality and age. J AbnormPsychol 1994;103:206–21.

69. Kolko DJ, Bukstein OG, Barron J.Methylphenidate and behavior modification inchildren with ADHD and comorbid ODD or CD:main and incremental effects across settings. J AmAcad Child Adolesc Psychiatry 1999;38:578–86.

70. Kratochvil CJ, Heiligenstein JH, Dittmann R,Spencer TJ, Biederman J, Wernicke J, et al.Atomoxetine and methylphenidate treatment in


135


children with ADHD: a prospective, randomized,open-label trial. J Am Acad Child Adolesc Psychiatry2002;41:776–84.

71. Kupietz SS, Winsberg BG, Richardson E, Maitinsky S. Effects of methylphenidate dosage inhyperactive reading-disabled children: I. Behaviorand cognitive performance effects. J Am Acad ChildAdolesc Psychiatry 1988;27:70–7.

72. Manos MJ, Short EJ, Findling RL. Differentialeffectiveness of methylphenidate and Adderall® inschool-age youths with attention-deficit/hyperactivity disorder. J Am Acad ChildAdolesc Psychiatry 1999;38:813–19.

73. Michelson D, Faries D, Wernicke J, Kelsey D,Kendrick K, Sallee FR, et al. Atomoxetine in thetreatment of children and adolescents withattention-deficit/hyperactivity disorder: arandomized, placebo-controlled, dose–responsestudy. Pediatrics 2001;108:E83.

74. Michelson D, Allen AJ, Busner J, Casat C, Dunn D,Kratochvil C, et al. Once-daily atomoxetinetreatment for children and adolescents withattention deficit hyperactivity disorder: arandomized, placebo-controlled study. Am JPsychiatry 2002;159:1896–901.

75. Michelson D, Buitelaar JK, Danckaerts M, Gillberg C, Spencer TJ, Zuddas A, et al. Relapseprevention in pediatric patients with ADHDtreated with atomoxetine: a randomized, double-blind, placebo-controlled study. J Am Acad ChildAdolesc Psychiatry 2004;43:896–904.

76. Jensen PS, Arnold LE, Richters JE, Severe JB,Vereen D, Vitiello B, et al. A 14-month randomizedclinical trial of treatment strategies for attention-deficit/hyperactivity disorder. Arch Gen Psychiatry1999;56:1073–86.

77. Pelham WEJ, Sturges J, Hoza J, Schmidt C,Bijlsma JJ, Milich R, et al. Sustained release andstandard methylphenidate effects on cognitive andsocial behavior in children with attention deficitdisorder. Pediatrics 1987;80:491–501.

78. Pelham WEJ, Greenslade KE, Vodde-Hamilton M,Murphy DA, Greenstein JJ, Gnagy EM, et al.Relative efficacy of long-acting stimulants onchildren with attention deficit-hyperactivity disorder:a comparison of standard methylphenidate,sustained-release dextroamphetamine, andpemoline. Pediatrics 1990;86:226–37.

79. Pelham WEJ, Carlson C, Sams SE, Vallano G,Dixon MJ, Hoza B. Separate and combined effectsof methylphenidate and behaviour modificationon boys with attention deficit-hyperactivitydisorder in the classroom. J Consult Clin Psychol1993;61:506–15.

80. Pelham WE, Aronoff HR, Midlam JK, Shapiro CJ,Gnagy EM, Chronis AM, et al. A comparison ofRitalin and Adderall: efficacy and time-course in

children with attention-deficit/hyperactivitydisorder. Pediatrics 1999;103:e43.

81. Pelham WE, Gnagy EM, Chronis AM, Burrows-MacLean L, Fabiano GA, Onyango AN, et al. A comparison of morning-only andmorning/late afternoon Adderall to morning-only,twice-daily, and three times-daily methylphenidatein children with attention-deficit/hyperactivitydisorder. Pediatrics 1999;104:1300–11.

82. Pelham WE, Gnagy EM, Burrows-Maclean L,Williams A, Fabiano GA, Morrisey SM, et al. Once-a-day Concerta methylphenidate versus three-times-daily methylphenidate in laboratory andnatural settings. Pediatrics 2001;107:E105.

83. Pliszka SR, Browne RG, Olvera RL, Wynne SK. A double-blind, placebo-controlled study ofAdderall and methylphenidate in the treatment ofattention-deficit/hyperactivity disorder. J Am AcadChild Adolesc Psychiatry 2000;39:619–26.

84. Quinn D. Celltech integrated clincial study report. Amulti-centre, double-blind, three arm, parallel groupstudy comparing the efficacy of immediate releasemethylphenidate (Ritalin®) and modified releasemethylphenidate with placebo in children with attentiondeficit/hyperactivity disorder. Slough: CelltechPharmaceuticals; 2003.

85. Rapport MD, DuPaul GJ, Kelly KL. Attentiondeficit hyperactivity disorder andmethylphenidate: the relationship between grossbody weight and drug response in children.Psychopharmacol Bull 1989;25:285–90.

86. Schachar RJ, Tannock R, Cunningham C, CorkumPV. Behavioural, situational, and temporal effectsof treatment of ADHD with methylphenidate. J AmAcad Child Adolesc Psychiatry 1997;36:754–63.

87. Diamond IR, Tannock R, Schachar RJ. Responseto methylphenidate in children with ADHD andcomorbid anxiety. J Am Acad Child Adolesc Psychiatry1999;38:402–9.

88. Smith BH, Pelham WE, Evans S, Gnagy E, Molina B, Bukstein O, et al. Dosage effects ofmethylphenidate on the social behaviour ofadolescents diagnosed with attention-deficithyperactivity disorder. Exp Clin Psychopharmacol1998;6:187–204.

89. Spencer T, Heiligenstein JH, Biederman J,Faries DE, Kratochvil CJ, Conners CK, et al.Results from 2 proof-of-concept, placebo-controlled studies of atomoxetine in children withattention-deficit/hyperactivity disorder. J ClinPsychiatry 2002;63:1140–7.

90. Steele M, Weiss M, Swanson J, Wang J, Riccardelli R, Binder C. A randomized, controlled,effectiveness trial of OROS–methylphenidate compared tousual care with immediate-release-methylphenidate inattention-deficit-hyperactivity disorder. A submission tothe National Institute for Clinical Excellence.London, Ontario: Janssen-Ortho; 2004.

References

136

91. Stein MA, Blondis TA, Schnitzler ER, O’Brien T,Fishkin J, Blackwell B, et al. Methylphenidatedosing: twice daily versus three times daily.Pediatrics 1996;98:748–56.

92. Stein MA, Sarampote CS, Waldman ID, Robb AS,Conlon C, Pearl PL, et al. A dose–response study ofOROS methylphenidate in children with attention-deficit/hyperactivity disorder. Pediatrics2003;112:E404–13.

93. Tervo RC, Azuma S, Fogas B, Fiechtner H.Children with ADHD and motor dysfunctioncompared with children with ADHD only. Dev MedChild Neurol 2002;44:383–90.

94. Weiss M, Tannock R, Kratochvil C, Dunn D, Velez-Borras J, Thomason C, et al. A randomized,placebo-controlled study of once-daily atomoxetine in theschool setting in children with ADHD. A submission tothe National Institute for Clinical Excellence.Indianapolis, IN: Lilley Research Laboratories;2004.

95. Wernicke JF, Adler L, Spencer T, West SA, Allen AJ, Heiligenstein J, et al. Changes insymptoms and adverse events afterdiscontinuation of atomoxetine in children andadults with attention deficit/hyperactivity disorder:a prospective, placebo-controlled assessment. J Clin Psychopharmacol 2004;24:30–5.

96. Werry JS, Aman MG, Diamond E. Imipramine andmethylphenidate in hyperactive children. J ChildPsychol Psychiatry 1980;21:27–35.

97. Wolraich ML, Greenhill LL, Pelham W, Swanson J,Wilens T, Palumbo D, et al. Randomized,controlled trial of OROS methylphenidate once aday in children with attention-deficit/hyperactivitydisorder. Pediatrics 2001;108:883–92.

98. Zeiner P, Bryhn G, Bjercke C, Truyen K, Strand G.Response to methylphenidate in boys withattention-deficit hyperactivity disorder. ActaPaediatr 1999;88:298–303.

99. Kemner JE, Starr L, Ciccone PE, Hooper-Wood CG,Crockett S. Treatment outcomes of OROS®

methylphenidate compared with atomoxetine in childrenwith ADHD: a multicenter, randomized, prospectivestudy. A submission to the National Institute forClinical Excellence. Fort Washington, PA: McNeilConsumer and Speciality Pharmaceuticals; 2004.

100. Pliszka SR, Browne RG, Olvera RL, Wynne SK. A double-blind, placebo-controlled study of Adderall and methylphenidate in the treatment ofattention-deficit/hyperactivity disorder. J Am AcadChild Adolesc Psychiatry 2000;39:619–26.

101. Zeiner P. Do the beneficial effects of extendedmethylphenidate treatment in boys with attention-deficit hyperactivity disorder dissipate rapidlyduring placebo treatment? Nord J Psychiatry1999;53:55–66.

102. Albaret JM, Soppelsa R, Marquet-Doleac J.Interest of the Matching Figure Test in thediagnostic of the impulsivity component in theattention deficit/hyperactivity disorder:comparative study. Approche Neuropsychologique desApprentissages chez l’Enfant 2000;12:98–102.

103. Jensen PS. Fact versus fancy concerning themultimodal treatment study for attention-deficithyperactivity disorder. Can J Psychiatry1999;44:975–80.

104. Taylor E. Development of clinical services forattention-deficit/hyperactivity disorder. Arch GenPsychiatry 1999;56:1097–9.

105. Moderators and mediators of treatment responsefor children with attention-deficit/hyperactivitydisorder: the Multimodal Treatment Study ofchildren with attention-deficit/hyperactivitydisorder. Arch Gen Psychiatry 1999;56:1088–96.

106. Jensen PS, Hinshaw SP, Kraemer HC, Lenora N,Newcorn JH, Abikoff HB, et al. ADHDcomorbidity findings from the MTA study:comparing comorbid subgroups. J Am Acad ChildAdolesc Psychiatry 2001;40:147–58.

