Introduction
Obesity and eating disorders are both complicated by serious, short- and long-term health problems(1–Reference Steinbeck, Lister and Gow3). The prevalence of both is increasing(4,Reference Galmiche, Déchelotte and Lambert5) , with some data suggesting rates of combined obesity and eating disorders are increasing faster than the prevalence of either obesity or eating disorders alone(Reference Da Luz, Hay and Touyz6). For example, between 1995 and 2015 in a community sample of Australian adults, prevalence of obesity alone increased 1·7-fold, and binge eating episodes increased 3·5-fold, while prevalence of combined obesity and recurrent binge eating episodes increased 5·7-fold(Reference da Luz, Sainsbury and Mannan7). Co-existence of obesity and a range of eating disorders is seldom acknowledged(Reference Haines, Kleinman and Rifas-Shiman8), with these conditions commonly stereotyped as existing on opposite ends of an eating disordered spectrum. A key distinction is that obesity is defined by a physical metric, whereas eating disorders are defined by well-characterised cognitive and behavioural phenotypes(9). Obesity and eating disorders share many risk factors (e.g. weight concern, dieting)(Reference Haines, Kleinman and Rifas-Shiman8) and treatment approaches (e.g. health professional support, self-monitoring, goal setting and normalising eating patterns)(Reference Lister, Baur and Paxton10). Of concern, the focus on weight loss during obesity treatment may lead to under-diagnosis of eating disorders in people with obesity(Reference Da Luz, Hay and Touyz6). While only a small proportion of individuals may develop or have exacerbated eating disorder symptoms during weight management(Reference Jebeile, Gow and Baur11), the potential burden for this risk is high. Nevertheless, there is a paucity of research and limited treatment pathways for those affected by both obesity and eating disorders. Additionally eating disorders may develop over several years(Reference Stice and Van Ryzin12), and weight management interventions may be only a single experience in the eating disorder development pathway. However, this is a unique point of engagement with health services, where risk can be identified and addressed, thus representing an important research and practice gap.
The aim of this review is to provide the background, rationale and study designs for the Eating Disorders In weight-related Therapy (EDIT) Collaboration. We describe considerations for the nuances of eating disorder development, including risk factors from observational studies, and describe how such risk factors may be influenced by behavioural weight management. The EDIT Collaboration will bring together individual participant data (IPD) from relevant trials to understand how individual characteristics and components of weight management interventions may contribute to eating disorder risk. We hypothesise that individuals undertaking weight management interventions will vary in their eating disorder risk profile, on the basis of personal characteristics and intervention strategies available to them. Further, we propose individual characteristics can be identified and intervention strategies can be adapted to reduce eating disorder risk.
State of the literature: obesity, weight management and eating disorders
Eating disorder prevalence and complications in people with obesity
Worldwide, in 2016, 39% of men and 40% of women were affected by overweight, while 11% of men and 15% of women were affected by obesity(13); and prevalence is predicted to rise by the year 2030(Reference Lobstein and Brinsden14). For children and adolescents (5–19 years), the prevalence of obesity was 5·6% in girls and 7·8% for boys(4). Complications of obesity include type 2 diabetes, non-alcoholic fatty liver disease, cardiovascular disease, sleep apnoea and depression(Reference Bray, Kim and Wilding15,Reference Pereira-Miranda, Costa and Queiroz16) . There is growing evidence that prevalence of obesity is higher in populations with culturally diverse and lower socioeconomic backgrounds(13,Reference Bann, Johnson and Li17) .
