omputers in

C

Computers in Human Behavior 23 (2007) 850–859

www.elsevier.com/locate/comphumbeh

Human Behavior

Does therapist guidance improve uptake,adherence and outcome from a CD-ROM

based cognitive-behavioral interventionfor the treatment of bulimia nervosa?

Kathryn Murray a, Ulrike Schmidt b,*,Maria-Guadelupe Pombo-Carril a, Miriam Grover a,

Joana Alenya a, Janet Treasure c, Christopher Williams d

a Eating Disorders Unit, Maudsley Hospital, Denmark Hill, London SE5 8AZ, UKb Section of Eating Disorders, Institute of Psychiatry, De Crespigny Park, PO Box 59,

London SE5 8AF, UKc Department of Academic Psychiatry, 5th floor, Thomas Guy House, Guys Hospital, London SE1 9RT, UK

d Section of Psychological Medicine, Academic Centre, Gartnavel Royal Hospital,

1055 Great Western Road, Glasgow G12 0XH, UK

Available online 10 December 2004

Abstract

Background: We recently demonstrated the efficacy and feasibility of a novel CD-ROM

based cognitive-behavioral multi-media self-help intervention for the treatment of bulimia

nervosa. What is not known in CD-ROM treatments is how to best to deliver and support

such packages in clinical practice. In particular, it is of great importance to identify to what

extent such packages can be offered stand alone, and to what extent additional support from

a practitioner is required.

Objective: The aim of the present study was to examine whether the addition of therapist

support to the CD-ROM intervention would improve treatment uptake, adherence and

outcome.

0747-5632/$ - see front matter � 2004 Elsevier Ltd. All rights reserved.

doi:10.1016/j.chb.2004.11.014

* Corresponding author. Fax: +44 2078480182.

E-mail address: u.schmidt@iop.kcl.ac.uk (U. Schmidt).

K. Murray et al. / Computers in Human Behavior 23 (2007) 850–859 851

Method: Two cohorts of patients with full or partial bulimia nervosa referred to a catchment

area based eating disorder service were offered an eight session CD-ROM-based cognitive-

behavioral self-help treatment (‘‘Overcoming Bulimia’’). The first cohort received minimal

guidance only and the second cohort were offered three brief focused support sessions with

a therapist. The two cohorts were compared on treatment uptake, adherence and outcome.

Results: Patients in both groups improved significantly. There were no significant differences

between the two groups in terms of treatment uptake, adherence or outcome, except that the

therapist guidance group more often achieved remission from excessive exercise at follow-up.

Discussion: These findings provide further support for the acceptability and efficacy of the

CD-ROM intervention for bulimia nervosa. Brief focused therapist guidance did not confer

any significant additional benefits. This result has important implications for the widespread

adoption of such approaches.

� 2004 Elsevier Ltd. All rights reserved.

Keywords: Bulimia nervosa; Eating disorder; CD-ROM; Treatment; Self-help; Computerised cognitive-

behavior therapy

1. Introduction

Systematic reviews have demonstrated the efficacy of cognitive behavioral ther-

apy (CBT) for the treatment of bulimia nervosa (Hay & Bacaltchuk, 2003a,

2003b), and recently published clinical guidelines for the treatment of eating disor-

ders recommend that CBT should be offered to most people with this disorder (Na-

tional Collaborating Centre for Mental Health, 2004). However, CBT is expensive

and trained therapists are in limited supply. Consequently, alternative methods of

delivery of CBT need to be developed to make this intervention more accessible.Computerised CBT (CCBT) programs have previously been shown to be accept-

able and effective in the treatment of many psychological disorders such as depres-

sion, anxiety disorders, and obsessive-compulsive disorder (e.g. Kaltenthaler et al.,

2002; Marks, Kenwright, McDonough, Whittaker, & Mataix-Cols, 2004; Proud-

foot et al., 2003a, 2003b). We recently piloted the use of a multi-media cogni-

tive-behavioral CD-ROM intervention (Overcoming Bulimia; Williams, Aubin,

Cottrell, & Harkin, 1998) in adults with BN in an open study. Patients accessed

the CD-ROM in the clinic, but had only minimal guidance from a practitionerwho showed them how to operate the program. High levels of patient satisfaction

and significantly reduced bingeing and vomiting were found (Bara-Carril et al.,

2004; Murray et al., 2003).

One important question is whether clinician guidance, supporting patients� in

their use of the program, might improve the efficacy of the CD-ROM intervention.

