Skip to main content
NCBI home page
BMJ Clinical Evidence logoLink to BMJ Clinical Evidence
. 2015 Mar 4;2015:1009.

Bulimia nervosa: online interventions

Phillipa J Hay 1,#, Angélica Medeiros Claudino 2,#
PMCID: PMC4356174  PMID: 25735611

Abstract

Introduction

Up to 1% of people in the community may have bulimia nervosa, characterised by an intense preoccupation with body weight, binge-eating episodes, and use of extreme measures to counteract the feared effects of overeating. People with bulimia nervosa are of normal weight or are overweight, making the condition distinct from anorexia nervosa. After 10 years, about half of people with bulimia nervosa will have recovered fully, one third will have made a partial recovery, and 10% to 20% will still have symptoms.

Methods and outcomes

We conducted a systematic review and aimed to answer the following clinical question: What are the effects of online interventions for people with bulimia nervosa? We searched: Medline, Embase, The Cochrane Library, and other important databases up to April 2014 (BMJ Clinical Evidence reviews are updated periodically; please check our website for the most up-to-date version of this review). We included harms alerts from relevant organisations such as the US Food and Drug Administration (FDA) and the UK Medicines and Healthcare products Regulatory Agency (MHRA).

Results

We found eight studies that met our inclusion criteria. We performed a GRADE evaluation of the quality of evidence for interventions.

Conclusions

In this systematic review we present information relating to the effectiveness and safety of the following interventions: applications (apps) or online programmes used as an adjunct to face-to-face therapy, delivery of self-help online, and delivery of therapy online.

Key Points

Up to 1% of people at any one time may have bulimia nervosa, characterised by an intense preoccupation with body weight, binge-eating episodes, and use of extreme measures to counteract the feared effects of overeating.

  • People with bulimia nervosa are of normal weight or are overweight, making the condition distinct from anorexia nervosa.

  • Obesity has been associated with both an increased risk of bulimia nervosa and a worse prognosis, as have personality disorders and substance misuse.

  • After 10 years, about half of people with bulimia nervosa will have recovered fully, one third will have made a partial recovery, and 10% to 20% will still have symptoms.

In this review, we have considered interventions delivered online.

With online therapies, it is often hard to make rigorous diagnoses as therapy is not being delivered in a clinic and, in the real world, such online therapies rely on a degree of self-referral and self-report of symptoms, especially where used for initial therapy versus relapse prevention.

For the purposes of this review, we have included interventions delivered via the internet (including real-time videoconferencing), by e-mail, or by text, but not those delivered by other routes (such as by speaking by telephone) or delivered by other means without an internet component.

The trials we found were generally small, included mixed populations, used different delivery systems, assessed different interventions, and reported different outcome measures. This makes it difficult to draw reliable conclusions or to generalise results.

Although we have reported an ITT analysis where possible, some trials had large numbers of drop-outs, which may affect the robustness of results.

We don’t know whether online therapy involving contact with a person online is more effective than placebo, sham therapy, waiting list control, or no online therapy in people with bulimia nervosa or eating disorder not otherwise specified.

  • We found insufficient evidence on an internet cognitive behavioural therapy (CBT)-based programme plus e-mail guidance compared with guided bibliotherapy plus e-mail guidance.

  • We also found insufficient evidence on the effects of CBT delivered by telemedicine compared with face-to-face contact.

One study found no evidence that a self-help therapeutic writing task delivered online was more effective than a control writing task delivered online. However, the trial was small, and participants did not need a formal diagnosis of bulimia nervosa.

We found limited evidence that a 16-week text messaging intervention (SMS) that delivered a tailored feedback message may improve abstinence and reduce full bulimic symptoms at 8 months, compared with treatment as usual, in women who had been discharged after specialist inpatient care.

  • However, the trial included a mixed population of women with full and sub-threshold bulimia nervosa as well as women with eating disorder not otherwise specified.

  • In addition, the trial was limited to one hospital site and used a non-standard intervention, which may limit its generalisability.

There is a need for further high-quality studies in this area.

Clinical context

General background

Bulimia nervosa is a common condition but many sufferers do not access evidence based psychological therapies. Online interventions have subsequently been developed to improve access to therapies such as cognitive behaviour therapy. These range from therapies that simulate face-to face sessions with a therapist online using video to self-directed self-help programmes that have no therapist contact.

Focus of the review

In this review we included all interventions that used an online medium grouped by intensity and therapist contact.

Comments on evidence

Variance in type and intensity of the use of online approaches makes assessing evidence difficult. The informality inherent in internet use and interventions without therapist contact also limits the accuracy of clinical diagnostic assessment of participants and comparability with outcomes in trials with valid diagnostic assessment.

Search and appraisal summary

The literature search was carried out in April 2014. For more information on the electronic databases searched and criteria applied during assessment of studies for potential relevance to the review, please see the Methods section. After deduplication and removal of conference abstracts, 146 records were screened for inclusion in the review. Appraisal of titles and abstracts led to the exclusion of 123 studies and the further review of 23 full publications. Of the 23 full articles evaluated, one systematic review and four additional RCTs were included. One systematic review published since the search date found no new RCTs and will be assessed fully at the next update of this BMJ Clinical Evidence overview.

About this condition

Definition

Bulimia nervosa is an intense pre-occupation with body weight and shape, with regular episodes of overeating (binge eating) associated with extreme measures to counteract the feared effects of the overeating. If a person also meets the diagnostic criteria for anorexia nervosa, then the diagnosis of anorexia nervosa takes precedence. The latest version of the Diagnostic and Statistical Manual for Mental Disorders (DSM-5) has broadened the previous criteria in the DSM-IV used for diagnosing bulimia nervosa by including people with a lower frequency of bingeing and purging symptoms (now at least once a week for 3 months). This change allowed for many cases that were previously considered in DSM-IV as having an eating disorder not otherwise specified (EDNOS) and represented, in fact, less severe cases of bulimia nervosa (those presenting with lower than twice a week episodes), to now be diagnosed as having bulimia nervosa. For this reason, this review includes studies with mixed samples of participants with bulimia nervosa or EDNOS of bulimic type that were diagnosed before changes to DSM-5. Bulimia nervosa can be difficult to identify because of extreme secrecy about binge eating and purgative behaviour. While current weight may be normal, there is often a history of anorexia nervosa or of restrictive dieting and weight suppression. Some people alternate between anorexia nervosa and bulimia nervosa. Nearly all cases of bulimia nervosa identified in a national community survey featured an additional psychiatric disorder, and common comorbidities were mood, anxiety, impulse control, and substance-misuse disorders. Some RCTs included people with sub-threshold bulimia nervosa, or with a related eating disorder, binge-eating disorder. Where possible, only results relevant to bulimia nervosa are reported in this review. In this review, we have considered interventions delivered online. For the purposes of this review, we have included interventions delivered via the internet (including real-time videoconferencing), by e-mail, or by text. We have not included interventions delivered by other means, such as by speaking by telephone or by CD/DVD delivered without an internet component.