Eating disorders include anorexia nervosa, atypical anorexia nervosa, bulimia nervosa, binge eating disorder and several other categories of feeding and eating disorders(18). Eating disorders are severe mental and physical health conditions with a long duration(Reference Schmidt, Adan and Böhm19) and high morbidity(Reference Klump, Bulik and Kaye2,Reference Arcelus, Mitchell and Wales20,Reference Ágh, Kovács and Supina21) . Despite misconceptions that eating disorders are diseases of individuals with lower body weight, the prevalence of eating disorders is higher in both men and women with obesity compared with their healthy weight peers(Reference Duncan, Ziobrowski and Nicol22). In a survey of 12 337 adults in the United States, the lifetime prevalence, that is, the proportion of people who had any eating disorder at any point in their life, was 2·2% in men and 4·9% in women(Reference Duncan, Ziobrowski and Nicol22). However, men and women with obesity had a higher prevalence of eating disorders compared with the general population, at 3·8% and 7·6%, respectively(Reference Duncan, Ziobrowski and Nicol22). Similarly in adolescents, the prevalence of eating disorders is associated with higher body mass index (BMI). For example, a study of 3043 Canadian adolescents found 9·3% of male adolescents and 20·2% of female adolescents with obesity had a sub- or full-threshold eating disorder compared with respectively 2·1% and 8·4% of adolescents with a BMI in the normal range(Reference Flament, Henderson and Buchholz23). Data from 5191 Australian adolescents show those with overweight or obesity were more likely to experience an eating disorder(Reference Mitchison, Mond and Bussey24). Further, several studies have identified an increase in disordered eating behaviours over time in community samples(Reference da Luz, Sainsbury and Mannan7,Reference Darby, Hay and Mond25) , particularly among individuals with overweight and obesity. It has been reported that approximately one in four adolescents with obesity engages in binge eating behaviours or experiences loss of control with eating(Reference He, Cai and Fan26,Reference Goossens, Braet and Decaluwé27) . Binge eating and a loss of control with eating are also associated with weight gain and symptoms of the metabolic syndrome and are important drivers of continuing weight gain(Reference Klump, Bulik and Kaye2,Reference Tanofsky-Kraff, Shomaker and Stern28) . Moreover, emerging evidence suggests sociocultural factors such as food insecurity and childhood adversity influence both eating disorders(Reference Huryk, Drury and Loeb29,Reference Rasmusson, Lydecker and Coffino30) and obesity(Reference Schroeder, Schuler and Kobulsky31,Reference St. Pierre, Ver Ploeg and Dietz32) , and both eating disorder symptoms and weight gain have been exacerbated by the coronavirus disease 2019 pandemic lockdowns(Reference Sideli, Lo Coco and Bonfanti33–Reference Phillipou, Meyer and Neill35).
Compared with the general population, individuals with an eating disorder have an elevated risk of premature mortality, with a German study showing that the highest mortality risk (standardised mortality ratio) associated with anorexia nervosa was 5·35, compared with bulimia nervosa (1·49) and binge eating disorder (1·50)(Reference Fichter and Quadflieg36). Indeed, complications of anorexia nervosa are equally severe for individuals with weight within or above the normal range (i.e. atypical anorexia nervosa compared with anorexia nervosa)(Reference Walsh, Hagan and Lockwood37). People with binge eating disorder experience a high prevalence of both psychiatric comorbidities (e.g. mood, anxiety and substance use disorders)(Reference Araujo, Santos and Nardi38,Reference Goossens, Braet and Van Vlierberghe39) and physical comorbidities (e.g. type 2 diabetes, hypertension and chronic pain)(Reference Mitchell40). These comorbidities are also associated with obesity(Reference McCuen-Wurst, Ruggieri and Allison41,Reference Thornton, Watson and Jangmo42) ; however, higher weight only partly explains the association with binge eating disorder(Reference Thornton, Watson and Jangmo42). Almost 30% of adults with binge eating disorder also report a history of childhood obesity(Reference Jacobi, Hayward and de Zwaan43). Thus, identifying eating disorders and reducing risk during weight management has potential to reduce or prevent a range of physical and psychological complications.
While there may be a growing recognition that people with obesity are at increased risk of developing eating disorders(Reference Da Luz, Hay and Touyz6,Reference Jebeile, Lister and Baur44) , there has been limited progress in the identification, prevention and treatment of eating disorders in the context of weight management.