107. Jensen P, Hinshaw S, Swanson J, Greenhill L,Conners C, Arnold L, et al. Findings from theNIMH multimodal treatment study of ADHD(MPA): implications and applications for primarycare providers. J Dev Behav Pediatr 2001;22:60–73.

108. Vitiello B, Severe J, Greenhill L, Arnold L, Abikoff H, Bukstein O, et al. Methylphenidatedosage for children with ADHD over time undercontrolled conditions: lessons from the MTA. J AmAcad Child Adolesc Psychiatry 2001;40:188–96.

109. Greenhill LL, Swanson JM, Vitiello B, Davies M,Clevenger W, Wu M, et al. Impairment anddeportment responses to differentmethylphenidate doses in children with ADHD:the MTA titration trial. J Am Acad Child AdolescPsychiatry 2001;40:180–7.

110. Rieppi R, Greenhil lLL, Ford RE, Chuang S,Wu M, Davies M, et al. Socioeconomic status as amoderator of ADHD treatment outcomes. J AmAcad Child Adolesc Psychiatry 2002;41:269–77.

111. Arnold LE, Elliot M, Sachs L, Bird H, Kraemer HC, Wells KC, et al. Effects of ethnicityon treatment attendance, stimulant response/dose,and 14-month outcome in ADHD. J Consult ClinPsychol 2003;71:713–27.

112. Galanter CA, Carlson GA, Jensen PS, Greenhill LL, Davies M, Li W, et al. Response tomethylphenidate in children with attention deficithyperactivity disorder and manic symptoms in themultimodal treatment study of children withattention deficit hyperactivity disorder titrationtrial. J Child Adolesc Psychopharmacol2003;13:123–36.


137


113. MTA Cooperative Group. National Institute ofMental Health Multimodal Treatment Study ofADHD follow-up: 24-month outcomes of treatmentstrategies for attention-deficit/hyperactivitydisorder. Pediatrics 2004;113:754–61.

114. Rapport MD, Moffitt C. Attentiondeficit/hyperactivity disorder andmethylphenidate. A review of height/weight,cardiovascular, and somatic complaint side effects.Clin Psychol Rev 2002;22:1107–31.

115. Drummond M, O’Brien B, Stoddart G, TorranceG. Methods for the economic evaluation of health careprogrammes. 2nd ed. Oxford: Oxford MedicalPublications; 1997.

116. Markwardt FC. Peabody Individual Achievement Test(PIAT). Circle Pines, MN: American GuidanceService; 1998.

117. Pelham WE, Bender ME, Caddell J, Booth S,Moorer SH. Methylphenidate and children withattention deficit disorder. Arch Gen Psychiatry1985;42:948–52.

118. Conners CK, Epstein JN, March JS, Angold A,Wells KC, Klaric J, et al. Multimodal treatment ofADHD in the MTA: an alternative outcomeanalysis. J Am Acad Child Adolesc Psychiatry2001;40:159–67.

119. Rosser R, Cottee M, Rabin R, Selai C. Index ofhealth related quality of life. In Hopkins A, editor.Measures of the quality of life. London: Royal Collegeof Doctors; 1992.

120. Landgraf J, Abetz L, Ware J. The Child HealthQuestionnaire user’s manual. Boston, MA: HealthAct;1999.

121. Zwi M, Ramchandani P, Joughin C. Evidence andbelief in ADHD. BMJ 2000;321:975–6.

122. Dolan P, Gudex C, Kind P, Williams A. A social tarifffor the EurQol: results from a UK general populationsurvey. York: Centre for Health Economics,University of York; 1995.

123. Gilmore A, Milne R. Methylphenidate in childrenwith hyperactivity: review and cost–utility analysis.Pharmacoepidemiol Drug Saf 2001;10:85–94.

124. Gilmore A, Best L, Milne R. Methylphenidate inchildren with hyperactivity. Report No. 78.Southampton: Wessex Institute for HealthResearch and Development; 1998.

125. Matza LS, Secnik K, Rentz AM, Mannix S, Revicki DA, Sallee FR. Development andassessment of health state utilities for attentiondeficit/hyperactivity disorder in children usingparent proxy report. Value Health 2003;6:697–8.

126. Zupancic JAF, Miller A, Raina P, Lee SK, Klassen A, Olsen L. A review of therapies for attentiondeficit/hyperactivity disorder. Ottawa: CanadianCoordinating Office for Health TechnologyAssessment; 1998.

127. Shukla VK, Otten N. Assessment of attentiondeficit/hyperactivity disorder therapy: a Canadianperspective. Ottawa: Canadian Coordinating Officefor Health Technology Assessment (CCOHTA);1999. URL: http://www.ccohta.ca/

128. Marchetti A, Magar R, Lau H, Murphy EL, Jensen PS, Conners CK, et al. Pharmacotherapiesfor attention-deficit/hyperactivity disorder:expected-cost analysis. Clin Ther 2001;23:1904–21.

129. Vanoverbeke N, Annemans L, Ingham M,Adriaenssen I. A cost analysis of the managementof attention-deficit/hyperactivity disorder (ADHD)in children in the UK. J Med Econ 2003;6:79–94.

130. Barkley RA. Child behaviour rating scales andchecklists. In Rutter M, Tuma AH, Lann IS,editors. Assessment and diagnosis in childpsychopathology. New York: Guildford Press; 1988.

131. Monthly index of medical specialities (MIMS) [August].London: Haymarket Medical Publications, 1997.

132. Foster N. Comorbidity as a moderator of cost-effectiveness. York: University of York; 2004.

133. Swanson JM, Kraemer HC, Hinshaw SP,Arnold LE, Conners CK, Abikoff HB, et al. Clinicalrelevance of the primary findings of the MTA:success rates based on severity of ADHD and ODDsymptoms at the end of treatment. J Am Acad ChildAdolesc Psychiatry 2001;40:168–79.

134. National Institute for Clinical Excellence (NICE).Methylphenidate (Ritalin, Equasym) for attentiondeficit/hyperactivity disorder (ADHD) in childhood.London: NICE; 2000. URL:http://www.nice.org.uk/article.asp?a=11667

135. Netten A, Curtis L. Unit costs of health and socialcare. Canterbury: Personal Social Services ResearchUnit (PSSRU), University of Kent; 2003. URL:www.pssru.ac.uk

136. Coghill D, Spender Q, Barton J, Hollis C, Yuen C,Cleemput I, et al. Measuring quality of life inchildren with attention-deficit-hyperactivity-disorder in the United Kingdom. In 16th WorldCongress of the International Association of Child andAdolescent Psychiatry and Allied Professions(IACAPAP), 23–26 August 2004, Berlin. URL:http://www.iacapap-berlin.de/guest/?MIval=AbstractView&ABSID=4112

137. Fenwick E, Claxton K, Sculpher M. Representinguncertainty: the role of cost-effectivenessacceptability curves. Health Econ 2001;10:779–89.

138. Kauffman RE, Smith-Wright D, Reese CA,Simpson R, Jones F. Medication compliance inhyperactive children. Pediatr Pharmacol1981;1:231–7.

139. Dirksen SJ, D’Imperio JM, Birdsall D, Hatch SJ. A postmarketing clinical experience study ofMetadate CD. Curr Med Res Opin 2002;18:371–80.

References

138

140. Hill P, Taylor E. An auditable protocol for treatingattention deficit/hyperactivity disorder. Arch DisChild 2001;84:404–9.


142. Biederman J, Heiligenstein JH, Faries DE, Galil N,Dittmann R, Emslie GJ, et al. Efficacy ofatomoxetine versus placebo in school-age girlswith attention-deficit/hyperactivity disorder.Pediatrics 2002;110:e75.

143. Efron D, Jarman F, Barker M. Side effects ofmethylphenidate and dexamphetamine inchildren with attention deficit hyperactivitydisorder: a double-blind, crossover trial. Pediatrics1997;100:662–6.

144. Weiss M, Tannock R, Kratochvil C, Dunn D, Velez-Boras J, Thomason C, et al. Placebo-controlled study of once-daily atomoxetine in theschool setting. Eur Neuropsychopharmacol 2003;13(suppl 4):S456.

145. Secnik K, Cottrell S, Matza LS, Edgell E, Aristides M, Tilden D, et al. Assessment of healthstate utilities for attention-deficit/hyperactivitydisorder in children using parent-based standardgamble scores. In 9th Annual International Meetingof the International Society for Pharmacoeconomics andOutcomes Research (ISPOR), 16–19 May 2004,Arlington, VA.

146. Briggs AH. Handling uncertainty in cost-effectiveness models. Pharmacoeconomics 2000;17:479–500.

147. Spiegelhalter DJ, Thomas A, Best NG, Gilks WR.BUGS: Bayesian inference using Gibbs sampling,version 0.5, (version ii). Cambridge: MRCBiostatistics Unit; 1996.

148. Hill JC, Schoener EP. Age-dependent decline ofattention deficit hyperactivity disorder. Am JPsychiatry 1996;153:1143–6.

149. Sharp WS, Walter JM, Marsh WL, Ritchie GF,Hamburger SD, Castellanos FX. ADHD in girls:clinical comparability of a research sample. J AmAcad Child Adolesc Psychiatry 1999;38:40–7.

150. Higgins JPT, Whitehead A. Borrowing strengthfrom external trials in a meta-analysis. Stat Med1996;15:2733–49.

151. Ades AE, Cliffe S. Markov Chain Monte Carloestimation of a multi-parameter decision model:consistency of evidence and the accurateassessment of uncertainty. Med Decis Making2002;22:359–71.

152. Claxton AJ, Cramer J, Pierce C. A systematicreview of the associations between dose regimensand medication compliance. Clin Ther2001;23:1296–310.


154. Celltech Pharmaceuticals Limited. Appraisal ofdrugs used in attention deficit hyperactivity disorder. A submission to the National Institute for ClinicalExcellence. Slough: Celltech Pharmaceuticals; 2004.