Incidence/ Prevalence

In community-based studies, the point prevalence of bulimia nervosa is between 0.5% and 1.0% in people, with a lifetime prevalence of up to 2% in women and with an even social-class distribution. About 90% of people diagnosed with bulimia nervosa are women. The numbers presenting with bulimia nervosa in industrialised countries increased during the decades after its recognition in the late 1970s, although the incidence has plateaued or even fallen since then, with an incidence of new diagnoses at 6.6 per 100,000 in 2000 in young women from the general population, and around 20.7 per 100,000 new cases in females aged 10 to 49 in UK primary care registers. A 'cohort effect', with an increasing incidence, has been reported in community surveys. The prevalence of eating disorder features such as weight/shape overconcern, as found in bulimia nervosa, appears to be increasing in non-industrialised populations but may vary across ethnic groups. African-American women have a lower rate of restrictive dieting compared with white American women, but they have a similar rate of recurrent binge eating.

Aetiology/ Risk factors

The aetiology of bulimia nervosa is complex, but sociocultural pressures to be thin and the promotion of dieting seem to increase risk. One community-based case-control study compared 102 people with bulimia nervosa with 204 healthy controls and found higher rates of obesity, mood disorder, sexual and physical abuse, parental obesity, substance misuse, low self-esteem, perfectionism, disturbed family dynamics, parental weight/shape concern, and early menarche in people with the eating disorder. Heritability is high, ranging from 28% to 83% in one review, although it has been suggested that genotypic variations map onto intermediate phenotypes, such as traits of affective instability and impulsivity, rather than onto a 'gross' bulimia nervosa phenotype. Personality traits such as perfectionism may thus be important moderators of clinical features of an eating disorder.

Prognosis

A large study (222 people) from a trial of antidepressants and structured, intensive group psychotherapy found that, after a mean follow-up of 11.5 years, 11% still met criteria for bulimia nervosa, whereas 70% were in full or partial remission. One study (102 women) of the natural course of bulimia nervosa found that 31% continued to have the disorder at 15 months and 15% continued to have the disorder at 5 years. Only 28% received treatment during the follow-up period. A 5-year naturalistic study of 23 people with bulimia nervosa found a 74% remission at 5 years, with a 47% probability of relapse within the 5-year follow-up study in those in remission. A large review of 79 studies found an overall recovery rate of 45%, 27% partial improvement, and 23% with a chronic course, and crude mortality rate of 0.32%. There are very few consistent predictors of long-term outcome. A systematic review found a family history of obesity to predict poor outcome for treatment of bulimia nervosa, and a good prognosis to be associated with shorter illness duration and good interpersonal relationships. However, on the whole there were many more inconsistent positive and negative predictors than consistent positive predictors of outcome. A consistent post-treatment predictor of a better outcome is an early response to treatment.

Aims of intervention

To reduce symptoms of bulimia nervosa; to improve general psychiatric symptoms; to improve social functioning and quality of life; to minimise the adverse effects of treatment.

Outcomes

Symptom improvement frequency of binge eating or bingeing, abstinence from binge eating or bingeing, frequency of behaviours to reduce weight and counter the effects of binge eating, severity of extreme weight and shape preoccupation, severity of general psychiatric symptoms, severity of depression, improvement in social and adaptive functioning, remission rates, relapse rates, withdrawal rates; quality of life; and adverse effects.

Methods

BMJ Clinical Evidence search and appraisal April 2014. The following databases were used to identify studies for this systematic review: Medline 1966 to April 2014, Embase 1980 to April 2014, and The Cochrane Database of Systematic Reviews 2014, issue 4 (1966 to date of issue). Additional searches were carried out in the Database of Abstracts of Reviews of Effects (DARE) and the Health Technology Assessment (HTA) database. We also searched for retractions of studies included in the review. Titles and abstracts identified by the initial search, run by an information specialist, were first assessed against predefined criteria by an evidence scanner. Full texts for potentially relevant studies were then assessed against predefined criteria by an evidence analyst. Studies selected for inclusion were discussed with an expert contributor. All data relevant to the review were then extracted by an evidence analyst. Study design criteria for inclusion in this review were published RCTs and systematic reviews of RCTs in the English language, at least single-blinded, and containing 20 or more individuals (10 in each arm), of whom more than 80% were followed up. There was no minimum length of follow-up. We excluded all studies described as 'open', 'open label', or not blinded unless blinding was impossible. We included RCTs and systematic reviews of RCTs where harms of an included intervention were studied, applying the same study design criteria for inclusion as we did for benefits. In addition, we use a regular surveillance protocol to capture harms alerts from organisations such as the FDA and the MHRA, which are added to the reviews as required. To aid readability of the numerical data in our reviews, we round many percentages to the nearest whole number. Readers should be aware of this when relating percentages to summary statistics such as relative risks (RRs) and odds ratios (ORs). We have performed a GRADE evaluation of the quality of evidence for interventions included in this review (see table). The categorisation of the quality of the evidence (high, moderate, low, or very low) reflects the quality of evidence available for our chosen outcomes in our defined populations of interest. These categorisations are not necessarily a reflection of the overall methodological quality of any individual study, because the Clinical Evidence population and outcome of choice may represent only a small subset of the total outcomes reported, and population included, in any individual trial. For further details of how we perform the GRADE evaluation and the scoring system we use, please see our website (www.clinicalevidence.com).

Table.

GRADE Evaluation of interventions for Bulimia nervosa: online interventions.

Important outcomes Quality of life, Symptom improvement
Studies (Participants) Outcome Comparison Type of evidence Quality Consistency Directness Effect size GRADE Comment
What are the effects of online interventions for people with bulimia nervosa?
2 (146) Symptom improvement Delivery of therapy online versus placebo, sham therapy online, or waiting list control 4 –3 0 –1 0 Very low Quality points deducted for sparse data, weak methods, and incomplete reporting of results; directness point deducted for mixed population (BN or EDNOS)
3 (352) Symptom improvement Delivery of therapy online versus delivery of therapy not online 4 –2 0 –2 0 Very low Quality points deducted for weak methods and incomplete reporting of results; directness points deducted for mixed population (BN or EDNOS) and high drop-out rate
1 (128) Quality of life Delivery of therapy online versus delivery of therapy not online 4 –2 0 –2 0 Very low Quality points deducted for sparse data and incomplete reporting of results; directness points deducted for mixed population (BN or EDNOS) and high drop-out rate
1 (94) Symptom improvement Delivery of self-help online versus placebo, sham therapy online, or waiting list control 4 –2 0 –1 0 Very low Quality points deducted for sparse data and incomplete reporting of results; directness point deducted for no formal DSM criteria for diagnosis
1 (165) Symptom improvement Apps or online programmes used as an adjunct to face-to-face therapy versus no adjunct therapy 4 –2 0 –2 0 Very low Quality points deducted for sparse data and weak methods; directness points deducted for mixed population (BN or EDNOS) and single site/non-standard intervention affecting generalisability
Open in a new tab

We initially allocate 4 points to evidence from RCTs, and 2 points to evidence from observational studies. To attain the final GRADE score for a given comparison, points are deducted or added from this initial score based on preset criteria relating to the categories of quality, directness, consistency, and effect size. Quality: based on issues affecting methodological rigour (e.g., incomplete reporting of results, quasi-randomisation, sparse data [<200 people in the analysis]). Consistency: based on similarity of results across studies. Directness: based on generalisability of population or outcomes. Effect size: based on magnitude of effect as measured by statistics such as relative risk, odds ratio, or hazard ratio.