Behavioural weight management
Multicomponent behavioural interventions are first-line treatment for adolescents and adults affected by obesity(45–Reference Alman, Lister and Garnett47). These interventions typically recommend a combination of diet physical activity and behavioural modifications. A 2018 systematic review of adult behavioural weight management interventions of at least 12 months in duration showed that they were likely to produce significantly more weight loss compared with standard care (mean difference in weight change (MD), −2·39 kg (95% CI −2·86, −1·93); 67 studies; n = 22,065). Eligible studies included participants recruited from primary care or a health care system, and intervention groups experienced less weight regain during the follow-up periods (beyond 12–18 months, pooled MD compared with control, −1·59 kg (95% CI −2·38, −0·79); n = 1408)(Reference LeBlanc, Patnode and Webber48). Further, the risk of developing diabetes over 1–9 years was substantially reduced (pooled risk ratio, 0·67 (95% CI 0·51, 0·89); 9 trials; n = 3140). Clinical practice guidelines for adolescents with overweight or obesity recommend a family-based approach to multicomponent behavioural interventions that address dietary, sedentary and sleep behaviours(Reference Alman, Lister and Garnett47). A 2017 Cochrane review(Reference Al-Khudairy, Loveman and Colquitt49) found multicomponent behavioural interventions for adolescents (aged 12–17 years) with overweight or obesity resulted in a mean change in body weight of −3·67 kg (95% CI −5·21, −2·13; 20 trials; n = 1993) and BMI of −1·18 kg/m2 (95% CI −1·67, −0·69; 28 trials; n = 2774). These effects were maintained at 24-month follow-up. A 2012 systematic review reported significant improvements in low-density lipoprotein cholesterol (−0·30 mmol/L, 95% CI −0·45, −0·15), triglycerides (−0·15 mmol/L, 95% CI −0·24, −0·07), fasting insulin (−55·1 pmol/L, 95% CI −71·2, −39·1) and blood pressure up to 1 year from baseline following lifestyle interventions for children and adolescents. Hence, multicomponent intensive behavioural weight management interventions can effectively reduce body weight and cardiometabolic risk in both adolescents and adults.
An emerging area of research and practice are dietary interventions with the potential to induce greater weight loss and improve cardiometabolic complications of obesity(Reference Mulholland, Nicokavoura and Broom50–Reference Parretti, Jebb and Johns52). These interventions may include very-low-energy diets (<800 kcal/d), very-low-carbohydrate diets (<50 g carbohydrate/d) or intermittent energy restriction, and require both medical and dietetic supervision(Reference Hoare, Jebeile and Garnett53–Reference Gow, Garnett and Baur55). Such interventions are recommended for adolescents and adults with obesity and associated complications, or with severe obesity(Reference Alman, Lister and Garnett47,Reference Jensen, Ryan and Apovian56) . These restrictive approaches play an important role in effectively managing weight and cardiometabolic risk, particularly as an obesity management approach prior to bariatric surgery or when pharmacological and surgical approaches are not available or contraindicated. However, the effect of these interventions on eating disorder risk is unclear.
Weight management and eating disorder risk
There is concern that dietary interventions, the cornerstone of behavioural weight management, may promote disordered eating and worsen psychological health(57,Reference da Luz, Hay and Gibson58) in some individuals. This is informed by longitudinal data showing that dieting is an important step within eating disorder development(Reference Neumark-Sztainer, Wall and Guo59,Reference Stice, Marti and Durant60) . However, some of these studies have poorly characterised the population sampled and the definition of ‘dieting’. Nevertheless, energy restriction may trigger binge eating in some people, and it is thought that dietary and weight monitoring may trigger a preoccupation with food, weight and shape. Data from intervention studies including dietary components are described below.