Previous research into manual-based cognitive-behavioral self-help for bulimia nerv-

osa has shown that the efficacy of such interventions can be significantly increased if

delivered with therapist guidance (for review see Birchall & Palmer, 2002; Perkins &Schmidt, 2004, in press). This question is of great importance because at present ac-

cess to cognitive-behavior therapy for BN is often focused within specialised eating

852 K. Murray et al. / Computers in Human Behavior 23 (2007) 850–859

services. If wider access to CBT for BN is to occur, information on how much prac-

titioner support is required for effective use is essential.

During our first pilot study, many bulimic patients requested therapist guidance

with the computer program. A qualitative analysis of patients� initial attitudes and

expectations towards the CD-ROM program for BN showed that some peoplewho did not engage with the computer treatment saw it as an inferior replacement

of therapist-led treatment, rather than the first step in a treatment plan, which could

go on to therapist-led help if required (Murray et al., 2003). These people stated their

preference for a therapist because they saw human interaction as more flexible and

sensitive to their individual needs, and they placed importance on the support, empa-

thy and opportunity for expression of problems offered within a therapeutic

relationship.

We hypothesised that analogous to previous research into manual-based self-helpin BN the addition of therapist guidance to the CD-ROM treatment program might

improve treatment uptake, adherence and outcome. The aim of the present paper

was to test this hypothesis.

2. Method

2.1. Sample

Patients were recruited from two consecutive cohorts of patients newly referred

to the Eating Disorders Unit of the South London and Maudsley NHS Trust

which provides a service to a catchment area population of two million. Patients

with a diagnosis of BN or eating disorder not otherwise specialised (EDNOS) were

eligible for participation. Patients with EDNOS were included, as many of them

closely resemble bulimia nervosa, with severe, distressing and long lasting symp-

toms, even though they do not meet the precise diagnostic criteria (Fairburn &Harrison, 2003). We defined EDNOS as any clinically relevant eating disorder

(i.e. with significant impairment of physical health or psychosocial functioning)

where the patient met all the criteria for bulimia nervosa except that the binge eat-

ing and/or inappropriate compensatory mechanisms occurred at a frequency of less

than twice a week or for a duration of less than three months. All diagnoses were

assigned by experienced and trained clinicians using a semi-structured clinical inter-

view, devised in our Unit, for DSM-IV diagnoses. The exclusion criteria were:

insufficient knowledge of English, insufficient literacy level, severe learning disabil-ity, anorexia nervosa, severe depression or acute suicidality, or alcohol/substance

dependence.

Informed written consent was sought from patients at initial assessment. The

study was approved by the research ethics committee at the Institute of Psychiatry

and South London and Maudsley NHS Trust. Patients who were eligible and con-

sented to participation were offered the CD-ROM treatment as a first step in treat-

ment, with the option of a therapist intervention 6–8 weeks after completion of the

program. The first cohort, described in more detail in a previous paper (Bara-Carril

K. Murray et al. / Computers in Human Behavior 23 (2007) 850–859 853

et al., 2004) completed the package with minimal guidance and the second cohort

received therapist guidance (TG) as detailed below.

2.2. Procedure

After their initial assessment, patients were introduced to the research team, who

were responsible for administering the program and booking appointments. Treat-

ment consisted of eight interactive computerised modules making up the Overcom-

ing Bulimia disc. The program uses a cognitive-behavioral therapy format,

incorporating educational and motivational strategies, with personalised feedback,

homework tasks, accompanying manuals and an audio relaxation tape (Williams

et al., 1998). Patients were required to complete sessions in sequential order and were

encouraged to attend on a weekly basis, though they were able to book appoint-ments at their convenience within working hours. The CD-ROM treatment was con-

ducted in a small self-contained clinic room and patients were left to work through

each session alone at their own pace.

The minimal guidance (MG) group had low-key interaction with a researcher who

introduced them to the program at their first appointment, and logged them onto the

computer and booked subsequent appointments with them. The therapist guidance

(TG) group received three 20-min sessions of therapist guidance after sessions 1, 3

and 8 as has been recommended by the developers of the CD-Rom. The sessionswere delivered by several professionals with different degrees of experience, including

nurse therapists and assistant psychologists. The aims of these sessions were to help

people make the best use of the CD-ROM sessions by looking at how the materials

applied to them, to increase and maintain motivation to complete the program; to

clarify any information on the program that the patient had not understood and

to foster compliance with homework. Thus, the guidance sessions focussed predom-

inantly on issues raised in the CD-ROM modules rather than on any additional top-

ics. If clients raised other topics, they were either told that this would be covered lateron in the program, or where that was not the case that they might wish to discuss this

issue with their therapist when they began face-to-face therapy at a later date. The

therapists giving the guidance received regular clinical supervision of their cases by

one of the authors (US).