155. Janssen-Cilag. For health technology appraisal:methylphenidate, atomoxetine and dexamphetamine forattention-deficit/hyperactivity disorder in children andadolescents, including review of existing guidancenumber 13. A submission to the National Institutefor Clinical Excellence. High Wycombe: Janssen-Cilag Limited; 2004.

156. Eli Lilly. Atomoxetine for the treatment of ADHD. Asubmission to the National Institute for ClinicalExcellence. Basingstoke: Eli Lilly and CompanyLimited; 2004.

157. National Institute for Health and ClinicalExcellence (NICE). National Institute for ClinicalExcellence methodological guidance: economicevaluations. London: NICE; 2004.

158. Biederman J, Mick E, Faraone S. Age-dependentdecline of symptoms of attention deficithyperactivity disorder: impact of remissiondefinition and symptom type. Am J Psychiatry2000;157:816–18.

159. Moher D, Schulz KF, Altman DG. The CONSORTstatement: revised recommendations forimproving the quality of reports of parallel-grouprandomised trials. Lancet 2001;357:1191–4.

160. Aggarwal A, Lillystone D. A follow-up pilot studyof objective measures in children with attentiondeficit hyperactivity disorder. J Paediatr ChildHealth 2000;36:134–8.

161. Akhondzadeh S, Tavakolian R, Davari-Ashtiani R,Arabgol F, Amini H. Selegiline in the treatment ofattention deficit hyperactivity disorder in children:a double blind and randomized trial. ProgNeuropsychopharmacol Biol Psychiatry 2003;27:841–5.

162. Akhondzadeh S, Mohammadi MR, Khademi M.Zinc sulfate as an adjunct to methylphenidate forthe treatment of attention deficit hyperactivitydisorder in children: a double blind andrandomized trial [ISRCTN64132371]. BMCPsychiatry 2004;4:9.

163. Allen AJ, Spencer TJ, Heiligenstein JH, Faries DE,Kelsey DK, Laws HF, et al. Safety and efficacy oftomoxetine for ADHD in two double-blind,placebo-controlled trials. Biol Psychiatry2001;49:325–35.

164. Allen AJ, Michelson D, Harder Donald R, Trapp NJ, Kelsey DK. Efficacy of atomoxetine inchildren and adolescents with ADHD. In 155thAnnual Meeting of the American Psychiatric Association,18–23 May 2002, Philadelphia, PA.


139


165. Aman MG, Langworthy KS. Pharmacotherapy forhyperactivity in children with autism and otherpervasive developmental disorders. J Autism DevDisord 2000;30:451–9.

166. Aman MG, Buican B, Arnold LE. Methylphenidatetreatment in children with borderline IQ andmental retardation: analysis of three aggregatedstudies. J Child Adolesc Psychopharmacol2003;13:29–40.

167. Baren M, Swanson JM, Wigal SB. Lack of effect ofdifferent breakfast conditions on thepharmacokinetics and efficacy of OROS®

methylphenidate HCl extended-release tablets inchildren with ADHD. Pediatr Res 2000;47:129.

168. Bedard AC, Ickowicz A, Tannock R.Methylphenidate improves Stroop naming speed,but not response interference, in children withattention deficit hyperactivity disorder. J ChildAdolesc Psychopharmacol 2002;12:301–9.

169. Bedard AC, Martinussen R, Ickowicz A, TannockR. Methylphenidate improves visual-spatialmemory in children with attention-deficit/hyperactivity disorder. J Am Acad ChildAdolesc Psychiatry 2004;43:260–8.

170. Berman T. Effects of methylphenidate on complexcognitive processing in attention-deficit hyperactivitydisorder. Thesis. Montreal: McGill University; 1998.

171. Berman T, Douglas VI, Barr RG. Effects ofmethylphenidate on complex cognitive processingin attention-deficit hyperactivity disorder. J AbnormPsychol 1999;108:90–105.

172. Biederman J. Effectiveness and safety of the once-daily OROS formulation of methylphenidate inadolescents with attention-deficit/hyperactivitydisorder. Eur Neuropsychopharmacol 2003;13:S448.

173. Biederman J. Evaluation of once-daily OROSmethylphenidate in children with attention-deficit/hyperactivity disorder: effects on sleep,appetite, and tics. Eur Neuropsychopharmacol2003;13 (suppl 4):S447–8.

174. Biederman J, Quinn D, Weiss M, Markabi S,Weidenman M, Edson K, et al. Efficacy and safetyof Ritalin LA, a new, once daily, extended-releasedosage form of methylphenidate, in children withattention deficit hyperactivity disorder. PaediatrDrugs 2003;5:833–41.

175. Caballero J, Nahata MC. Atomoxetinehydrochloride for the treatment of attention-deficit/hyperactivity disorder. Clin Ther2003;25:3065–83.

176. Chen J, Chen YY, Wang XM. Clinical study ontreatment of children attention deficithyperactivity disorder by jiangqian granule.Zhongguo Zhong Xi Yi Jie He Za Zhi2002;22:258–60.

177. Chronis AM, Pelham WEJR, Gnagy EM,Roberts JE, Aronoff HR. The impact of late-afternoon stimulant dosing for children withADHD on parent and parent–child domains. J ClinChild Adolesc Psychol 2003;32:118–26.

178. Connor DF. Preschool attention deficithyperactivity disorder: a review of prevalence,diagnosis, neurobiology, and stimulant treatment.J Dev Behav Pediatr 2002;23 (Suppl 1):S1–9.

179. Connor DF, Barkley RA, Davis HT. A pilot study ofmethylphenidate, clonidine, or the combination inADHD comorbid with aggressive oppositionaldefiant or conduct disorder. Clin Pediatr2000;39:15–25.

180. Cox DJ, Merkel RL, Penberthy JK, Kovatchev B,Hankin CS. Impact of methylphenidate deliveryprofiles on driving performance of adolescentswith attention-deficit/hyperactivity disorder: a pilotstudy. J Am Acad Child Adolesc Psychiatry2004;43:269–75.

181. Davidovitch M, Manning-Courtney P,Hartmann LA, Watson J, Lutkenhoff M,Oppenheimer S. The prevalence of attentionalproblems and the effect of methylphenidate inchildren with myelomenigocele. Pediatr Rehabil1999;3:29–35.

182. Denney CB, Rapport MD. Predictingmethylphenidate response in children with ADHD:theoretical, empirical, and conceptual models.J Am Acad Child Adolesc Psychiatry 1999;38:393–401.

183. Ding GA, Yu GH, Chen SF. Assessment on effect oftreatment for childhood hyperkinetic syndrome bycombined therapy of yizhi mixture and ritalin.Zhongguo Zhong Xi Yi Jie He Za Zhi 2002;22:255–7.

184. Donnelly M, Rapoport JL, Potter WZ, Oliver J,Keysor CS, Murphy DL. Fenfluramine anddextroamphetamine treatment of childhoodhyperactivity: clinical and biochemical findings.Arch Gen Psychiatry 1989;46:205–12.

185. Duggan CM, Mitchell G, Nikles CJ, Glasziou PP,Del Mar CB, Clavarino A. Managing ADHD ingeneral practice. N of 1 trials can help! Aust FamPhysician 2000;29:1205–9.

186. Ebell M. How safe and effective is short-actingmethylphenidate for the treatment of attention-deficit disorder in children and adolescents? EvidBased Pract 2002;5:9.

187. Eiland LS, Guest AL. Atomoxetine treatment ofattention-deficit/hyperactivity disorder. AnnPharmacother 2004;38:86–90.

188. Firestone P, Musten LM, Pisterman S, Mercer J,Bennett S. Short-term side effects of stimulantmedication are increased in preschool childrenwith attention-deficit/hyperactivity disorder: adouble-blind placebo-controlled study. J ChildAdolesc Psychopharmacol 1998;8:13–25.

References

140

189. Francis S, Fine J, Tannock R. Methylphenidateselectively improves story retelling in children withattention deficit hyperactivity disorder. J ChildAdolesc Psychopharmacol 2001;11:217–28.

190. Gadow KD, Nolan E, Sverd J, Sprafkin J, Paolicelli L. Methylphenidate in aggressive-hyperactive boys: I. Effects on peer aggression inpublic school settings. J Am Acad Child AdolescPsychiatry 1990;29:710–18.

191. Gadow KD, Sverd J, Sprafkin J, Nolan EE,Grossman S. Long-term methylphenidate therapyin children with comorbid attention-deficithyperactivity disorder and chronic multiple ticdisorder. Arch Gen Psychiatry 1999;56:330–6.

192. Gadow KD, Nolan EE, Sverd J, Sprafkin J,Schwartz J. Anxiety and depression symptoms andresponse to methylphenidate in children withattention-deficit hyperactivity disorder and ticdisorder. J Clin Psychopharmacol 2002;22:267–74.

193. Goldberg J. Clonidine and methylphenidate wereeffective for attention deficit hyperactivity disorderin children with comorbid tics. Evid Based MentHealth 2002;5:122.

194. Goldberg J. Clonidine and methylphenidate wereeffective for attention deficit hyperactivity disorderin children with comorbid tics. ACP J Club2002;137:70.

195. Grcevich S. Review: Adderall may have a smalladvantage over methylphenidate in attentiondeficit hyperactivity disorder. Evid Based MentHealth 2003;6:43.

196. Greenhill LL. Efficacy and safety of once-dailymethylphenidate HCl, standard methylphenidateand placebo in children with ADHD. In 153rdAnnual Meeting of the American Psychiatric Association,13–18 May 2002, Chicago, IL.

197. Greenhill LL. Evaluation of the efficacy and safetyof Concerta (methylphenidate HCl) extended-release tablets, ritalin, and placebo in childrenwith ADHD. Neurology 2000;54 (suppl 7):A420–1.

198. Greenhill LL, Halperin JM, Abikof H. Stimulantmedications. J Am Acad Child Adolesc Psychiatry1999;38:503–12.

199. Greenhill LL, Pliszka S, Dulcan MK, Bernet W,Arnold V, Beitchman J, et al. Practice parameterfor the use of stimulant medications in thetreatment of children, adolescents, and adults.J Am Acad Child Adolesc Psychiatry 2002;41 (2 Suppl):26S–49S.

