RESULTS: Evidence appraisal forms are available for allstudy designs related to each domain of a clinical question–ranging from randomized, controlled trials to case reports andqualitative studies in domains of treatment, diagnosis, andprognosis, to name a few. Based on the domain of the clinicalquestion, then the study design to which the study applies, theuser is guided to determine the aspects of the quality of thestudy as good quality, lesser quality, or lacking validity, reli-ability, or applicability. This quality level for each study rep-resents the study design for the domain and the quality of thestudy, and is then easily translated into the body of evidence.On each appraisal form, the user is then able to see the contextof each study by the domain of the question, the study designand evidence level, and the quality resulting from the appraisal.DISCUSSION (CONCLUSION): Although the questionson the forms are common, the approach to the group of formsis the seminal concept, accounting for multiple study designsand multiple types of clinical questions.TARGET AUDIENCE(S):

1. Evidence synthesizer, developer of systematic reviews ormeta-analyses

2. Guideline developer3. Developer of guideline-based products4. Quality improvement manager/facilitator5. Medical educator6. Medical providers and executives7. Nurses

P6– Pervasive developmental disorders: A shared

knowledge synthesis

Joelle M. Andre-Vert, MSc (Presenter) (HauteAutorite de Sante, La Plaine Saint Denis, France,Metropolitan); Muriel Dhenain, MD (Haute Autoritede Sante, La Plaine Saint Denis, France,Metropolitan)

PRIMARY TRACK: Evidence generation and synthesisSECONDARY TRACK: Evidence sharingBACKGROUND (INTRODUCTION): Pervasive develop-mental disorders (PDD) gather a large diversity of clinicalsituations that require education, health, and social manage-ment. As a first step before guidelines development, NationalPDD policy commissioned the French National Authority forHealth (HAS) to draft an evidence synthesis that would beshared among professionals and users’ representatives. Thepurpose was to identify shared key messages on definition,epidemiology, functioning specificities, diagnostic, evaluationtools, and intervention description.LEARNING OBJECTIVES (TRAINING GOALS):

1. Update current knowledge on pervasive developmentaldisorders.

2. Present a method to identify shared knowledge amongprofessional and users’ representatives.

METHODS: Formal consensus method was adapted. A crit-ical literature review was performed by searching medical andeducational databases (MEDLINE, PsycINFO, CINAHL, Co-

chrane Library, Base SantePsy, SAPHIR, ERIC; 2000-August2009; English or French). A steering committee drafted theevidence report based on clinical practice guidelines, system-atic review, and, if further information needed, on clinical trialsand extracted key message proposals. The evidence report wassubmitted to multidisciplinary peer reviewers and users’ asso-ciations to complete the report and proposals. An independentrating panel graded each proposal, using a 9-point numericalscale. Evidence synthesis was drafted by the steering commit-tee based on key messages, which were consensually judgedconforming to scientific evidence (strong agreement if allmarks were 7-9 after two rating tours and one interspersedmeeting).RESULTS: 209/249 proposals received a strong agreementamong professionals and users’ representatives. No consensuswas found on: French classification, exact prevalence of men-tal retardation associated with PDD, symptoms’ evolution dur-ing adolescence and adulthood, possible evolution toward psy-chotic disorders.DISCUSSION (CONCLUSION): Whereas discussionsamong professionals and users’ representatives frequentlypoint out divergences, the formal consensus method helpedthem to find consensual key messages. These messages will bespread by initial and continuing education about PDD. Thiswork will allow developing future guidelines on a sharedknowledge.TARGET AUDIENCE(S):

1. Evidence synthesizer, developer of systematic reviews ormeta-analyses

2. Guideline developer3. Medical educator4. Health care policy analyst/policy-maker5. Medical providers and executives6. Allied health professionals7. Consumers and patients representatives

P7– Considerations for development of

Generalized Anxiety Disorder Guidelines (GAD) in

the US

Jonathan Davidson, MD (Duke University, SeabrookIsland, South Carolina); Mark Pollack, MD (Harvard,Boston, Massachusetts); Cheryl Brewster, MHA(Presenter) (EPI-Q, Inc., Oak Brook, Illinois)

PRIMARY TRACK: Evidence generation and synthesisSECONDARY TRACK: Other evidence generation and syn-thesisBACKGROUND (INTRODUCTION): No clinical guide-lines for Generalized Anxiety Disorder (GAD) have been de-veloped in the United States. Early guideline development inpsychiatric conditions outside the US has been hindered bylack of methodological quality and evidence, inadequate trans-lation of evidence into recommendations, and no systematicupdating. The GAD Guidelines Assessment (GAD-GA)Working Group was convened to identify recent credible clin-ical practice guidelines (CPGs), conduct guideline content syn-