2.3. Measures

Demographic information and baseline measurements of eating disorder symp-tomatology were collected from patients prior to starting the computer program.

The main outcome measure used was the short evaluation of eating disorder

symptoms (SEED), a patient self-rating instrument developed in the context of

a large naturalistic European multi-centre study of eating disorders (Kordy

et al., 1999). This scale gives separate ratings for frequency of bingeing, self-induced

vomiting, laxative/diuretic use, excessive exercise and food restriction. The scale

points are as follows: ‘‘1 = not at all’’, 2 = up to 1 · week’’, 3 = 2/3 · week’’,

‘‘4 = daily’’, ‘‘5 = more than 1 · day’’. Although full validation of the SEED

854 K. Murray et al. / Computers in Human Behavior 23 (2007) 850–859

has not yet been published, preliminary analyses on data from approximately

180 patients with a range of eating disorder diagnoses for whom ratings made

independently by both the patient and experienced and trained clinicians are

available and indicate excellent inter-rater reliability. The j values for those

symptoms included in this report are all above 0.70. The SEED was completedby patients at assessment, session 3 and at follow-up (6–8 weeks after completing

the CD-ROM).

2.4. Statistical analysis

The data were analysed using SPSS-11 for Windows, using parametric and

non-parametric tests where appropriate. Symptom remission (defined as being be-

low the DSM-IV frequency threshold of twice-weekly occurrence of a particularbehaviour) was calculated for bingeing, vomiting, laxative abuse, dieting and

excessive exercise pre-treatment, from the SEED scores at week 3 and at fol-

low-up. Change scores from pre-treatment to week 3 and from pre-treatment to

follow-up (6–8 weeks after completing the CD-ROM) were calculated according

to improvement (= �1), no change (= 0) or deterioration (= 1) in terms of whether

a patient stayed the same or moved above/below the DSM threshold frequency

for each symptom. Independent samples Mann–Whitney tests were then carried

out to look for differences between the MG and TG groups at each of these timepoints. The week 3 change score was used because we wanted to know whether

there might be differences in the speed of improvement in the two groups and

the follow-up score was used to assess whether there were any more lasting differ-

ences. SEED data on 73% of participants were available at week 3 and on 85.5%

of participants at follow-up.

3. Results

3.1. Baseline characteristics

A total of 103 patients were eligible for the study and agreed to participate. Eighty

two of these patients (79 female, 3 male) took up the computer programme (43/53

(81%) of patients in the MG group and 39/50 (78%) of the TG group). Patients

who took up the computer treatment had a diagnosis of BN binge–purge type in

82% of cases; 12% suffered from BN non-purging type and 6% had eating disordernot otherwise specialised. The age range of participants was between 18 and 62 years,

and the mean duration of the eating disorder was 11 years (SD 9.7). Thirty-two

(39%) of the 82 patients were receiving antidepressant medication at assessment. This

information was missing for 12 patients. There were no significant differences be-

tween the two groups on demographic characteristics. Table 1 gives the baseline de-

tails of patients who took up the CD-ROM treatment. The only significant baseline

difference in eating symptomatology was that the TG group had a significantly higher

rate of laxative use (p < 0.046).

Table 1

Pre-treatment symptomatology

Minimal guidance

group (n = 43)

Therapist guidance

group (n = 39)

p

Mean BMI (SD) 21.4 (8.7) 22.2 (4.3) 0.59a

Eating disorder symptoms

(percent below DSM-IV frequency threshold)

% %

Bingeing 20.9 13.9 0.557b

Vomiting 35 21.6 0.217b

Laxative use 80.5 58.3 0.046b

Excessive exercise 72.5 52.8 0.097b

Food restriction 17.1 27.8 0.284b

a Independent samples t-tests.b Fisher�s exact test.

K. Murray et al. / Computers in Human Behavior 23 (2007) 850–859 855

3.2. Session attendance

There were no significant differences between the groups on session attendance

(Table 2). Twenty patients (47%) of the MG group and seventeen patients (44%)

of the TG group completed all eight sessions.