Glossary

Beck Depression Inventory

Standardised scale to assess depression. This instrument consists of 21 items to assess the intensity of depression. Each item is a list of 4 statements (rated 0, 1, 2, or 3), arranged in increasing severity, about a particular symptom of depression. The range of scores possible are 0 = least severe depression to 63 = most severe depression. It is recommended for people aged 13 to 80 years. Scores of more than 12 or 13 indicate the presence of depression.

Binge eating

Modified from DSM-IV. Eating, in a discrete period (e.g., hours), a large amount of food, accompanied by a lack of control over eating during the episode.

Cognitive behavioural therapy (CBT)

Brief structured treatment using relaxation and exposure procedures, and aimed at changing dysfunctional beliefs and negative automatic thoughts (typically 20 sessions over 12–16 weeks).

Remission

Sustained abstinence (longer than 1 month) from binge eating.

Short Form 36 (SF-36)

A scale that assesses health-related quality of life across eight domains: limitations in physical activities (physical component); limitations in social activities; limitations in usual role activities due to physical problems; pain; psychological distress and wellbeing (mental health component); limitations in usual role activities because of emotional problems; energy and fatigue; and general health perceptions.

Very low-quality evidence

Any estimate of effect is very uncertain.

Disclaimer

The information contained in this publication is intended for medical professionals. Categories presented in Clinical Evidence indicate a judgement about the strength of the evidence available to our contributors prior to publication and the relevant importance of benefit and harms. We rely on our contributors to confirm the accuracy of the information presented and to adhere to describe accepted practices. Readers should be aware that professionals in the field may have different opinions. Because of this and regular advances in medical research we strongly recommend that readers' independently verify specified treatments and drugs including manufacturers' guidance. Also, the categories do not indicate whether a particular treatment is generally appropriate or whether it is suitable for a particular individual. Ultimately it is the readers' responsibility to make their own professional judgements, so to appropriately advise and treat their patients. To the fullest extent permitted by law, BMJ Publishing Group Limited and its editors are not responsible for any losses, injury or damage caused to any person or property (including under contract, by negligence, products liability or otherwise) whether they be direct or indirect, special, incidental or consequential, resulting from the application of the information in this publication.

Contributor Information

Phillipa J Hay, School of Medicine, Campbelltown Campus, Western Sydney University, Sydney, Australia.

Angélica Medeiros Claudino, Federal University of Sao Paulo, Sao Paulo, Brazil.