Evidence from systematic reviews
Systematic reviews have examined the association between behavioural weight management and the change in eating disorder risk. A 2017 systematic review examining weight management interventions for adults identified five randomised controlled trials (RCTs), all of which reported beneficial outcomes for eating disorder symptoms, including a reduction in binge eating(Reference Peckmezian and Hay61). Similarly, systematic reviews of pre-post studies and RCTs conducted in children and adolescents found no change or a small reduction in eating disorder symptoms, including binge eating and loss of control, following behavioural weight management(Reference Jebeile, Gow and Baur11,Reference Moustafa, Quigley and Wadden62–Reference De Giuseppe, Di Napoli and Porri64) . Other related eating disorder risk factors, including depression, anxiety, and low body image and self-esteem, have also been improved following weight management in both adults and adolescents(Reference Fabricatore, Wadden and Higginbotham65–Reference Jones, Lawlor and Birch69). The effect of restrictive dietary approaches, including low- or very-low-energy diets, on binge eating has been examined in a systematic review of ten studies including 805 adults(Reference da Luz, Hay and Gibson58). In participants with pre-treatment binge eating disorder, studies reported a reduction in binge eating. The evidence was mixed in studies with participants with sub-clinical or no binge eating symptoms prior to treatment, with some studies showing a reduction in symptoms and others showing no change or an increase. Importantly, two studies included in this review reported an increase in binge eating or the onset of binge eating disorder in 10–15% of participants(Reference da Luz, Hay and Gibson58). To our knowledge, the evidence of restrictive dietary approaches on eating disorder risk in adolescents is yet to be synthesised. These interventions are likely to include delivery features very different from behavioural weight management interventions included in the reviews in adults above, and the implications of this are unknown.
In summary, previous evidence suggests that eating disorder risk is reduced for most participants following professionally supervised behavioural weight management; however, individual studies have reported a small subset of participants who develop an eating disorder during the intervention or in the years following intervention(Reference Epstein, Valoski and Wing70–Reference Telch and Agras75). For example, scores increased to above a clinical cut-point in seven children in one study, and three of fifty-six participants followed up at 6 years had developed binge eating disorder in another study(Reference Goossens, Braet and Verbeken71,Reference Epstein, Paluch and Saelens72) . The mechanisms by which behavioural weight management may increase or decrease eating disorder risk at the individual level are not clear.
Individual variation in responses to weight management interventions
Few studies have investigated individuals’ characteristics for associations with eating disorder outcomes following behavioural weight management interventions. For some individuals within the general population, dietary restraint (a proxy marker of dieting behaviours) is associated with the development of symptoms of binge eating disorder and bulimia nervosa(Reference Stice76,Reference Decaluwé and Braet77) . In contrast, in others it may be an important behaviour that enables improvements in weight management and cardiometabolic health(Reference Schaumberg, Anderson and Anderson78). This divergence in response may be in part explained by the difficulty in distinguishing between flexible (i.e. gradual reduction, foods are limited in quantity rather than eliminated) and rigid restraint (extreme, all-or-nothing mentality)(Reference Schaumberg, Anderson and Anderson78). Psychosocial predictors thought to play a role in the development of eating disorders include poor self-esteem, depression, anxiety, bulimic behaviours (i.e. binge eating with compensatory behaviours), body dissatisfaction and drive for thinness(Reference Stice, Marti and Durant60,Reference Russo, Brennan and Walkley79) . It is possible that the interactions or clustering of such individual characteristics with dietary restraint may be important for eating disorder development. For example, Stice’s dual pathway model hypothesises that pressure for thinness increases risk for body dissatisfaction, which in turn increases the risk for dietary restriction and/or negative affect, thereby increasing the risk for subsequent onset of binge eating-related disorders(Reference Wonderlich, Mitchell and Steiger80). However, aetiological models including these risk factors do not consistently predict onset of eating disorders(Reference Stice and Van Ryzin12), suggesting varied individual responses. Importantly, research identifying eating disorder risk factors has been conducted in predominantly healthy weight populations; risk factors specific to individuals with obesity for the full spectrum of eating disorder diagnoses are needed(Reference Jebeile, Lister and Baur44,Reference Hilbert, Pike and Goldschmidt81,Reference Striegel-Moore, Fairburn and Wilfley82) .