3.3. Symptom change

Table 3 gives the percentages of each group below the DSM-IV frequency

thresholds and the symptom changes at weeks 3 and and at follow-up. There

was a trend for patients in the TG group to improve more rapidly on bingeing,

vomiting and exercise from baseline to week 3 with p-values of 0.118, 0.163 and

0.154, respectively. At follow-up, the only significant difference between the

groups was that the TG group showed significantly more improvement in exercise

(p = 0.006).

Table 2

Number of CD-ROM sessions attended

Number of sessions attended GROUP minimal guidance group Therapist guidance group Total

1 3 (7%) 2 (5%) 5 (6%)

2 6 (14%) 6 (15%) 12 (15%)

3 5 (12%) 3 (8%) 8 (9%)

4 7 (16%) 6 (15%) 13 (16%)

5 1 (2%) 3 (8%) 4 (5%)

6 1 (2%) 2 (5%) 3 (4%)

8 20 (47%) 17 (44%) 37 (45%)

Total 43 39 82

Table 3

Percentages below DSM-IV threshold and symptom changes

Week 3 Minimal guidance group Therapist guidance group p

% Below DSM threshold Mean ranks of symptom

changes baseline to week 3

% Below DSM threshold Mean ranks of symptom

changes baseline to week 3

Bingeing 23.1 25.2 42.1 20 0.118a

Vomiting 50.0 24.8 66.7 20 0.163a

Dieting 32.0 24.7 47.8 24.3 0.816a

Exercise 76.0 26.7 87.0 21.2 0.154a

Laxative 91.7 22 60.0 20.9 0.359a

Follow-up (6–8 weeks

after the end

of the CD-ROM treatment)

Mean rank symptom

changes baseline to

follow-up

Mean rank symptom

changes baseline to

follow-up

Bingeing 31.6 37.5 41.2 33.2 0.335a

Vomiting 71.1 34.7 57.6 33.2 0.811a

Dieting 29.7 34.8 48.5 33 0.688a

Exercise 80.6 37.9 82.4 28.8 0.006a

Laxative 91.7 34.2 75.0 30.6 0.315a

a Mann–Whitney U test.

85

6K

.M

urra

yet

al.

/C

om

pu

tersin

Hu

ma

nB

eha

vior

23

(2

00

7)

85

0–

85

9

K. Murray et al. / Computers in Human Behavior 23 (2007) 850–859 857

4. Discussion and conclusions

Patients in both groups showed significant improvements in frequency of bingeing

and a range of reversing behaviours. The only significant difference in terms of eating

disorder-related outcomes between the groups was that a larger proportion of peoplewere in remission from excessive exercise in the therapist-guidance group at follow-

up. This is interesting, but on its own is probably not enough to justify the addition

of therapist guidance sessions.

Contrary to our hypothesis the addition of limited therapist guidance did not seem

to confer any additional benefit in terms of improving treatment uptake, retention in

treatment or outcome from the CD-ROM program. One possibility is that the failure

to find any differences between the two groups is simply due to the relatively small

sample size. Alternatively, it must be considered that the therapist guidance wasnot intensive enough to make a significant difference. However, part of the attractive-

ness of computerised delivery of treatment is that this can reduce therapist time and

we therefore wanted to keep the time investment quite low. The rather rigid timing of

the guidance sessions may have been less helpful than a more flexible access to guid-

ance sessions if and when needed. In a recent open evaluation of computer-aided CBT

for anxiety and depression, patients were able to access six brief scheduled telephone

or face-to-face contacts for advice as needed, provided it was within office hours

(Marks et al., 2003). Over 12 weeks of computer-aided treatment the average time ta-ken for support was just 1 h, but patients reported high levels of satisfaction with this

form of therapist support. In our study, the minimal guidance group had some (albeit

very limited) contact with the research workers in the clinic and the need to come to

the clinic to schedule appointments to access the computer together with a friendly

face may have been enough �to keep most people going�. Overall the findings of the

study provide further support for the acceptability and preliminary efficacy of the

CD-ROM program.

5. Limitations

The study has a number of limitations. First, this is not a randomised controlled

trial and it is possible that the groups differed systematically on variables other than

eating disorder symptoms and that this may have obscured any treatment differ-

ences. Moreover, we did not tape-record the guidance sessions and check the quality

of the guidance and clinician�s adherence to their task. Thirdly, outcomes were mea-sured by self-report rather than through an interview-based measure, such as the eat-

ing disorders examination. Finally, the follow-up period was relatively brief.