200. Gross-Tsur V, Shalev RS, Badihi N, Manor O.Efficacy of methylphenidate in patients withcerebral palsy and attention-deficit hyperactivitydisorder (ADHD). J Child Neurol 2002;17:863–6.

201. Handen BL, Johnson CR, Lubetsky M. Efficacy ofmethylphenidate among children with autism and

symptoms of attention-deficit hyperactivitydisorder. J Autism Dev Disord 2000;30:245–55.

202. Heiligenstein JH, Biederman J, Christopher J,Galil N, West Scott A, Emslie Graham J. Efficacy oftomoxetine versus placebo in school-age girls withADHD. In 155th Annual Meeting of the AmericanPsychiatric Association, 18–23 May 2002;Philadelphia, PA.

203. Heiligenstein JH, Biederman J, Kratochvil CJ,Galil N, West SA, Emslie GJ. Efficacy oftomoxetine versus placebo in school-age girls withADHD. In Annual Meeting of the American PsychiatricAssociation 5–10 May 2001, New Orleans, LA.

204. Heiligenstein JH, Dunn DW, Busner J, Stein M,McDougle CJ. Efficacy of tomoxetine versusplacebo in school-age children with ADHD whofailed psychostimulant treatment. In AnnualMeeting of the American Psychiatric Association5–10 May 2001, New Orleans, LA.

205. Heiligenstein JH, Wagner KD, Casat CD, Busner J,Newcorn JH, Saylor KE, et al. Efficacy oftomoxetine versus placebo in school-age childrenwith ADHD: inattentive subtype. In Annual Meetingof the American Psychiatric Association 5–10 May2001, New Orleans, LA.

206. Hoffman M, Pelham WE, Burrows-MacLean L.Dose-ranging studies demonstrating efficacy ofonce-daily transdermal methylphenidate, time-course of action, and effect in combination withbehavioral modification in pediatric patients withADHD. Pediatr Res 2003;53:556A.

207. Jadad AR, Boyle M, Cunningham C.Pharmacological interventions are more effectivethan non-pharmacological for attention deficithyperactivity disorder. Evid Based Med 2000;5:179.

208. Kent MA, Camfield CS, Camfield PR. Double-blind methylphenidate trials – practical, useful,and highly endorsed by families. Arch PediatrAdolesc Med 1999;153:1292–6.

209. Klein C, Fischer B, Fischer B, Hartnegg K. Effectsof methylphenidate on saccadic responses inpatients with ADHD. Exp Brain Res2002;145:121–5.

210. Kollins SH, MacDonald EK, Rush CR. Assessingthe abuse potential of methylphenidate innonhuman and human subjects: a review.Pharmacol Biochem Behav 2001;68:611–27.

211. Konrad K, Gunther T, Hanisch C, Herpertz-Dahlmann B. Differential effects ofmethylphenidate on attentional functions inchildren with attention-deficit/hyperactivitydisorder. J Am Acad Child Adolesc Psychiatry2004;43:191–8.

212. Kurlan R, Goldberg J. Clonidine andmethylphenidate were effective for attentiondeficit hyperactivity disorder in children withcomorbid tics. Evid Based Med 2002;7:157.


141


213. Lewis D, Linder S, Kurlan R, Winner P, Dunn D,Sallee FR, et al. Atomoxetine for the treatment ofattention deficit hyperactivity disorder andcomorbid tics in children. Ann Neurol 2003;54(Suppl 7):S106.

214. Li X, Chen Z. Clinical comparative observation onduodongning and Ritalin in treating childhyperkinetic syndrome. Zhongguo Zhong Xi Yi JieHe Za Zhi 1999;19:410–11.

215. Lieberman SG, Christophersen ER. The effect ofan afternoon dose of methylphenidate on theaccuracy and on-tack behavior of the homeworkcompleted by children with attention deficithyperactivity disorder. Pediatr Res 2000;47:29A.

216. Loo SK, Specter E, Smolen A, Hopfer C, Teale PD, Reite ML. Functional effects of theDAT1 polymorphism on EEG measures in ADHD.J Am Acad Child Adolesc Psychiatry 2003;42:986–93.

217. Lopez F, Silva R, Pestreich L, Muniz R.Comparative efficacy of two once dailymethylphenidate formulations (Ritalin LA andConcerta) and placebo in children with attentiondeficit hyperactivity disorder across the school day.Paediatr Drugs 2003;5:545–55.

218. Malone MA, Rosenfield JD, Roberts SW.Methylphenidate (Ritalin) effects on the grapho-motor artwork of children with attention deficithyperactivity disorder. Canadian Art TherapyAssociation Journal 2002;15:7–13.

219. Mannuzza S, Klein RG, Moulton JL III. Doesstimulant treatment place children at risk for adultsubstance abuse? A controlled, prospective follow-up study. J Child Adolesc Psychopharmacol2003;13:273–82.

220. Manos MJ, Short EJ, Findling RL. Dose responsecurves across ADHD subtypes: differential effectsbetween Adderall and methylphenidate. PediatrRes 2000;47:172.

221. McBurnett K. Treatment of adolescent ADHD withOROS® methylphenidate. Pediatr Res2003;53:3143.

222. Michelson D, Faries D, Wernicke J, Spencer T.Atomoxetine in the treatment of children withADHD: a randomized, placebo-controlled, dose-response study. Eur Neuropsychopharmacol2001;11:S319.

223. Michelson D, Faries DE, Wernicke JF, Spencer T.Atomoxetine in the treatment of children withADHD: a randomized, placebo-controlled dose-response study. Eur Neuropsychopharmacol2001;11:S319–20.

224. Mohammadi MR, Kashani L, Akhondzadeh S,Izadian ES, Ohadinia S. Efficacy of theophyllinecompared to methylphenidate for the treatment ofattention-deficit hyperactivity disorder in children

and adolescents: a pilot double-blind randomizedtrial. J Clin Pharm Ther 2004;29:139–44.

225. Montiel Nava C, Pena JA, Espina Marines G,Ferrer-Hernandez ME, Lopez-Rubio A, Puertas-Sanchez S, et al. A pilot study of methylphenidateand parent training in the treatment of childrenwith attention-deficit hyperactivity disorder. RevNeurol 2002;35:201–5.

226. Newcorn JH. Atomoxetine in the treatment ofchildren and adolescents with attention-deficit/hyperactivity disorder: a randomized,placebo-controlled dose–response study. CurrPsychiatry Rep 2003;5:85.

227. Nolan EE, Gadow KD, Sprafkin J. Stimulantmedication withdrawal during long-term therapyin children with comorbid attention-deficithyperactivity disorder and chronic multiple ticdisorder. Pediatrics 1999;103:730–7.

228. Overtoom CC, Verbaten MN, Kemner C,Kenemans JL, van Engeland H, Buitelaar JK, et al.Effects of methylphenidate, desipramine, and L-dopa on attention and inhibition in children withattention deficit hyperactivity disorder. Behav BrainRes 2003;145:7–15.

229. Palumbo DR, Starr HL. Impact of ADHDtreatment with a once-daily OROS® formulation ofmethylphenidate on tics. Pediatr Res 2003;53:555A–6A.

230. Pearson DA, Santos CW, Roache JD, Casat CD,Loveland KA, Lachar D, et al. Treatment effects ofmethylphenidate on behavioral adjustment inchildren with mental retardation and ADHD. J AmAcad Child Adolesc Psychiatry 2003;42:209–16.

231. Pearson DA, Lane DM, Santos CW, Casat CD,Jerger SW, Loveland KA, et al. Effects ofmethylphenidate treatment in children withmental retardation and ADHD: individualvariation in medication response. J Am Acad ChildAdolesc Psychiatry 2004;43:686–98.

232. Pearson DA, Santos CW, Casat CD, Lane DM,Jerger SW, Roache JD, et al. Treatment effects ofmethylphenidate on cognitive functioning inchildren with mental retardation and ADHD. J AmAcad Child Adolesc Psychiatry 2004;43:677–85.

233. Pelham WE, Hoza B, Pillow DR, Gnagy EM, KippHL, Greiner AR, et al. Effects of methylphenidateand expectancy on children with ADHD: behavior,academic performance, and attributions in asummer treatment program and regular classroomsettings. J Consult Clin Psychol 2002;70:320–35.

234. Rapport MD, Randall R, Moffitt C. Attention-deficit/hyperactivity disorder andmethylphenidate: a dose–response analysis andparent-child comparison of somatic complaints.J Attention Disord 2002;6:15–24.

References

142

235. Rhodes SM, Coghill DR, Matthews K.Neuropsychological profile and response tomethylphenidate in girls with hyperkineticdisorder (ADHD). J Psychopharmacol 2003;17 (3Suppl):A73.

236. Riyad CE, Skowronsk iGA, Turkall RM, Abdel-Rahman MS. A study of safety factors for riskassessment of drugs used for treatment ofattention deficiency hyperactivity disorder insensitive populations. Hum Ecol Risk Assess2002;8:823–40.

237. Rubia K, Noorloos J, Smith A, Gunning B,Sergeant J. Motor timing deficits in communityand clinical boys with hyperactive behavior: theeffect of methylphenidate on motor timing.J Abnorm Child Psychol 2003;31:301–13.

238. Schachter HM, Pham B, King J, Langford S,Moher D. How efficacious and safe is short-actingmethylphenidate for the treatment of attention-deficit disorder in children and adolescents? Ameta-analysis. CMAJ 2001;165:1475–88.

239. Scheres A, Oosterlaan J, Swanson J, Morein-ZamirS, Meiran N, Schut H, et al. The effect ofmethylphenidate on three forms of responseinhibition in boys with AD/HD. J Abnorm ChildPsychol 2003;31:105–20.

240. Sharp WS, Walter JM, Marsh WL, Ritchie GF,Hamburger SD, Castellanos FX. ADHD in girls:clinical comparability of a research sample. InHertzig ME, Farber EA, editors. Annual progress inchild psychiatry and child development 2000–2001.New York: Brunner Routledge; 2003. pp. 263–77.

241. Shaughnessy A. Does Adderall produce similarresults as methylphenidate in children withattention deficit-hyperactivity disorder? Evid BasedPract 1999;2:7.