References

  • 1.American Psychiatric Association. Diagnostic and statistical manual of mental disorders, 5th ed. (DSM-5). Washington, DC: American Psychiatric Publishing; 2013. [Google Scholar]
  • 2.Hudson JI, Hiripi E, Pope HG, Jr, et al. The prevalence and correlates of eating disorders in the National Comorbidity Survey Replication. Biol Psychiatry 2007;61:348–358. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Oakley Browne MA, Wells JE, Scott KM, et al; New Zealand Mental Health Survey Research Team. Lifetime prevalence and projected lifetime risk of DSM-IV disorders in Te Rau Hinengaro: the New Zealand Mental Health Survey. Aust N Z J Psychiatry 2006;40:865–874. [DOI] [PubMed] [Google Scholar]
  • 4.Garfinkel PE, Lin B, Goering P, et al. Bulimia nervosa in a Canadian community sample: prevalence, co-morbidity, early experiences and psychosocial functioning. Am J Psychiatry 1995;152:1052–1058. [DOI] [PubMed] [Google Scholar]
  • 5.Gard MCE, Freeman CP. The dismantling of a myth: a review of eating disorders and socioeconomic status. Int J Eat Disord 1996;20:1–12. [DOI] [PubMed] [Google Scholar]
  • 6.Preti A, Girolamo GD, Vilagut G, et al; ESEMeD-WMH Investigators. The epidemiology of eating disorders in six European countries: results of the ESEMeD-WMH project. J Psychiatr Res 2009;43:1125–1132. [DOI] [PubMed] [Google Scholar]
  • 7.Favaro A, Ferrara S, Santonastaso P. The spectrum of eating disorders in young women: a prevalence study in a general population sample. Psychosom Med 2003;65:701–708. [DOI] [PubMed] [Google Scholar]
  • 8.Keski-Rahkonen A, Hoek HW, Susser ES, et al. Epidemiology and course of anorexia nervosa in the community. Am J Psychiatry 2007;164:1259–1265. [DOI] [PubMed] [Google Scholar]
  • 9.Striegel-Moore RH, Dohm FA, Kraemer HC, et al. Eating disorders in white and black women. Am J Psychiatry 2003;160:1326–1331. [DOI] [PubMed] [Google Scholar]
  • 10.Wade TD, Bergin JL, Tiggemann M, et al. Prevalence and long-term course of lifetime eating disorders in an adult Australian twin cohort. Aust N Z J Psychiatry 2006;40:121–128. [DOI] [PubMed] [Google Scholar]
  • 11.Hoek HW. Incidence, prevalence mortality of anorexia nervosa and other eating disorders. Curr Opin Psychiatry 2006;19:389–394. [DOI] [PubMed] [Google Scholar]
  • 12.Micali N, Hagberg KW, Petersen I, et al. The incidence of eating disorders in the UK in 2000-2009: findings from the General Practice Research Database. BMJ Open. 2013;3:e002646. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Hall A, Hay PJ. Eating disorder patient referrals from a population region 1977–1986. Psychol Med 1991;21:697–701. [DOI] [PubMed] [Google Scholar]
  • 14.Kendler KS, Maclean C, Neale M, et al. The genetic epidemiology of bulimia nervosa. Am J Psychiatry 1991;148:1627–1637. [DOI] [PubMed] [Google Scholar]
  • 15.Mishra SK, Mukhopadhyay S. Eating and weight concerns among Sikkimese adolescent girls and their biocultural correlates: an exploratory study. Public Health Nutr 2011;14:853–859. [DOI] [PubMed] [Google Scholar]
  • 16.Striegel-Moore RH, Wifley DE, Caldwell MB, et al. Weight-related attitudes and behaviors of women who diet to lose weight: a comparison for black dieters and white dieters. Obes Res 1996;4:109–116. [DOI] [PubMed] [Google Scholar]
  • 17.Stice E. Risk and maintenance factors for eating pathology: a meta-analytic review. Psychol Bull 2002;128:825–848. [DOI] [PubMed] [Google Scholar]
  • 18.Fairburn CG, Welch SL, Doll HA, et al. Risk factors for bulimia nervosa. A community-based case-control study. Arch Gen Psychiatry 1997;54:509–517. [DOI] [PubMed] [Google Scholar]
  • 19.Steiger H, Bruce KR. Phenotypes, endophenotypes, and genotypes in bulimia spectrum eating disorders. Can J Psychiatry 2007;52:220–227. [DOI] [PubMed] [Google Scholar]
  • 20.Bulik CM. Exploring the gene-environment nexus in eating disorders. J Psychiatry Neurosci 2005;30:335–339. [PMC free article] [PubMed] [Google Scholar]
  • 21.Mitchison D, Hay PJ. The epidemiology of eating disorders: genetic, environmental, and societal factors. Clin Epidemiol 2014;6:89–97. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Steinhausen HC, Weber S. The outcome of bulimia nervosa: findings from one-quarter century of research. Am J Psychiatry 2009;166:1331–1341. [DOI] [PubMed] [Google Scholar]
  • 23.Fairburn C, Cooper Z, Doll H, et al. The natural course of bulimia nervosa and binge eating disorder in young women. Arch Gen Psychiatry 2000;57:659–665. [DOI] [PubMed] [Google Scholar]
  • 24.Grilo CM, Pagano ME, Skodol AE, et al. Natural course of bulimia nervosa and of eating disorder not otherwise specified: 5-year prospective study of remissions, relapses, and the effects of personality disorder psychopathology. J Clin Psychiatry 2007;68:738–746. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.National Institute for Care Excellence. Eating disorders: core interventions in the treatment and management of anorexia nervosa, bulimia nervosa and related disorders. London: National Institute for Care Excellence (NICE), 2004:35. Search date 2002. [Google Scholar]
  • 26.Fairburn CG, Agras WS, Walsh BT, et al. Prediction of outcome in bulimia nervosa by early change in treatment. Am J Psychiatry 2004;161:2322–2324. [DOI] [PubMed] [Google Scholar]
  • 27.Walsh BT, Sysko R, Parides MK. Early response to desipramine among women with bulimia nervosa. Int J Eat Disord 2006;39:72–75. [DOI] [PubMed] [Google Scholar]
  • 28.Aardoom JJ, Dingemans AE, Spinhoven P, et al. Treating eating disorders over the internet: a systematic review and future research directions. Int J Eat Disord 2013;46:539–552. [DOI] [PubMed] [Google Scholar]
  • 29.Dolemeyer R, Tietjen A, Kersting A, et al. Internet-based interventions for eating disorders in adults: a systematic review. BMC Psychiatry 2013;13:207. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Ruwaard J, Lange A, Broeksteeg J, et al. Online cognitive behavioural treatment of bulimic symptoms: a randomised controlled trial. Clin Psychol Psychother 2012;20:308–318. [DOI] [PubMed] [Google Scholar]
  • 31.Sánchez-Ortiz VC, Munro C, Stahl D, et al. A randomised controlled trial of internet based cognitive-behavioural therapy for bulimia nervosa or related disorders in a student population. Psychol Med 2011;41:407–417. [DOI] [PubMed] [Google Scholar]
  • 32.Wagner G, Penelo E, Wanner C, et al. Internet-delivered cognitive behavioural therapy v conventional guided self-help for bulimia nervosa: long-term evaluation of a randomised controlled trial. Br J Psychiatry 2013;202:135–141. [DOI] [PubMed] [Google Scholar]
  • 33.Mitchell JE, Crosby RD, Wonderlich SA, et al. A randomized trial comparing the efficacy of cognitive-behavioral therapy for bulimia nervosa delivered via telemedicine versus face-to-face. Behav Res Ther 2008;46:581–592. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Loucas CE, Fairburn CG, Whittington C, et al. E-therapy in the treatment and prevention of eating disorders: a systematic review and meta-analysis. Beh Res Ther 2014;63:122–131. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Johnston O, Startup H, Lavender A, et al. Therapeutic writing as an intervention for symptoms of bulimia nervosa: effects and mechanism of change. Int J Eat Disord 2010;43:405–419. [DOI] [PubMed] [Google Scholar]
  • 36.Bauer S, Okon E, Meermann R, et al. Technology-enhanced maintenance of treatment gains in eating disorders: efficacy of an intervention delivered via text messaging. J Consult Clin Psychol 2012;80:700–706. [DOI] [PubMed] [Google Scholar]
  • 37.American Psychiatric Association. Diagnostic and statistical manual of mental disorders, 4th ed. Washington DC: American Psychiatric Press, 1994. [Google Scholar]
BMJ Clin Evid. 2015 Mar 4;2015:1009.

Delivery of therapy online (involving contact with a person online)

Summary

We don’t know whether online therapy involving contact with a person is more effective than placebo, sham therapy, or waiting list control.

One RCT found insufficient evidence on an internet cognitive behavioural therapy (CBT) -based programme plus e-mail guidance compared with guided bibliotherapy plus e-mail guidance.

Another RCT found insufficient evidence on the effects of CBT delivered by telemedicine compared with face-to-face contact.

However, both trials were small and included a mixed population of people who also had eating disorder not otherwise specified.

Although this may limit the generalisability of any results to people with bulimia nervosa alone, as DSM-5 has broader criteria than DSM-IV, these populations may actually be more valid than they first appear.

Benefits and harms

Delivery of therapy online versus placebo, sham therapy online, or waiting list control:

We found two systematic reviews (search date 2013; 2012), which identified two RCTs. The first RCT (105 people, 99% female) compared online CBT (based on existing CBT manuals plus 25 scheduled therapist feedback sessions, taking 13 hours in total to complete, given over 20 weeks) and a waiting list group (which started online CBT after 20 weeks). The third arm evaluated bibliotherapy. There was no face-to-face contact. The RCT aimed to generalise findings to a population with varying levels of bulimic symptoms, and a formal diagnosis of bulimia nervosa (BN) was not an inclusion requirement. Participants had to report recurrent binge eating, inappropriate weight control behaviour, and elevated concern with body shape and weight. Of the included population, average age was 31 years, 81/105 (77%) were of normal weight, 65/105 (62%) had received treatment before (not further specified), and most had chronic symptoms (mean 10 years, 11 years, and 13 years in the 3 groups). DSM-IV criteria were not reported. The majority (80%) engaged in purging behaviour. The RCT reported an ITT analysis with last observation carried forward. The second RCT compared a guided self-help treatment programme with a waiting list control group. It included 76 female students who met DSM-IV criteria for BN (39 women; 51% of total) or an eating disorder not otherwise specified (EDNOS) (37 women; 49% of total). In the study, EDNOS was specified as recurrent episodes of binge eating and/or compensatory behaviours that occurred less than 24 times during the previous 3 months or an eating disorder where the abstinence period between episodes was more than 2 weeks. This included people with purging disorder. No minimum number of binge or purge episodes was specified for inclusion, but where these behaviours occurred less than once per week, eating disorder symptoms in general were reviewed to see whether the participant warranted inclusion in the trial. Only two women with EDNOS fell below a weekly frequency of behaviours (as DSM-5 uses weekly, not twice-weekly, frequency criteria, most women in this study would meet DSM-5 criteria for BN). At baseline, the mean age was 24 years, mean duration of symptoms was 6.6 years, mean Eating Disorder Examination-Global score was 3.6, half the sample had not been diagnosed with an eating disorder previously, and 72% had not had any form of psychological treatment previously. The online intervention comprised a cognitive behavioural interactive multimedia programme ('Overcoming bulimia online') consisting of eight sessions requiring 45 minutes each, based on CBT. In addition, therapists sent e-mails every 1 to 2 weeks and responded to any e-mail received. They supported and encouraged participants to use the package. The intervention lasted 8 to 12 weeks. The people in the control group did not receive the programme until 12 weeks. Assessments were conducted by telephone interview and questionnaires completed online. One review re-calculated data based on ITT assuming a zero effects for study drop-outs. We have reported these results here.