6. Conclusions and implications

The present study provides further preliminary support for the acceptability

and efficacy of a CD-ROM based intervention for bulimia nervosa. If offered in a

858 K. Murray et al. / Computers in Human Behavior 23 (2007) 850–859

structured clinical environment minimal low-key support seems to be sufficient and

the addition of formal therapist guidance sessions does not seem to be necessary.

However, it is not known what the acceptability or efficacy of the same program

would be if delivered entirely without support, e.g. in the patient�s home or other

public settings such as libraries or via the internet. Further research is needed to elu-cidate these questions.

Conflict of interest

Dr. C. Williams holds IPR in the Calipso Overcoming Bulimia package.

Acknowledgements

K.M. was supported by the Psychiatry Research Trust and the South Thames

Specialist Audit Programme. M.G.P.-C. was supported by the Psychiatry Research

Trust. J.A. was supported by the South Thames Specialist Audit Programme. We

thank Dr. Sabine Landau for statistical advice.

References

Bara-Carril, N., Williams, C., Pombo-Carril, M. G., Reid, Y., Murray, K., Treasure, J., et al. (2004).

A preliminary investigation into the feasibility and efficacy of a CD-ROM based cognitive-

behavioral self-help intervention for bulimia nervosa. International Journal of Eating Disorders, 35,

538–548.

Birchall, H., & Palmer, B. (2002). Doing it by the book: what place for guided self-help for bulimic

disorders?. European Eating Disorders Review, 10, 379–385.

Fairburn, C. G., & Harrison, P. J. (2003). Eating disorders. The Lancet, 361, 407–416.

Hay, P. J., & Bacaltchuk, J. (2003a). Psychotherapy for bulimia nervosa and binging. Cochrane Database

Systematic Reviews CD000562.

Hay, P. J., & Bacaltchuk, J. (2003b). Bulimia nervosa. Clinical Evidence, 9, 997–1009.

Kaltenthaler, E., Shackley, P., Stevens, K., Beverley, C., Parry, G., & Chilcott, J. (2002). A systematic

review and economic evaluation of computerised cognitive behaviour therapy for depression and

anxiety. Health Technology Assessments, 6, 1–89.

Kordy, H., Richard, M., Herrmann, A., Murphy, F., Treasure, J., & Charpentier, P. (1999). A computer

assisted eating disorder specific quality management system: EQUAL-TREAT. European Eating

Disorders Review, 7, 239–258.

Marks, I., Mataix-Cols, D., Kenwright, M., Cameron, R., Hirsh, S., & Gega, L. (2003). Pragmatic

evaluation of computer-aided self-help for anxiety and depression. British Journal of Psychiatry, 183,

57–65.

Marks, I. M., Kenwright, M., McDonough, M., Whittaker, M., & Mataix-Cols, D. (2004). Saving

clinicians� time by delegating routine aspects of therapy to a computer: a randomized controlled trial in

phobia/panic disorder. Psychological Medicine, 34, 9–17.

Murray, K., Pombo-Carril, M. G., Bara-Carril, N., Grover, M., Reid, Y., Langham, C., et al. (2003).

Factors determining uptake of a CD-ROM based CBT self-help treatment for bulimia: patient

characteristics and subjective appraisals of self-help treatment. European Eating Disorders Review, 11,

243–260.

K. Murray et al. / Computers in Human Behavior 23 (2007) 850–859 859

National Collaborating Centre for Mental Health. (2004). Eating Disorders. Core Interventions in the

treatment and management of anorexia nervosa, bulimia nervosa and related eating disorders. British

Psychological Society and Royal College of Psychiatrists.

Perkins, S., & Schmidt, U. (2004). Self-Help for eating disorders. Academy of Eating Disorders. Annual

Review (in press).

Proudfoot, J., Swain, S., Widmer, S., Watkins, E., Goldberg, D., Marks, I., et al. (2003a). The

development and beta-test of a computer-therapy program for anxiety and depression: hurdles and

lessons. Computers in Human Behaviour, 19, 277–289.

Proudfoot, J., Goldberg, D., Mann, A., Everitt, B., Marks, I., & Gray, J. A. (2003b). Computerized,

interactive, multimedia cognitive-behavioural program for anxiety and depression in general practice.

Psychological Medicine, 33, 217–227.

Williams, C. J., Aubin, S. D., Cottrell, D., & Harkin, P. J. R. (1998). Overcoming bulimia: a self-help

package. University of Leeds. Available from www.calipso.co.uk. Media Innovations 0044113

2849220. Media Innovations, 3 Gemini Business Park, Sheepscar Way, Leeds LS7 3JB, England, UK.