242. Smith BH, Waschbusch DA, Willoughby MT, EvansS. The efficacy, safety, and practicality oftreatments for adolescents with attention-deficit/hyperactivity disorder (ADHD). Clin ChildFam Psychol Rev 2000;3:243–67.

243. Smith R, Larsen D, Derby KM, McLaughlin TF,Weber KP, Brown K, et al. A comparison of teacherchecklists used over 15 days and a one-dayantecedent analysis to conduct a medication trial.Psychol Schools 2004;41:235–40.

244. Spencer TJ. Atomoxetine treatment in childrenwith attention-deficit/hyperactivity disorder andcomorbid tic disorders. Biol Psychiatry 2004;55 (1S-242S):185.

245. Sunohara GA, Malone MA, Rovet J, Humphries T,Roberts W, Taylor MJ. Effect of methylphenidateon attention in children with attention deficithyperactivity disorder (ADHD): ERP evidence.Neuropsychopharmacology 1999;21:218–28.

246. Swanson JM, Wigal SB, Lemer MA. Comparison ofthe efficacy and safety of OROS methylphenidateHCl with methylphenidate tid and placebo inchildren with ADHD. Pediatr Res 2000;47:34A.

247. Swanson JM. Evaluation of differentmethylphenidate (MPH) delivery patterns for thetreatment of children with ADHD. Neurology2000;54 (Suppl 7):A420.

248. Swanson J, Wigal S, Greenhill L, Browne R,Waslick B, Lerner M, et al. Objective and subjectivemeasures of the pharmacodynamic effects ofAdderall® in the treatment of children with ADHDin a controlled laboratory classroom setting.Psychopharmacol Bull 1998;34:55–60.

249. Swanson JM, Wigal S, Greenhill LL, Browne R,Waslik B, Lerner M, et al. Analog classroomassessment of Adderall® in children with ADHD. J Am Acad Child Adolesc Psychiatry 1998;37:519–26.

250. Swanson JM, Gupta S, Williams L, Agler D,Lerner M, Wigal S. Efficacy of a new pattern ofdelivery of methylphenidate for the treatment ofADHD: effects on activity level in the classroomand on the playground. J Am Acad Child AdolescPsychiatry 2002;41:1306–14.

251. Swanson J, Gupta S, Guinta D, Flynn D, Agler D,Lerner M, et al. Acute tolerance tomethylphenidate in the treatment of attentiondeficit hyperactivity disorder in children. ClinPharmacol Ther 1999;66:295–305.

252. Swanson J, Sadeh A, Lerner MA, Wigal SB.Comparison of the impact of OROSmethylphenidate HCI with methylphenidate tidand placebo on the sleep of children with ADHD.J Dev Behav Pediatr 2000;21:387–8.

253. Swanson J, Wigal S, Lerner M. Treatment with acontrolled-release formulation of methylphenidatefor attention-deficit/hyperactivity disorder: onsetand duration of effect. Eur Neuropsychopharmacol2002;12 (Suppl 3):S414.

254. Szobot CM, Ketzer C, Cunha RD, Parente MA,Langleben DD, Acton PD, et al. The acute effect ofmethylphenidate on cerebral blood flow in boyswith attention-deficit/hyperactivity disorder. Eur JNucl Med Mol Imaging 2003;30:423–6.

255. Tenreiro K. Methylphenidate–placebo: a trial forattention deficit disorders. International Journal ofPharmaceutical Computing 2001;5:21–2.

256. The Tourette’s Syndrome Study Group. Treatmentof ADHD in children with tics: a randomizedcontrolled trial.comment. Neurology2002;58:527–36.

257. Tillery KL, Katz J, Keller WD. Effects ofmethylphenidate (Ritalan) on auditoryperformance in children with attention andauditory processing disorders. J Speech Lang HearRes 2000;43:893–901.


143


258. van der Meere J, Gunning B, Stemerdink N. Theeffect of methylphenidate and clonidine onresponse inhibition and state regulation inchildren with ADHD. J Child Psychol Psychiatry1999;40:291–8.

259. Wernicke JF, Fouche DS, Spencer T, Dunn D,Faries DE, Kelsey DK, et al. Safety of atomoxetinein pediatric patients treated for ADHD. EurNeuropsychopharmacol 2001;11 (Suppl 3):S319.

260. Wernicke JF, Dunn D, Faries DE, Heiligenstein JH,Kelsey DK, Kendrick KL, et al. Safety ofatomoxetine in placebo-controlled pediatricattention-deficit hyperactivity disorder trials. AnnNeurol 2001;50 (Suppl 3):S123–4.

261. Wernicke JF, Allen AJ, Faries D, Heiligenstein JH,Kelsey D, Kendrick K, et al. Safety of tomoxetinein clinical trials. Biol Psychiatry 2001;49:159S.

262. Whalen CK, Henker B, Buhrmester D, HinshawSP, Huber A, Laski K. Does stimulant medicationimprove the peer status of hyperactive children?J Consult Clin Psychol 1989;57:545–9.

263. Wigal SB. OROS® formulation of methylphenidatein treatment for ADHD: duration of effect. PediatrRes 2002;51:120.

264. Wigal SB, Swanson JM, Greenhill L, Waslick B,Cantwell D, Clevenger W, et al. Evaluation ofindividual subjects in the analog classroom setting:II. Effects of dose of amphetamine (Adderall®).Psychopharmacol Bull 1998;34:833–8.

265. Wigal SB, Sanchez DY, DeCory HH, D’ImperioJM, Swanson JM. Selection of the optimal doseratio for a controlled-delivery formulation ofmethylphenidate. J Appl Res 2003;3:46–63.

266. Wigal S, Lerner M, Swanson J. Once-dailymethylphenidate formulation: impact on academicproductivity and activity levels of children withattention-deficit/hyperactivity disorder. EurNeuropsychopharmacol 2002;12 (Suppl 3):S416.

267. Wigal T, Swanson JM, Regino R, Lerner MA,Soliman I, Steinhoff K, et al. Stimulantmedications for the treatment of ADHD: efficacyand limitations. Ment Retard Dev Disab Res Rev1999;5:215–24.

268. Wilens TE. Long-term safety and effectiveness ofConcerta (methylphenidate HCI) in children withADHD. Neurology 2000;54 (Suppl 7):A420.

269. Wilens TE, Faraone SV, Biederman J,Gunawardene S. Does stimulant therapy ofattention-deficit/hyperactivity disorder beget latersubstance abuse? A meta-analytic review of theliterature. Pediatrics 2003;111:179–85.

270. Wilens TE, Biederman J, Lerner M. Effects ofonce-daily osmotic-release methylphenidate onblood pressure and heart rate in children withattention-deficit/hyperactivity disorder. Resultsfrom a one-year follow-up study. J ClinPsychopharmacol 2004;24:36–41.

271. Wolraich ML. Evaluation of efficacy and safety ofOROS® methylphenidate HCl (MPH) extended-release tablets, methylphenidate tid, and placeboin children with ADHD. Pediatr Res 2000;47:207.

272. Amery B, Minichiello MD, Brown GI. Aggressionin hyperactive boys: response to D-amphetamine. J Am Acad Child Adolesc Psychiatry 1984;23:291–4.

273. Barrickmann LL, Perry PJ, Allen AJ, Kuperman S,Arndt SV, Herrmann KJ, et al. Bupropion versusmethylphenidate in the treatment of attention-deficit hyperactivity disorder. J Am Acad ChildAdolesc Psychiatry 1995;34:649–57.

274. Castellanos FX, Giedd JN, Elia J, Marsh WL,Ritchie GF, Hamburger SD, et al. Controlledstimulant treatment of ADHD and comorbidTourette’s syndrome: effects of stimulant and dose.J Am Acad Child Adolesc Psychiatry 1997;36:589–96.

275. Donnelly C, Faries D, Swensen A, Ruff D,Michelson D, Matza I, et al. The effect ofatomoxetine on the social and family functioningof children and adolescents with attention-deficit/hyperactivity disorder (ADHD). EurNeuropsychopharmacol 2002;12 (Suppl 3):S437.

276. Gadow KD, Nolan E, Sverd J. Methylphenidate inhyperactive boys with comorbid tic disorder: II.Short-term behavioural effects in school settings.J Am Acad Child Adolesc Psychiatry 1992;31:462–71.

277. Gadow KD, Nolan E, Sprafkin J, Sverd J. Schoolobservations of children with attention-deficithyperactivity disorder and comorbid tic disorder:effects of methylphenidate treatment. J Dev BehavPediatr 1995;16:167–76.

278. Gadow KD, Sverd J, Sprafkin J, Nolan EE, Ezor SN. Efficacy of methylphenidate forattention-deficit hyperactivity disorder in childrenwith tic disorder. Arch Gen Psychiatry1995;52:444–55.

279. Gadow KD, Sverd J, Sprafkin J, Nolan EE, EzorSN. Erratum: Efficacy of methylphenidate forattention-deficit hyperactivity disorder in childrenwith tic disorder. Arch Gen Psychiatry 1995;52:836.

280. Garfinkel BD, Webster CD, Sloman L. Responsesto methylphenidate and varied doses of caffeine inchildren with attention deficit disorder. Can JPsychiatry 1981;26:395–401.

281. Gittelman-Klein R, Landa B, Mattes JA, Klein DF.Methylphenidate and growth in hyperactivechildren. Arch Gen Psychiatry 1988;45:1127–30.

282. Handen BL, Feldman H, Gosling A, Breaux AM,McAuliffe S. Adverse side effects ofmethylphenidate among mentally retardedchildren with ADHD. J Am Acad Child AdolescPsychiatry 1991;30:241–5.

283. Hinshaw SP, Henker B, Whalen CK. Cognitive-behavioural and pharmacologic interventions for

References

144

hyperactive boys: comparative and combinedeffects. J Consult Clin Psychol 1984;52:739–49.

284. Hinshaw SP, Buhrmester D, Heller T. Angercontrol in response to verbal provocation: effectsof stimulant medication for boys with ADHD.J Abnorm Child Psychol 1989;17:393–407.