81Poster

thesis and gap analysis, and provide information about criticalissues involved with guideline development, use, and impact.LEARNING OBJECTIVES (TRAINING GOALS):

1. Define the GAD-GA process to examine evidence inexisting GAD guidelines.

2. Understand the process for recognizing guideline gaps.3. Identify areas in need of future evidence generation to

inform guideline development.METHODS: The GAD-GA process included: a MEDLINEsearch for existing GAD guidelines, limited to the Englishlanguage, published in the previous 10 years; a systematicguideline assessment for quality, evidence strength and gaps;recommendations to inform future US guideline development.RESULTS: Thirteen GAD guidelines developed in the lastdecade were identified primarily from Canada, Europe, Asia,South Africa and the Pacific Rim. Eleven GAD guidelines,available in English, were evaluated and numerous evidencegaps were noted: 1) Limited information on duration of GADtreatment beyond initial 6-12 months. 2) No consideration ofimportant GAD subpopulations (e.g., elderly, medically ill,children, and comorbid with other mental illnesses). 3) Lack ofevidence for treatment failures beyond initial therapy. 4) Inad-equate acknowledgment for role and duration of non-pharma-cologic treatments (i.e., psychotherapy).DISCUSSION (CONCLUSION): Significant gaps exist inexisting international GAD guidelines; efforts in the US shouldaddress GAD, considering notable gaps exist.TARGET AUDIENCE(S):

1. Clinical researcher2. Evidence synthesizer, developer of systematic reviews or

meta-analyses3. Guideline developer4. Guideline implementer5. Developer of guideline-based products6. Quality improvement manager/facilitator7. Medical educator8. Health care policy analyst/policy-maker9. Health insurance payers and purchasers10. Medical providers and executives11. Allied health professionals12. Consumers and patients representatives13. Nurses

P8– How can evidence be presented in a graphical

form?

Jonathan Nyong, MPH (Presenter) (NICE,Manchester, England, United Kingdom);Tarang Sharma, MPH (NICE, Manchester, England,United Kingdom); Elizabeth J. Shaw, MS(Manchester, England, United Kingdom)

PRIMARY TRACK: Evidence generation and synthesisSECONDARY TRACK: Other evidence generation and syn-thesisBACKGROUND (INTRODUCTION): Guideline develop-ers are often confronted with the challenge of summarizing a

wide range of disparate studies, particularly in areas whereRCTs are lacking. Often, a narrative summary is used but thiscan be lengthy and may not easily allow understanding andinterpretation of the totality of evidence.LEARNING OBJECTIVES (TRAINING GOALS):

1. To encourage the use of graphical presentation in guide-line development.

2. To describe the use of graphical summaries of data.METHODS: During the development of national guidelines,we wished to present evidence in a graphical format to com-plement the GRADE and evidence tables.RESULTS: We considered different graphical representationsto summarize and present 69 studies included in an evidencereview of ablative therapies in Barrett’s esophagus. Initially,we considered the use of a linear representation (that is, asimple diagram showing which interventions were evaluatedand, where appropriate, any comparisons). Although this pro-vided a useful model for RCTs, it proved less applicable to thetype of evidence considered in the ablative therapies guideline,primarily case series. We therefore used a Venn diagram ap-proach, which allowed the representation of studies with orwithout a comparison, and the representation of studies with acombination of interventions.DISCUSSION (CONCLUSION): Graphical representationallowed us to present evidence in a simple and easily under-standable way. Although we did not undertake a formal as-sessment, anecdotally, guideline group members found it ex-tremely helpful in navigating the “evidence landscape.” It wasalso useful for the technical team as an audit tool throughout.

Guideline developers could be more innovative in usinggraphical representation, especially when there is a wide rangeof evidence to be considered and statistical methods of sum-marizing the data are not possible.TARGET AUDIENCE(S):

1. Evidence synthesizer, developer of systematic reviews ormeta-analyses.

2. Guideline developer.3. Developer of guideline-based products.

P9– Developing the Epilepsies Guideline using

network meta-analysis

Vanessa D. Nunes, MSc (Presenter) (NationalClinical Guideline Centre, London, England, UnitedKingdom)

PRIMARY TRACK: Evidence generation and synthesisSECONDARY TRACK: Synthesizing evidence (e.g., meta-analysis, decision modeling)BACKGROUND (INTRODUCTION): Recent advances inthe development of anti-epileptic drugs (AEDs) have war-ranted an update of the pharmacological management sectionof the Epilepsies Guideline published by the National Institutefor Health and Clinical Development (NICE) in 2004. NICEhas commissioned the National Clinical Guideline Centre forAcute and Chronic Conditions (NCGC-ACC) to undertake thiswork. In addition to the results of conventional meta-analysis

82 Otolaryngology–Head and Neck Surgery, Vol 143, No 1S1, July 2010