Beyond individual characteristics, there may be components of behavioural weight management interventions that influence eating disorder risk. Behavioural weight management interventions typically include a combination of intervention strategies (e.g. related to diet, movement, eating behaviour), delivered through various approaches (e.g. in terms of mode of delivery, session frequency/duration). Some commonly used strategies within weight management interventions are considered disordered behaviours in the context of eating disorder development, or may be components of aetiological models of eating disorders. Two examples are the restriction of energy intake and self-monitoring of weight(Reference Pacanowski, Linde and Neumark-Sztainer83). In behavioural weight management interventions, the prescription of restriction of energy intake (i.e. reduced food intake) is thought to equate to dietary restriction, while a focus on monitoring of weight is likened to pressure for thinness and weight preoccupation. Hence these two components of behavioural weight management are thought to tie into the dual pathway sequence to promote disordered eating behaviours(Reference Jebeile, Lister and Baur44). This is a contentious theory, with intervention studies of restricted energy intake for up to 2 years in individuals without obesity (such as the CALERIE trials(Reference Williamson, Martin and Anton84,Reference Dorling, van Vliet and Huffman85) ) demonstrating increases in dietary restraint without increases in binge eating or eating disorders(Reference Stewart, Martin and Williamson86). Indeed, this leads to the conceptualisation of dietary restraint as a successful self-regulation strategy. Schaumburg et al. (Reference Schaumberg, Anderson and Anderson78) posit that, when this self-regulation fails and is followed by a period of disinhibition, eating disorder risk is increased(Reference Schaumberg, Anderson and Anderson78).
Moreover, intervention components such as self-monitoring improve weight-related outcomes and long-term weight maintenance for most people(Reference Burke, Wang and Sevick87). Large-scale behavioural programmes, such as the Diabetes Prevention Program (DPP) and Look AHEAD, demonstrate that self-monitoring of weight and dietary intake is positively related to weight loss(Reference Knowler, Fowler and Hamman88–90); however, these factors are also associated with disordered eating in community samples(Reference Pacanowski, Linde and Neumark-Sztainer83). Furthermore, factors such as increased self-esteem and self-efficacy around healthy eating are thought to be protective against disordered eating(Reference Neumark-Sztainer, Wall and Haines91), suggesting that changes in individual characteristics may mediate changes in eating disorder risk during interventions. Indeed, eating disorder treatment programmes may also include components of regular weighing and the development of healthy eating behaviours. This highlights the need to understand the complexities of interevention components of weight management interventions for eating disorder risk. There are also many components of behavioural weight management interventions which may be protective against disordered eating, including regular contact and support from healthcare professionals, and strategies such as promotion of self-efficacy, realistic goal setting, supported problem solving, and strategies for stimulus control or social support(Reference Hayes, Fitzsimmons-Craft and Karam92). Thus, a complex interaction between individual characteristics and intervention components is likely to influence eating disorder risk during weight management. To inform models of care for clinical practice it is vital to identify and understand components that may increase or decrease eating disorder risk for different individuals who participate in professionally supervised behavioural weight management programmes. Addressing this is the central goal of the EDIT Collaboration.
Stakeholder engagement in obesity and eating disorder research
To identify individual- or intervention-level attributes that influence eating disorder risk in the context of weight management, it is essential to bring together stakeholders including those with lived experience of obesity and eating disorders. Engagement of consumers in research development and dissemination is an important process for evidence-based medicine(Reference Qaseem, Forland and Macbeth93–Reference Jarrett and Unit95).