285. Hinshaw SP, Henker B, Whalen CK, Erhardt D,Dunnington REJ. Aggressive, prosocial, andnonsocial behaviour in hyperactive boys: doseeffects of methylphenidate in naturalistic settings.J Consult Clin Psychol 1989;57:636–43.

286. Klein RG, Abikoff H, Klass E, Ganeles D, Seese LM,Pollack S. Clinical efficacy of methylphenidate inconduct disorder with and without attention deficithyperactivity disorder. Arch Gen Psychiatry1997;54:1073–80.

287. Long N, Rickert VI, Ashcraft EW. Bibliotherapy asan adjunct to stimulant medication in thetreatment of attention-deficit hyperactivitydisorder. J Pediatr Health Care 1993;7:82–8.

288. Lufi D, Parish-Plass J, Gai E. The effect ofmethylphenidate on the cognitive and personalityfunctioning of ADHD children. Isr J PsychiatryRelat Sci 1997;34:200–9.

289. Matochik JA, Liebenauer LL, King C, SzymanskiHV, Cohen RM, Zametkin AJ. Cerebral glucosemetabolism in adults with attention deficithyperactivity disorder after chronic stimulanttreatment. Am J Psychiatry 1994;151:658–64.

290. McBride MC. An individual double-blindcrossover trial for assessing methylphenidateresponse in children with attention deficitdisorder. J Pediatr 1988;113:137–45.

291. Pelham WE, Hoza B, Kipp HL, Gnagy EM, Trane ST. Effects of methylphenidate andexpectancy of ADHD children’s performance, self-evaluations, persistence, and attributions on acognitive task. Exp Clin Psychopharmacol1997;5:3–13.

292. Quinn PO, Rapoport JL. One-year follow-up ofhyperactive boys treated with imipramine ormethylphenidate. Am J Psychiatry 1975;132:241–5.

293. Rapport MD, Carlson GA, Kelly KL, Pataki C.Methylphenidate and desipramine in hospitalizedchildren: I. Separate and combined effects oncognitive function. J Am Acad Child AdolescPsychiatry 1993;32:333–42.

294. Solanto MV, Wender EH, Bartell SS. Effects ofmethylphenidate and behavioral contingencies onsustained attention in attention-deficithyperactivity disorder: a test of the rewarddysfunction hypothesis. J Child AdolescPsychoparmacol 1997;7:123–36.

295. Spencer T, Wilens T, Biedermanm J, Faraone SV,Ablon JS, Lapey K. A double-blind, crossover

comparison of methylphenidate and placebo inadults with childhood-onset attention-deficithyperactivity disorder. Arch Gen Psychiatry1995;52:434–43.

296. Wender PH, Reimherr FW, Wood D, Ward M. A controlled study of methylphenidate in thetreatment of attention deficit disorder, residualtype, in adults. Am J Psychiatry 1985;142:547–52.

297. Winsberg BG, Press M, Bialer I, Kupietz S.Dextroamphetamine and methylphenidate in thetreatment of hyperactive-aggressive children.Pediatrics 1974;53:236–41.

298. Zametkin A, Rapoport JL, Murphy DL, Linnoila M, Ismond D. Treatment of hyperactivechildren with monoamine oxidase inhibitors: I,clinical efficacy. Arch Gen Psychiatry 1985;42:962–6.

299. Arnold LE, Pinkham SM, Votolato N. Does zincmoderate essential fatty acid and amphetaminetreatment of attention-deficit/hyperactivity disorder?J Child Adolesc Psychopharmacol 2000;10:111–17.

300. Efron D, Jarman FC, Barker MJ. Methylphenidatevs dexamphetamine in children with attentiondeficit hyperactivity disorder: a double-blind cross-over trial. J Paediatr Child Health 1997;33:A21.

301. Efron D, Jarman FC, Barker MJ. Child and parentperceptions of stimulant medication treatment inattention deficit hyperactivity disorder. J PaediatrChild Health 1998;34:288–92.

302. Castellanos FX, Elia J, Kruesi MJ, Marsh WL,Gulotta CS, Potter WZ, et al. Cerebrospinal fluidhomovanillic acid predicts behavioral response tostimulants in 45 boys with attentiondeficit/hyperactivity disorder.Neuropsychopharmacology 1996;14:125–37.

303. Handen BL, Feldman HM, Lurier A, Murray PJH.Efficacy of methylphenidate among preschoolchildren with developmental disabilities andADHD. In Hertzig ME, Farber E, editors. Annualprogress in child psychiatry and child development:2000–2001. New York: Brunner Routledge; 2003.pp. 247–62.

304. Klorman R, Coons HW, Borgstedt AD. Effects ofmethylphenidate on adolescents with a childhoodhistory of attention deficit disorder: I. Clinicalfindings (Erratum). J Am Acad Child AdolescPsychiatry 1987;26:820.

305. Dittmann R, Fouche DS, Kratochvil C,Heiligenstein J, Spencer T, Biederman J.Atomoxetine and methylphenidate treatment inchildren with ADHD: a prospective, randomized,open-label trial. Eur Neuropsychopharmacol 2001;11(Suppl 3):S318.

306. Findling RL, Short EJ, Manos MJ. Short-termcardiovascular effects of methylphenidate andAdderall. J Am Acad Child Adolesc Psychiatry2001;40:525–9.


145


307. Findling RL, Short EJ, Manos MJ. Developmentalaspects of psychostimulant treatment in childrenand adolescents with attention-deficit/hyperactivitydisorder. J Am Acad Child Adolesc Psychiatry2001;40:1441–7.

308. Faraone SV, Pliszka SR, Olvera RL, Biederman J.Adderall and methylphenidate in ADHD. In 2001Annual Meeting of the American Psychiatric Association,5–10 May 2001, New Orleans, LA.

309. Faraone SV, Short EJ, Biederman J, Findling RL,Roe C, Manos MJ. Efficacy of Adderall andmethylphenidate in attention deficit hyperactivitydisorder: a drug–placebo and drug–drug responsecurve analysis of a naturalistic study. Int JNeuropsychopharmacol 2002;5:121–9.

310. Newcorn J, Spencer TJ, Biedermann J, Milton D,Michelson D. Atomoxetine treatment in children andadolescents with attention-deficit/hyperactivity disorderand comorbid oppositional defiant disorder.Indianapolis, IL: Eli Lilly and Company; 2004.

311. Matza LS, Rentz AM, Secnik K, Swensen AR,Revicki DA, Michelson D, et al. The link betweenhealth-related quality of life and clinical symptomsamong children with attention-deficit hyperactivitydisorder. J Dev Behav Pediatr 2004;25:166–74.

312. Dunn D, Michelson D, Allen AJ, Kelsey D,Wernicke JF. Once-daily atomoxetine treatment forchildren and adolescents with attention-deficit/hyperactivity disorder. Ann Neurol 2002;52(Suppl 3):S116.

313. Bukstein OG. Once-daily atomoxetine may reduceattention deficit hyperactivity disorder symptomsin children and adolescents. Evid Based MentHealth 2003;6:42.

314. Michelson D. Once-daily administration ofatomoxetine: a new treatment for ADHD. In 155thAnnual Meeting of the American Psychiatric Association,18–23 May 2002, Philadelphia, PA.

315. Michelson D, Buitelaar J, Danckaerts M, Gillberg C, Spencer TJ, Zuddas A, et al.Atomoxetine in the long-term prevention ofrelapse in ADHD. Eur Neuropsychopharmacol2003;13 (Suppl 4):S455–6.

316. Pelham WE, Hoffman MT, Lock T. Evaluation ofonce-a-day OROS methylphenidate HCI (MPH)extended-release tablets versus MPH tid inchildren with ADHD in natural school settings.Pediatr Res 2000;47:31A.

317. Pelham WE, Hoffman MT, Lock T. Evaluation ofonce-a-day OROS® methylphenidate HCl (MPH)

extended-release tablets vs MPH tid in childrenwith ADHD in a laboratory setting. Pediatr Res2000;47:180.

318. Connor D. Once a day Concerta methylphenidatewas equivalent to 3 times daily methylphenidate inchildren with ADHD. Evid Based Ment Health2002;5:20.

319. Pliszka SR, Wynne SK, Olvera RL, Browne RG.Comparing Adderall methylphenidate in ADHD. In 152nd Annual Meeting of the AmericanPsychiatric Association, 15–20 May 1999,Washington, DC.

320. Faraone SV, Pliszka SR, Olvera RL, Skolnik R,Biederman J. Efficacy of Adderall andmethylphenidate in attention deficit hyperactivitydisorder: a reanalysis using drug–placebo anddrug–drug response curve methodology. J ChildAdolesc Psychopharmacol 2001;11:171–80.

321. Law SF, Schachar RJ. Do typical clinical doses ofmethylphenidate cause tics in children treated for attention-deficit hyperactivity disorder? J Am Acad Child Adolesc Psychiatry 1999;38:944–51.

322. Evans SW, Pelham WE, Smith BH, Bukstein O,Gnagy EM, Greiner AR, et al. Dose-responseeffects of methylphenidate on ecologically validmeasures of academic performance and classroombehavior in adolescents with ADHD. Exp ClinPsychopharmacol 2001;9:163–75.

323. Smith BH, Pelham WE, Jr., Gnagy E, Molina B,Evans S. The reliability, validity, and uniquecontributions of self-report by adolescentsreceiving treatment for attention-deficit/hyperactivity disorder. J Consult Clin Psychol2000;68:489–99.

324. Stein M, Seymour K, Black D, Sarampote C, Robb A, Conlon C, et al. Effects and side effects ofConcerta methylphenidate (MPH) in children withADHD and comorbid internalizing symptoms.Pediatr Res 2003;53:555A.

325. Wolraich ML. Efficacy and safety of OROS®

methylphenidate HCl (MPH) extended-releasetablets (CONCERTA™), conventional MPH, andplacebo in children with ADHD. Int JNeuropsychopharmacol 2000;3 (Suppl 1):S329.

326. Wolraich ML, Greenhill LL, Pelham W, Swanson J,Wilens T, Palumbo D, et al. Randomized,controlled trial of OROS methylphenidate once aday in children with attention-deficit/hyperactivitydisorder. Pediatrics 2001;108:883–92.