Symptom improvement

Delivery of therapy online compared with placebo, sham therapy online, or waiting list control We don’t know whether the delivery of therapy online is more effective than a waiting list control in improving symptoms in a mixed population of adults with bulimia nervosa or eating disorder not otherwise specified, as we found insufficient evidence (very low-quality evidence).

Ref (type) Population Outcome, Interventions Results and statistical analysis Effect size Favours
Symptoms

RCT
3-armed trial 		105 people with symptoms of bulimia nervosa (BN) Symptoms measured by EDE-Q at 20 weeks
with online CBT
with waiting list control
Absolute results not reported
P = 0.003 		Effect size not calculated online CBT

RCT
3-armed trial 		105 people with symptoms of BN Binge eating at 20 weeks
with online CBT
with waiting list control
Absolute results not reported
P = 0.04 		Effect size not calculated online CBT

RCT
3-armed trial 		105 people with symptoms of BN Purging at 20 weeks
with online CBT
with waiting list control
Absolute results not reported
P = 0.02 		Effect size not calculated online CBT

Systematic review		 76 women with BN or eating disorder not otherwise specified (EDNOS)
Data from 1 RCT		 Binge eating
with online CBT
with waiting list control
Effect size +0.35
95% CI –0.10 to +0.80 		Not significant

Systematic review		 76 women with BN or EDNOS Purging
with online CBT
with waiting list control
Effect size +0.27
95% CI –0.18 to +0.72 		Not significant

Systematic review		 76 women with BN or EDNOS Vomiting
with online CBT
with waiting list control
Effect size +0.29
95% CI –0.16 to +0.74 		Not significant
Open in a new tab

Quality of life

No data from the following reference on this outcome.

Adverse effects

No data from the following reference on this outcome.

Delivery of therapy online versus delivery of therapy not online:

We found two systematic reviews (search date 2013; 2012), which identified two RCTs. We found one additional RCT. The first RCT was included in one review. It included 155 women aged 16 to 35 years who fulfilled the diagnostic criteria for bulimia nervosa (BN) purging type (DSM-IV-TR) and eating disorder not otherwise specified (EDNOS), with binge eating or purging behaviour between once and twice a week or for less than 3 months and a body mass index (BMI) above 18. The RCT reported that the diagnosis of BN was present in 90% of participants, while 10% had a diagnosis of EDNOS (absolute numbers not reported). The RCT only reported baseline data for people who actually received the intervention. Of these 126 women, the average age was 24 years, the average duration of eating disorder was 8 to 9 years, and 83/126 (66%) had previously had formal psychotherapeutic or in-patient treatment. It compared a CBT-based self-help programme that included seven modules consisting of lessons, exercises, and examples provided by an internet platform, which was used for a period of 4 to 7 months, with a conventional guided bibliotherapy intervention, which was a self-help manual based on CBT. The content of the manual and structure was similar to the internet intervention, but addressed additional topics such as drug misuse and sexuality, and the use of the book was more flexible than the internet intervention. In addition, both groups had similar support through weekly e-mails from psychologists or psychotherapists. The aim of the e-mail support was to motivate people, answer technical questions about the programme, or address other problems that arose. Assessments at 4 and 7 months were conducted face to face, while at 18 months, most interviews were conducted by telephone. Of 155 women initially randomised, 126 (81%) women actually received the intervention, and 87 (56%) women completed the evaluation at 18 months. The RCT reported an ITT analysis with last observation carried forward for missing data. The level of blinding for outcome assessment was unclear. The second RCT (105 people, 99% female) compared online CBT therapy (based on existing CBT manuals plus 25 scheduled therapist feedback sessions, taking 13 hours in total to complete, given over 20 weeks), a bibliotherapy group (which received a hard copy of a self-help book for BN based on the same principles as applied in online treatment but with no therapist support), and a waiting list group who started online CBT after 20 weeks. There was no face-to-face contact. The RCT aimed to generalise findings to a population with varying levels of bulimic symptoms, and a formal diagnosis of BN was not an inclusion requirement. Participants had to report recurrent binge eating, inappropriate weight control behaviour, and elevated concern with body shape and weight. Of the included population, average age was 31 years, 81/105 (77%) were of normal weight, 65/105 (62%) had received treatment before (not further specified), and most had chronic symptoms (mean 10 years, 11 years, and 13 years in the 3 groups). DSM-IV criteria were not reported. The majority (80%) engaged in purging behaviour. The RCT reported an ITT analysis with last observation carried forward. In the online group, 26/35 (74%) of people completed the treatment versus 17/35 (68%) in the bibliotherapy group. At 20 weeks, attrition was 6/35 (17%) in the online group versus 12/35 (34%) in the bibliotherapy group, while at 1 year it was 13/35 (37%) in the online group versus 16/35 (46%) in the bibliotherapy group. The additional RCT (128 people, 98% female) compared CBT delivered by a telemedicine system that linked to a regional healthcare system facility and involved interaction with the therapist through telemedicine only, and the therapy used was based on manual-based CBT, versus CBT delivered in person by a therapist. The therapy consisted of 20 sessions delivered over a 16-week period, and telemedicine units were placed so as to mimic the interpersonal distance and height equality used in face-to-face therapy and used real-time videoconferencing via T1 lines. Participants were at least 18 years of age and met DSM-IV criteria for BN (purging or non-purging subtype) or EDNOS with one of the following features: DSM-IV criteria for BN except binge eating/purging at a minimum frequency of once per week; DSM-IV criteria for BN with only subjective binge eating episodes. At baseline, 71/128 (55%) people had a diagnosis of BN and 57/128 (44%) people had a diagnosis of EDNOS, average age was 29 years, and average BMI was 23. The randomisation sequence was generated by an independent statistician and stratified by eating disorder diagnosis and antidepressant use. Allocation was concealed, assessors were blinded, and analysis was by intention to treat with the last observation carried forward. In total, of 128 people (98% female) initially randomised, 80 (62%) people were assessed at the end of treatment, 72 (56%) people were followed up at 3 months, and 52 (40%) people were followed up at 12 months.

Symptom improvement

Delivery of therapy online compared with delivery of therapy not online We don’t know whether the delivery of therapy online is more or less effective than the delivery of therapy not online in improving symptoms in a mixed population of adults with bulimia nervosa or eating disorder not otherwise specified, as we found insufficient evidence (very low-quality evidence).