References

146

This version of HTA monograph volume 10, number 23 does not include the 220 pages of appendices. This is to save download time from the HTA website. The printed version of this monograph also excludes the appendices. View/download the appendices (800 kbytes).


379

Health Technology AssessmentProgramme

Prioritisation Strategy GroupMembers

Chair,Professor Tom Walley, Director, NHS HTA Programme,Department of Pharmacology &Therapeutics,University of Liverpool

Professor Bruce Campbell,Consultant Vascular & GeneralSurgeon, Royal Devon & ExeterHospital

Dr Edmund Jessop, MedicalAdvisor, National Specialist,Commissioning Advisory Group(NSCAG), Department ofHealth, London

Professor Jon Nicholl, Director,Medical Care Research Unit,University of Sheffield, Schoolof Health and Related Research

Dr John Reynolds, ClinicalDirector, Acute GeneralMedicine SDU, RadcliffeHospital, Oxford

Dr Ron Zimmern, Director,Public Health Genetics Unit,Strangeways ResearchLaboratories, Cambridge

Director, Professor Tom Walley, Director, NHS HTA Programme,Department of Pharmacology &Therapeutics,University of Liverpool

Deputy Director, Professor Jon Nicholl,Director, Medical Care ResearchUnit, University of Sheffield,School of Health and RelatedResearch

HTA Commissioning BoardMembers

Programme Director, Professor Tom Walley, Director, NHS HTA Programme,Department of Pharmacology &Therapeutics,University of Liverpool

Chair,Professor Jon Nicholl,Director, Medical Care ResearchUnit, University of Sheffield,School of Health and RelatedResearch

Deputy Chair, Professor Jenny Hewison,Professor of Health CarePsychology, Academic Unit ofPsychiatry and BehaviouralSciences, University of LeedsSchool of Medicine

Dr Jeffrey AronsonReader in ClinicalPharmacology, Department ofClinical Pharmacology,Radcliffe Infirmary, Oxford

Professor Deborah Ashby,Professor of Medical Statistics,Department of Environmentaland Preventative Medicine,Queen Mary University ofLondon

Professor Ann Bowling,Professor of Health ServicesResearch, Primary Care andPopulation Studies,University College London

Dr Andrew Briggs, PublicHealth Career Scientist, HealthEconomics Research Centre,University of Oxford

Professor John Cairns, Professorof Health Economics, PublicHealth Policy, London School ofHygiene and Tropical Medicine,London

Professor Nicky Cullum,Director of Centre for EvidenceBased Nursing, Department ofHealth Sciences, University ofYork

Mr Jonathan Deeks, Senior Medical Statistician,Centre for Statistics inMedicine, University of Oxford

Dr Andrew Farmer, SeniorLecturer in General Practice,Department of Primary Health Care, University of Oxford

Professor Fiona J Gilbert,Professor of Radiology,Department of Radiology,University of Aberdeen

Professor Adrian Grant,Director, Health ServicesResearch Unit, University ofAberdeen

Professor F D Richard Hobbs,Professor of Primary Care &General Practice, Department ofPrimary Care & GeneralPractice, University ofBirmingham

Professor Peter Jones, Head ofDepartment, UniversityDepartment of Psychiatry,University of Cambridge

Professor Sallie Lamb, Professor of Rehabilitation,Centre for Primary Health Care, University of Warwick

Professor Stuart Logan,Director of Health & SocialCare Research, The Peninsula Medical School, Universities of Exeter &Plymouth

Dr Linda Patterson, Consultant Physician,Department of Medicine,Burnley General Hospital

Professor Ian Roberts, Professorof Epidemiology & PublicHealth, Intervention ResearchUnit, London School ofHygiene and Tropical Medicine

Professor Mark Sculpher,Professor of Health Economics,Centre for Health Economics,Institute for Research in theSocial Services, University of York

Dr Jonathan Shapiro, SeniorFellow, Health ServicesManagement Centre,Birmingham

Ms Kate Thomas,Deputy Director,Medical Care Research Unit,University of Sheffield

Ms Sue Ziebland,Research Director, DIPEx,Department of Primary HealthCare, University of Oxford,Institute of Health Sciences

Current and past membership details of all HTA ‘committees’ are available from the HTA website (www.hta.ac.uk)


Health Technology Assessment Programme

380

Diagnostic Technologies & Screening PanelMembers

Chair,Dr Ron Zimmern, Director ofthe Public Health Genetics Unit,Strangeways ResearchLaboratories, Cambridge

Ms Norma Armston,Lay Member, Bolton

Professor Max BachmannProfessor of Health Care Interfaces, Department of Health Policy and Practice,University of East Anglia

Professor Rudy BilousProfessor of Clinical Medicine &Consultant Physician,The Academic Centre,South Tees Hospitals NHS Trust

Dr Paul Cockcroft, Consultant MedicalMicrobiologist and ClinicalDirector of Pathology,Department of ClinicalMicrobiology, St Mary'sHospital, Portsmouth

Professor Adrian K Dixon,Professor of Radiology,University Department ofRadiology, University ofCambridge Clinical School

Dr David Elliman, Consultant Paediatrician/Hon. Senior Lecturer,Population Health Unit, Great Ormond St. Hospital,London

Professor Glyn Elwyn,Primary Medical Care Research Group,Swansea Clinical School,University of Wales Swansea

Mr Tam Fry, HonoraryChairman, Child GrowthFoundation, London

Dr Jennifer J Kurinczuk,Consultant ClinicalEpidemiologist,National PerinatalEpidemiology Unit, Oxford

Dr Susanne M Ludgate, MedicalDirector, Medicines &Healthcare Products RegulatoryAgency, London

Professor William Rosenberg,Professor of Hepatology, LiverResearch Group, University ofSouthampton

Dr Susan Schonfield, Consultantin Public Health, SpecialisedServices Commissioning NorthWest London, HillingdonPrimary Care Trust

Dr Phil Shackley, SeniorLecturer in Health Economics,School of Population andHealth Sciences, University ofNewcastle upon Tyne

Dr Margaret Somerville, PMSPublic Health Lead, PeninsulaMedical School, University ofPlymouth

Dr Graham Taylor, ScientificDirector & Senior Lecturer,Regional DNA Laboratory, TheLeeds Teaching Hospitals

Professor Lindsay WilsonTurnbull, Scientific Director,Centre for MR Investigations &YCR Professor of Radiology,University of Hull

Professor Martin J Whittle,Associate Dean for Education,Head of Department ofObstetrics and Gynaecology,University of Birmingham

Dr Dennis Wright, Consultant Biochemist &Clinical Director, Pathology & The KennedyGalton Centre, Northwick Park & St Mark’sHospitals, Harrow

Pharmaceuticals PanelMembers

Chair,Dr John Reynolds, ChairDivision A, The John RadcliffeHospital, Oxford RadcliffeHospitals NHS Trust

Professor Tony Avery, Head of Division of PrimaryCare, School of CommunityHealth Services, Division ofGeneral Practice, University ofNottingham

Ms Anne Baileff, ConsultantNurse in First Contact Care,Southampton City Primary CareTrust, University ofSouthampton

Professor Stirling Bryan,Professor of Health Economics,Health Services Management Centre,University of Birmingham

Mr Peter Cardy, ChiefExecutive, Macmillan CancerRelief, London

Professor Imti Choonara,Professor in Child Health,Academic Division of ChildHealth, University ofNottingham

Dr Robin Ferner, ConsultantPhysician and Director, WestMidlands Centre for AdverseDrug Reactions, City HospitalNHS Trust, Birmingham

Dr Karen A Fitzgerald,Consultant in PharmaceuticalPublic Health, National PublicHealth Service for Wales,Cardiff

Mrs Sharon Hart, Head of DTB Publications, Drug &Therapeutics Bulletin, London

Dr Christine Hine, Consultant inPublic Health Medicine, SouthGloucestershire Primary CareTrust

Professor Stan Kaye,Cancer Research UK Professor of Medical Oncology,Section of Medicine, The Royal Marsden Hospital,Sutton

Ms Barbara Meredith,Lay Member, Epsom

Dr Andrew Prentice, SeniorLecturer and ConsultantObstetrician & Gynaecologist,Department of Obstetrics &Gynaecology, University ofCambridge

Dr Frances Rotblat, CPMPDelegate, Medicines &Healthcare Products RegulatoryAgency, London

Professor Jan Scott, Professor of Psychological Treatments,Institute of Psychiatry,University of London

Mrs Katrina Simister, AssistantDirector New Medicines,National Prescribing Centre,Liverpool

Dr Richard Tiner, MedicalDirector, Medical Department,Association of the BritishPharmaceutical Industry,London

Dr Helen Williams,Consultant Microbiologist,Norfolk & Norwich UniversityHospital NHS Trust



381

Therapeutic Procedures PanelMembers

Chair, Professor Bruce Campbell,Consultant Vascular andGeneral Surgeon, Departmentof Surgery, Royal Devon &Exeter Hospital

Dr Aileen Clarke,Reader in Health ServicesResearch, Public Health &Policy Research Unit, Barts &the London School of Medicine& Dentistry, London

Dr Matthew Cooke, Reader inA&E/Department of HealthAdvisor in A&E, WarwickEmergency Care andRehabilitation, University ofWarwick

Dr Carl E Counsell, ClinicalSenior Lecturer in Neurology,Department of Medicine andTherapeutics, University ofAberdeen

Ms Amelia Curwen, ExecutiveDirector of Policy, Services andResearch, Asthma UK, London

Professor Gene Feder, Professorof Primary Care R&D,Department of General Practiceand Primary Care, Barts & theLondon, Queen Mary’s Schoolof Medicine and Dentistry,London

Professor Paul Gregg,Professor of OrthopaedicSurgical Science, Department ofGeneral Practice and PrimaryCare, South Tees Hospital NHSTrust, Middlesbrough

Ms Bec Hanley, Co-Director,TwoCan Associates,Hurstpierpoint

Ms Maryann L Hardy, Lecturer, Division ofRadiography, University ofBradford

Professor Alan Horwich,Director of Clinical R&D,Academic Department ofRadiology, The Institute ofCancer Research, London