Ref (type) Population Outcome, Interventions Results and statistical analysis Effect size Favours
Symptoms

RCT		 155 women with bulimia nervosa (BN) or eating disorder not otherwise specified (EDNOS)
In review 		Objective binge eating (scale not further defined), ANOVA analysis 0–18 months
with internet CBT-based programme plus e-mail guidance
with guided bibliotherapy intervention plus e-mail guidance
Absolute results not reported
P = 0.838
See Further information on studies 		Not significant

RCT		 155 women with BN or EDNOS
In review 		Vomiting (scale not further defined), ANOVA analysis 0–18 months
with internet CBT-based programme plus e-mail guidance
with guided bibliotherapy intervention plus email guidance
Absolute results not reported
P = 0.353
See Further information on studies 		Not significant

RCT		 155 women with BN or EDNOS
In review 		Fasting (scale not further defined), ANOVA analysis 0–18 months
with internet CBT-based programme plus e-mail guidance
with guided bibliotherapy intervention plus e-mail guidance
Absolute results not reported
P = 0.324
See Further information on studies 		Not significant

RCT
3-armed trial 		105 people with symptoms of BN Symptoms measured by EDE-Q at 20 weeks
with online CBT
with bibliotherapy
Absolute results not reported
P = 0.04 		Effect size not calculated online CBT

RCT
3-armed trial 		105 people with symptoms of BN Binge eating at 20 weeks
with online CBT
with bibliotherapy
Absolute results not reported
P = 0.003 		Effect size not calculated online CBT

RCT
3-armed trial 		105 people with symptoms of BN Purging at 20 weeks
with online CBT
with bibliotherapy
Absolute results not reported
P = 0.02 		Effect size not calculated online CBT

RCT
3-armed trial 		105 people with symptoms of BN Symptoms measured by EDE-Q at 1 year
with online CBT
with bibliotherapy
Absolute results not reported
P = 0.99 		Not significant

RCT
3-armed trial 		105 people with symptoms of BN Binge eating at 1 year
with online CBT
with bibliotherapy
Absolute results not reported
P = 0.99 		Not significant

RCT
3-armed trial 		105 people with symptoms of BN Purging at 1 year
with online CBT
with bibliotherapy
Absolute results not reported
P = 0.99 		Not significant

RCT		 128 adults with BN or EDNOS Abstinence rates (no objective binge eating or purging [vomiting, laxative abuse, diuretic abuse] in the previous 28 days) based on Eating Disorder Examination (EDE) interview at end of treatment
17/62 (27%) with telemedicine CBT
19/66 (29%) with face-to-face CBT
Reported as not significant
P value not reported 		Not significant

RCT		 128 adults with BN or EDNOS Abstinence rates (no objective binge eating or purging [vomiting, laxative abuse, diuretic abuse] in the previous 28 days) based on EDE interview at 12 months
14/62 (23%) with telemedicine CBT
19/66 (29%) with face-to-face CBT
Reported as not significant
P value not reported 		Not significant

RCT		 128 adults with BN or EDNOS Purging frequency at 12 months
with telemedicine CBT
with face-to-face CBT
Absolute results not reported
P = 0.011
Unclear as to what figures this analysis relates to 		Effect size not calculated face-to-face CBT

RCT		 128 adults with BN or EDNOS HAM-D scores (mean) at end of treatment
10.6 with telemedicine CBT
7.0 with face-to-face CBT
Absolute results not reported
P = 0.008 		Effect size not calculated face-to-face CBT

RCT		 128 adults with BN or EDNOS Beck Depression Inventory scores at end of treatment
with telemedicine CBT
with face-to-face CBT
Absolute results reported graphically
Reported as “no differences were found on level or pattern of change in BDI scores”
P value not reported
Open in a new tab

Quality of life

Delivery of therapy online compared with delivery of therapy not online We don’t know whether CBT delivered by telemedicine is more effective than CBT delivered face to face at improving quality of life (measured by SF-36 physical and mental component scores) at 16–52 weeks in a mixed population of adults with bulimia nervosa or eating disorder not otherwise diagnosed (very low-quality evidence).

Ref (type) Population Outcome, Interventions Results and statistical analysis Effect size Favours
Quality of life scores

RCT		 128 adults with bulimia nervosa (BN) or eating disorder not otherwise specified (EDNOS) SF-36 (scale not defined), mean physical component scores at end of treatment
54.1 with telemedicine CBT
56.2 with face-to-face CBT
Reported as “no differences between groups were found for SF-36 scores”
P value not reported

RCT		 128 adults with BN or EDNOS SF-36 (scale not defined), mean mental component scores at end of treatment
42.9 with telemedicine CBT
45.5 with face-to-face CBT
Reported as “no differences between groups were found for SF-36 scores”
P value not reported

RCT		 128 adults with BN or EDNOS SF-36 (scale not defined), mean physical component scores at 12 months
53.6 with telemedicine CBT
55.4 with face-to-face CBT
Reported as “no differences between groups were found for SF-36 scores”
P value not reported

RCT		 128 adults with BN or EDNOS SF-36 (scale not defined), mean mental component scores at 12 months
43.5 with telemedicine CBT
42.7 with face-to-face CBT
Reported as “no differences between groups were found for SF-36 scores”
P value not reported
Open in a new tab

Adverse effects

No data from the following reference on this outcome.

Further information on studies

The RCT of online CBT versus waiting list control reported that the online intervention did significantly improve outcomes for a range of secondary measures assessed by the Eating Disorder Questionnaire (EDQ), including restraint, eating concern, shape concern, weight concern, and total score (total score, effect size 1.09, 95% CI 0.61 to 1.57). It also significantly improved depression score (measured by HADS; effect size 0.97, 95% CI 0.49 to 1.45) and anxiety (measured by HADS; effect size 0.82, 95% CI 0.33 to 1.31). The RCT noted that to test the success of blinding, assessors were asked to guess the treatment group. The treatment group was guessed correctly in 69% of cases (i.e., blinding was not completely successful). Of the people randomised to the active intervention, 8/38 (21%) did not complete any session, while overall 67/76 (88%) people completed eating disorder examination scores and 51/76 (67%) returned data for other secondary outcomes (e.g., HADS).

The RCT did not use strict DSM-IV criteria; however, this was a purely online approach with community recruitment, and all people taking part had recurrent bingeing, weight control behaviours, and concerns with high levels of symptoms and would, therefore, probably meet DSM-5 criteria. The RCT also found no significant difference between the bibliotherapy and waiting list control groups at 20 weeks (bibliotherapy v waiting list: EDE-Q, P = 0.99; binge eating, P = 0.99; purging, P = 0.99).

The RCT also reported that there were no differences in abstinence and remission between the two groups at 7 months and 18 months, but these data were based on less than 80% of participants with no ITT analysis for these results, so we have not reported these results further.

The RCT noted the substantial drop-out rate in the study, which is common in bulimia nervosa treatment trials, but noted that it was particularly important for this trial as it reduced the ability to detect any differences between treatment groups in outcomes. Nonetheless, it noted that abstinence rates at the end of treatment were similar to the rates seen in other large multi-centre trials using CBT.

Comment

In this option we have included interventions delivered online (by the internet, by text, or e-mail) that included contact with a person as part of the intervention. In the area of investigating online therapies, it is often hard to make rigorous diagnoses as therapy is not being delivered in a clinic, and in the real world, such online therapies rely on a degree of self-referral and self-report of symptoms, especially where used for initial therapy versus relapse prevention – but if the aim is to demonstrate equivalence with face-to-face therapy, then accurate diagnosis is important in design.

In the RCT with telemedicine delivered versus face-to-face therapy, it is possible that the intensity of treatment given was much higher (and may nearly work as face-to-face therapy) compared to other forms of online therapy where the only feedbacks are given via e-mail or an avatar. Therefore, elements of this interaction may be different from less intensive interaction, and there should be some caution when interpreting the results of this study and in generalising these results to other online therapies.

Since the search date of this review, a further systematic review has been published. This will be assessed at the next update of this BMJ Clinical Evidence review.

Clinical guide

Guided self-help has been evaluated in an online format. Although there is as yet insufficient evidence of efficacy from controlled trials, it appears safe and acceptable to people with bulimia nervosa and has potential to increase accessibility of treatment.

Substantive changes

Delivery of therapy online (involving contact with a person online) New option. Two systematic reviews were added, which included three RCTs and one additional RCT Categorised as 'unknown effectiveness'.

BMJ Clin Evid. 2015 Mar 4;2015:1009.

Delivery of self-help online (not involving contact with a person as part of intervention)

Summary

One RCT found no evidence that a self-help therapeutic writing task delivered online was any more effective than a control writing task delivered online.

However, the RCT was small, and participants were included with a Bulimic Investigatory Test-Edinburgh (BITE) score of 10 or above and did not need a formal diagnosis of bulimia nervosa.

This may limit the generalisability of results to people with bulimia nervosa accurately diagnosed.

Benefits and harms

Delivery of self-help online versus placebo, sham therapy online, or waiting list control:

We found two systematic reviews. One review (search date 2013) found no RCTs of sufficient quality, while the second review (search date 2012) found one RCT. The RCT (94 people) identified by the review compared therapeutic writing as an intervention for symptoms of bulimia nervosa (BN) with a control writing task. Recruitment and evaluation was done by e-mail, with no interview or face-to-face contact. Participants were recruited primarily through UK universities and colleges, through e-mails, posters, and website advertisements. No formal assessment of participants against DSM criteria was reported. Participants aged at least 18 years were required to score at or above the medium range cut-off for bulimic symptoms (10 or above) on the Bulimic Investigatory Test-Edinburgh (BITE; 33-item self-rating measure assessing symptoms of BN). Individuals in treatment for an eating disorder were excluded, as were those reporting suicidal thoughts, risk of self-harm, serious physical illness, substance misuse, or with a BMI of less than 18.5. Participants were also reimbursed £10 for participation. The therapeutic intervention was unsupported self-directed writing. Participants were asked to set aside about 20 minutes for 3 consecutive days to write about "the things others don’t see, hear, or notice about me" and some instructions incorporated from the Pennebaker task among other elements, with subsequent instructions to further explore their emotions and thoughts. Participants in the control group were asked to write about superficial topics in a factual manner without exploring thoughts and feelings. In total, 94 people were randomised and 80 (85%) people were included in the analysis by the RCT. The RCT reported that in total, eight people discontinued the intervention in the therapeutic writing group, and two were lost to follow-up and a further four people discontinued the intervention in the control writing group. Of the remaining 80 people, the RCT reported that the average age was 29 years, 71 (89%) participants were female, mean BMI was 25.7, and mean BITE total score was 29.5. The review noted that it was unclear whether the outcome assessment in the RCT was blinded. The review re-calculated data based on an ITT analysis. We have presented these data rather than the original analysis from the RCT.

Symptom improvement

Delivery of self-help online compared with placebo, sham therapy online, waiting list control, or control We don’t know whether a therapeutic writing task delivered by e-mail is more effective than a control writing task delivered by e-mail at improving symptoms at 4 weeks in people with medium-severity or above bulimic symptoms assessed by BITE score but with no formal DSM-IV diagnosis of bulimia nervosa (very low-quality evidence).

Ref (type) Population Outcome, Interventions Results and statistical analysis Effect size Favours
Symptoms

Systematic review		 94 people reporting symptoms of bulimia nervosa, assessed by BITE scores
Data from 1 RCT		 Bulimic Investigatory Test-Edinburgh (BITE) severity score at 4 weeks
with therapeutic writing task
with control writing task
Absolute results not reported
Effect size +0.28
95% CI –0.13 to +0.69 		Not significant

Systematic review		 94 people reporting symptoms of bulimia nervosa, assessed by BITE scores
Data from 1 RCT		 Bulimic Investigatory Test-Edinburgh (BITE) symptom score at 4 weeks
with therapeutic writing task
with control writing task
Absolute results not reported
Effect size +0.03
95% CI –0.37 to +0.43 		Not significant

Systematic review		 94 people reporting symptoms of bulimia nervosa, assessed by BITE scores
Data from 1 RCT		 Bulimic Investigatory Test-Edinburgh (BITE) total score at 4 weeks
with therapeutic writing task
with control writing task
Absolute results not reported
Effect size +0.13
95% CI –0.28 to +0.54 		Not significant

Systematic review		 94 people reporting symptoms of bulimia nervosa, assessed by BITE scores
Data from 1 RCT		 HADS depression score after writing task
with therapeutic writing task
with control writing task
Absolute results not reported
Effect size +0.18
95% CI –0.23 to +0.59 		Not significant

Systematic review		 94 people reporting symptoms of bulimia nervosa, assessed by BITE scores
Data from 1 RCT		 HADS anxiety score after writing task
with therapeutic writing task
with control writing task
Absolute results not reported
Effect size +0.02
95% CI –0.38 to +0.42 		Not significant
Open in a new tab

Quality of life

No data from the following reference on this outcome.

Adverse effects

No data from the following reference on this outcome.

The delivery of self-help online versus the delivery of self-help not online:

We found two systematic reviews (search date 2013; 2012), which found no RCTs. We found no subsequent RCTs.

Further information on studies

The study also reported a qualitative analysis of feedback and transcripts. This reported enhanced emotional wellness, increased understanding, and motivation to change in the therapeutic writing task group. The control group also had beneficial effects of distraction and enhanced coping. In total, 26 participants in the therapeutic writing task group (72% of those who expressed a preference) expressed a preference for internet delivery as opposed to face-to-face discussion, feeling that this promoted disclosure and caused less shame, and the study found that those with high levels of body shame were most likely to benefit. However, there was no face-to-face group in the study to provide any direct comparison.