Dr Simon de Lusignan,Senior Lecturer, Primary Care Informatics,Department of CommunityHealth Sciences,St George’s Hospital MedicalSchool, London

Professor Neil McIntosh,Edward Clark Professor of Child Life & Health,Department of Child Life &Health, University of Edinburgh

Professor James Neilson,Professor of Obstetrics andGynaecology, Department ofObstetrics and Gynaecology,University of Liverpool

Dr John C Pounsford,Consultant Physician,Directorate of Medical Services,North Bristol NHS Trust

Karen Roberts, NurseConsultant, Queen ElizabethHospital, Gateshead

Dr Vimal Sharma, ConsultantPsychiatrist/Hon. Senior Lecturer,Mental Health Resource Centre,Cheshire and Wirral PartnershipNHS Trust, Wallasey

Dr L David Smith, ConsultantCardiologist, Royal Devon &Exeter Hospital

Professor Norman Waugh,Professor of Public Health,Department of Public Health,University of Aberdeen


Health Technology Assessment Programme

382Current and past membership details of all HTA ‘committees’ are available from the HTA website (www.hta.ac.uk)

Expert Advisory NetworkMembers

Professor Douglas Altman,Director of CSM & CancerResearch UK Med Stat Gp,Centre for Statistics inMedicine, University of Oxford,Institute of Health Sciences,Headington, Oxford

Professor John Bond,Director, Centre for HealthServices Research, University ofNewcastle upon Tyne, School ofPopulation & Health Sciences,Newcastle upon Tyne

Mr Shaun Brogan, Chief Executive, RidgewayPrimary Care Group, Aylesbury

Mrs Stella Burnside OBE,Chief Executive, Office of theChief Executive. TrustHeadquarters, AltnagelvinHospitals Health & SocialServices Trust, Altnagelvin AreaHospital, Londonderry

Ms Tracy Bury, Project Manager, WorldConfederation for PhysicalTherapy, London

Professor Iain T Cameron,Professor of Obstetrics andGynaecology and Head of theSchool of Medicine,University of Southampton

Dr Christine Clark,Medical Writer & ConsultantPharmacist, Rossendale

Professor Collette Clifford,Professor of Nursing & Head ofResearch, School of HealthSciences, University ofBirmingham, Edgbaston,Birmingham

Professor Barry Cookson,Director, Laboratory ofHealthcare Associated Infection,Health Protection Agency,London

Professor Howard Cuckle,Professor of ReproductiveEpidemiology, Department ofPaediatrics, Obstetrics &Gynaecology, University ofLeeds

Dr Katherine Darton, Information Unit, MIND – The Mental Health Charity,London

Professor Carol Dezateux, Professor of PaediatricEpidemiology, London

Mr John Dunning,Consultant CardiothoracicSurgeon, CardiothoracicSurgical Unit, PapworthHospital NHS Trust, Cambridge

Mr Jonothan Earnshaw,Consultant Vascular Surgeon,Gloucestershire Royal Hospital,Gloucester

Professor Martin Eccles, Professor of ClinicalEffectiveness, Centre for HealthServices Research, University ofNewcastle upon Tyne

Professor Pam Enderby,Professor of CommunityRehabilitation, Institute ofGeneral Practice and PrimaryCare, University of Sheffield

Mr Leonard R Fenwick, Chief Executive, Newcastleupon Tyne Hospitals NHS Trust

Professor David Field, Professor of Neonatal Medicine,Child Health, The LeicesterRoyal Infirmary NHS Trust

Mrs Gillian Fletcher, Antenatal Teacher & Tutor andPresident, National ChildbirthTrust, Henfield

Professor Jayne Franklyn,Professor of Medicine,Department of Medicine,University of Birmingham,Queen Elizabeth Hospital,Edgbaston, Birmingham

Ms Grace Gibbs, Deputy Chief Executive,Director for Nursing, Midwifery& Clinical Support Services, West Middlesex UniversityHospital, Isleworth

Dr Neville Goodman, Consultant Anaesthetist,Southmead Hospital, Bristol

Professor Alastair Gray,Professor of Health Economics,Department of Public Health,University of Oxford

Professor Robert E Hawkins, CRC Professor and Director ofMedical Oncology, Christie CRCResearch Centre, ChristieHospital NHS Trust, Manchester

Professor Allen Hutchinson, Director of Public Health &Deputy Dean of ScHARR,Department of Public Health,University of Sheffield

Dr Duncan Keeley,General Practitioner (Dr Burch& Ptnrs), The Health Centre,Thame

Dr Donna Lamping,Research Degrees ProgrammeDirector & Reader in Psychology,Health Services Research Unit,London School of Hygiene andTropical Medicine, London

Mr George Levvy,Chief Executive, MotorNeurone Disease Association,Northampton

Professor James Lindesay,Professor of Psychiatry for theElderly, University of Leicester,Leicester General Hospital

Professor Julian Little,Professor of Human GenomeEpidemiology, Department ofEpidemiology & CommunityMedicine, University of Ottawa

Professor Rajan Madhok, Medical Director & Director ofPublic Health, Directorate ofClinical Strategy & PublicHealth, North & East Yorkshire& Northern Lincolnshire HealthAuthority, York

Professor David Mant, Professor of General Practice,Department of Primary Care,University of Oxford

Professor Alexander Markham, Director, Molecular MedicineUnit, St James’s UniversityHospital, Leeds

Dr Chris McCall, General Practitioner, TheHadleigh Practice, Castle Mullen

Professor Alistair McGuire,Professor of Health Economics,London School of Economics

Dr Peter Moore, Freelance Science Writer, Ashtead

Dr Sue Moss, Associate Director,Cancer Screening EvaluationUnit, Institute of CancerResearch, Sutton

Mrs Julietta Patnick, Director, NHS Cancer ScreeningProgrammes, Sheffield

Professor Tim Peters,Professor of Primary CareHealth Services Research,Academic Unit of PrimaryHealth Care, University ofBristol

Professor Chris Price, Visiting Chair – Oxford, ClinicalResearch, Bayer DiagnosticsEurope, Cirencester

Professor Peter Sandercock,Professor of Medical Neurology,Department of ClinicalNeurosciences, University ofEdinburgh

Dr Eamonn Sheridan,Consultant in Clinical Genetics,Genetics Department,St James’s University Hospital,Leeds

Dr Ken Stein,Senior Clinical Lecturer inPublic Health, Director,Peninsula TechnologyAssessment Group, University of Exeter

Professor Sarah Stewart-Brown, Professor of Public Health,University of Warwick, Division of Health in theCommunity Warwick MedicalSchool, LWMS, Coventry

Professor Ala Szczepura, Professor of Health ServiceResearch, Centre for HealthServices Studies, University ofWarwick

Dr Ross Taylor, Senior Lecturer, Department ofGeneral Practice and PrimaryCare, University of Aberdeen

Mrs Joan Webster, Consumer member, HTA –Expert Advisory Network

How to obtain copies of this and other HTA Programme reports.An electronic version of this publication, in Adobe Acrobat format, is available for downloading free ofcharge for personal use from the HTA website (http://www.hta.ac.uk). A fully searchable CD-ROM is alsoavailable (see below).

Printed copies of HTA monographs cost £20 each (post and packing free in the UK) to both public andprivate sector purchasers from our Despatch Agents, York Publishing Services.

Non-UK purchasers will have to pay a small fee for post and packing. For European countries the cost is£2 per monograph and for the rest of the world £3 per monograph.

You can order HTA monographs from our Despatch Agents, York Publishing Services by:

– fax (with credit card or official purchase order) – post (with credit card or official purchase order or cheque)– phone during office hours (credit card only).

Additionally the HTA website allows you either to pay securely by credit card or to print out yourorder and then post or fax it.

Contact details are as follows:York Publishing Services Email: [email protected] Box 642 Tel: 0870 1616662YORK YO31 7WX Fax: 0870 1616663UK Fax from outside the UK: +44 1904 430868

NHS libraries can subscribe free of charge. Public libraries can subscribe at a very reduced cost of £100 for each volume (normally comprising 30–40 titles). The commercial subscription rate is £300 per volume. Please contact York Publishing Services at the address above. Subscriptions can only bepurchased for the current or forthcoming volume.

Payment methods

Paying by chequeIf you pay by cheque, the cheque must be in pounds sterling, made payable to York PublishingDistribution and drawn on a bank with a UK address.

Paying by credit cardThe following cards are accepted by phone, fax, post or via the website ordering pages: Delta, Eurocard,Mastercard, Solo, Switch and Visa. We advise against sending credit card details in a plain email.

Paying by official purchase orderYou can post or fax these, but they must be from public bodies (i.e. NHS or universities) within the UK.We cannot at present accept purchase orders from commercial companies or from outside the UK.

How do I get a copy of HTA on CD?

Please use the form on the HTA website (www.hta.ac.uk/htacd.htm). Or contact York PublishingServices (see contact details above) by email, post, fax or phone. HTA on CD is currently free of chargeworldwide.

The website also provides information about the HTA Programme and lists the membership of the variouscommittees.

HTA

Health Technology Assessm

ent 2006;Vol. 10: No. 23

The treatm

ent of attention deficit hyperactivity disorder in children and adolescents


S King, S Griffin, Z Hodges, H Weatherly, C Asseburg, G Richardson, S Golder, E Taylor, M Drummond and R Riemsma


HTAHealth Technology AssessmentNHS R&D HTA Programme

The National Coordinating Centre for Health Technology Assessment,Mailpoint 728, Boldrewood,University of Southampton,Southampton, SO16 7PX, UK.Fax: +44 (0) 23 8059 5639 Email: [email protected]://www.hta.ac.uk ISSN 1366-5278

FeedbackThe HTA Programme and the authors would like to know

your views about this report.

The Correspondence Page on the HTA website(http://www.hta.ac.uk) is a convenient way to publish

your comments. If you prefer, you can send your comments to the address below, telling us whether you would like

us to transfer them to the website.

We look forward to hearing from you.

July 2006

NHS R&D HTA Programme - LSHTM Research...

Documents

Transcript of NHS R&D HTA Programme - LSHTM Research...