Comment

In this option we have included interventions delivered online (by the internet or by text or e-mail) that did not include contact with a person as part of the intervention. That is, the intervention was self-help with no additional personal contact. In the area of investigating online therapies it is often hard to make rigorous diagnoses, as therapy is not being delivered in a clinic, and in the real world such online therapies rely on a degree of self-referral and self-report of symptoms, especially where used for initial therapy versus relapse prevention. But if the aim is to demonstrate equivalence with face-to-face therapy, then accurate diagnosis is important in design.

Since the search date of this review, a further systematic review has been published. This will be assessed at the next update of this BMJ Clinical Evidence review.

Clinical guide

Unsupported self-help therapies have insufficient evidence of efficacy and appear to have low effect sizes. Their role in treatment of bulimia nervosa is unclear, and they may delay accessing of more efficacious and comprehensive treatment.

Substantive changes

The delivery of self-help online (not involving contact with a person as part of intervention) New option. Two systematic reviews added, which included one RCT. Categorised as 'unknown effectiveness'.

BMJ Clin Evid. 2015 Mar 4;2015:1009.

Apps or online programmes used as an adjunct to face-to-face therapy

Summary

One RCT found limited evidence that a 16-week text messaging intervention (SMS) that delivered a tailored feedback message may improve abstinence and reduce full bulimic symptoms at 8 months, compared with treatment as usual, in women who had been discharged after specialist inpatient care.

However, the trial included a mixed population of women with full and sub-threshold bulimia nervosa, as well as women with eating disorder not otherwise specified.

In addition, the trial was limited to one hospital site, and used a non-standard intervention. This may limit its generalisability to women with bulimia nervosa alone or to other settings.

Benefits and harms

Apps or online programmes used as an adjunct to face-to-face therapy versus no adjunct therapy:

We found two systematic reviews (search date 2013; 2012), which found no RCTs of sufficient quality. We found one subsequent RCT. The reviews found no RCTs that evaluated an online therapy given at the same time as initial therapy. The subsequent RCT included 165 women aged 18 years and older who had been admitted to the same hospital for specialised inpatient care based on CBT. After discharge, women were randomised to a programme delivered via a short message service (SMS) intervention versus treatment as usual (during inpatient treatment, therapists routinely recommended that people utilise outpatient care; however, it was up to the patient to initiate this). Inclusion criteria for the trial included full or sub-threshold bulimia nervosa (BN) as defined by DSM-IV, or meeting the diagnostic criteria of eating disorder not otherwise specified (EDNOS), with a minimum of two binge eating episodes per week for a minimum duration of 1 month. All participants met level 3 criteria, psychiatric status ratings (PSR) 5 or 6, according to the Longitudinal Interval Follow-up Evaluation. About 56% and 61% in the SMS and control groups, respectively, had a diagnosis of BN, and 44% and 39% in SMS and control groups, respectively, had a diagnosis of EDNOS. At baseline, the average age was 30 years, 34% to 36% of women had an illness duration over 10 years, all had had binge eating prior to admission (defined as twice or more per week for a minimum of 4 weeks), and the average duration of in-patient treatment was 55 days. The text messaging intervention consisted of a weekly interaction for 16 weeks after discharge. Participants used their private mobile phones to report on body dissatisfaction, frequency of binge eating, and frequency of compensatory behaviours, and the software followed an algorithm to provide a tailored feedback message, which was sent to the mobile phone by a research assistant. In total, 10 to 15 messages were available for 64 possible symptom patterns. The feedback messages signalled social support and interest, reinforced improvements in symptoms, and reminded participants of CBT strategies they had learned to cope with difficult situations. Of 165 people initially randomised, 25 (24%) were lost to follow-up. The RCT reported both a completer and ITT analysis. We have reported the ITT analysis here. Abstinence was defined as absence of binge eating and compensatory behaviours for a minimum of 4 weeks, at 8 months, while remission was defined as a maximum of one binge episode per week and maximum of one compensatory behaviour per week for a duration of not less than 4 weeks, at 8 months. About half of the people in each group utilised outpatient treatment during the 8-month follow-up period. This meant that the remaining people in the control group had no professional support after discharge.

Symptom improvement

Apps or online programmes used as an adjunct to face to face therapy compared with no adjunct therapy A specific SMS intervention providing a tailored feedback message may be more effective than treatment as usual in improving abstinence levels at 8 months and reducing the proportion of women who still had full symptoms (still met DSM-IV criteria for bulimia nervosa or eating disorder not otherwise specified) in a mixed population of women with bulimia nervosa or eating disorder not otherwise specified who had just been discharged from hospital following in-patient care, but we don’t know about remission, and evidence was weak (very low-quality evidence).

Ref (type) Population Outcome, Interventions Results and statistical analysis Effect size Favours
Symptoms

RCT		 165 women with bulimia nervosa (BN) or eating disorder not otherwise specified (EDNOS) Abstinence (absence of binge eating and compensatory behaviours for a minimum of 4 weeks) at 8 months
31/82 (37.8%) with text intervention
15/83 (18.1%) with treatment as usual
P <0.01
Outcome assessed using self-report information via telephone interview; degree of blinding unclear 		Effect size not calculated text intervention

RCT		 165 women with BN or EDNOS Remission (maximum of one binge episode per week and maximum of one compensatory behaviour per week for a duration of not less than 4 weeks) at 8 months
42/82 (51.2%) with text intervention
30/83 (36.1%) with treatment as usual
P = 0.05
Borderline significant
Outcome assessed using self-report information via telephone interview; degree of blinding unclear 		Effect size not calculated text intervention

RCT		 165 women with BN or EDNOS Full symptoms (still met DSM-IV diagnostic criteria for BN or EDNOS) at 8 months
27/82 (32.9%) with text intervention
45/83 (54.2%) with treatment as usual
P <0.01
Outcome assessed using self-report information via telephone interview; degree of blinding unclear 		Effect size not calculated text intervention
Open in a new tab

Quality of life

No data from the following reference on this outcome.

Adverse effects

No data from the following reference on this outcome.

Further information on studies

The intervention was a non-standard intervention, and the RCT was limited to one hospital. Hence, the generalisability may be limited because of site-specific and setting-specific factors, as well as issues surrounding the reproducibility of the intervention.

Comment

In this option we have included RCTs of apps or online programmes (via the internet, by text, or e-mail) used as an adjunct to therapy (for example, recording diet, or recording purges, mood, additional prompts, etc). In the area of investigating online therapies it is often hard to make rigorous diagnoses, as therapy is not being delivered in a clinic, and in the real world such online therapies rely on a degree of self-referral and self-report of symptoms, especially where used for initial therapy versus relapse prevention. But if the aim is to demonstrate equivalence with face-to-face therapy, then accurate diagnosis is important in design.

Since the search date of this review, a further systematic review has been published. This will be assessed at the next update of this BMJ Clinical Evidence review.

Clinical guide

The use of online apps as adjunct to standard care is likely to increase in the future. While they as yet lack sufficient evidence, they appear promising.

Substantive changes

Apps or online programmes used as an adjunct to face-to-face therapy New option. Two systematic reviews added, and one subsequent RCT. Categorised as 'unknown effectiveness'.

Articles from BMJ Clinical Evidence are provided here courtesy of BMJ Publishing Group

RESOURCES