The engagement of consumers in obesity and eating disorder research must be sensitive to the stigma and potential harms for those affected by both conditions. Data on the lived experience of people with combined obesity and eating disorders are limited. However, qualitative studies in people with lived experience of obesity frequently identify the harm caused by obesity stigma and weight bias(Reference Pudney, Himmelstein and Puhl96,Reference Hollett and Carter97) . A 2017 systematic review of cross-sectional studies reported that more frequent weight stigma experiences were associated with poorer physiological and psychological health(Reference Wu and Berry98). Indeed, using the term obesity is an important consideration for researchers, with studies reporting mixed responses to the term from people with higher weights(Reference Brown and Flint99,Reference Puhl100) . Many people with obesity prefer neutral terms, such as ‘weight’ be used in clinical care. However, one study found all weight-related terms elicit negative emotions(Reference Brown and Flint99). On the other hand, obesity is a defined medical condition by a number of international health professional organisations, including the World Health Organization(101). Systematic reviews and international guidelines make suggestions and recommendations for reducing weight stigma and bias, such as the Joint international consensus statement for ending stigma of obesity(Reference Pont, Puhl and Cook102–Reference Rubino, Puhl and Cummings105). The EDIT Collaboration will use person-first language for scientific discourse, and language used within the collaboration will be reviewed and adjusted as required.
Summary and rationale for the EDIT collaboration
Behavioural weight management interventions form the first-line treatment approach for obesity(45). Clinical trials and systematic reviews show that, in addition to improved weight and cardiometabolic health, adolescents and adults with obesity who participate in supervised weight management interventions overall have improved eating behaviours and psychological outcomes(Reference Jebeile, Gow and Baur11,Reference Jebeile, Gow and Baur67,Reference Jones, Lawlor and Birch69,Reference Braet, Tanghe and Decaluwé106) . However, there is ongoing concern that these interventions may promote disordered eating and worsen psychological health in some individuals(57). Indeed, some studies report worsening eating disorder outcomes in some individual participants(Reference Jebeile, Gow and Baur11). This suggests that those who have poorer outcomes following weight management interventions are not captured when studies report aggregate risk scores. Due to the required large sample sizes and need for individual-level analysis (rather than pooling summary scores), the important research question of whether weight management increases or decreases eating disorder risk for an individual is difficult if not impossible to answer using a single trial, qualitative methods or standard aggregate data meta-analyses. Further, weight management interventions are complex and often poorly described. Research investigating whether certain intervention types or components of interventions may either increase or decrease eating disorder risk at an individual level is needed. It is likely that complex interactions between individual characteristics and intervention components influence eating disorder risk responses during weight management. Interventional evidence that addresses clinically supervised behavioural weight management for people with obesity should be examined to address these concerns.
Research programme
The EDIT Collaboration will bring together clinicians, researchers, biostatisticians and individuals with lived experience from around the world to improve treatment for people affected by obesity and eating disorders. The EDIT Collaboration aims to: (1) understand which participants experience a change in eating disorder risk, or related symptoms, during and following weight management interventions; (2) understand which intervention components may contribute to eating disorder risk; (3) identify predictive pathways for increased or decreased eating disorder risk during weight management; and (4) develop resources and recommendations to reduce eating disorder development during obesity treatment. To achieve these aims, the EDIT Collaboration will conduct five related studies (Fig. 1). Detailed methodologies will be published separately, but a brief overview is provided below.
Scientific and stakeholder engagement
The work of the EDIT Collaboration is guided by Scientific and Stakeholder Advisory panels with international representation. The Scientific Advisory Panel includes experienced researchers and clinicians from the fields of obesity and eating disorders or those working across both conditions. The panel is responsible for overall programme oversight and will provide strategic advice relating to the scientific rigour of each included study, contribute to protocol development and scientific publications, and guide project output and the translation of project outcomes. The Stakeholder Advisory Panel comprises consumers with a lived experience of eating disorders, obesity or both conditions. The Stakeholder Advisory Panel will provide strategic advice, contribute to protocol development, and guide project outputs and translation from the viewpoint of the end consumers of health services.
Eligible trials
Systematic searches of electronic databases and trial registries(Reference Hunter, Webster and Page107) are being conducted to identify trials that meet our inclusion criteria: (1) randomised controlled trial of behavioural weight management intervention; (2) for adolescents and/or adults with obesity; (3) report at least one measure of eating disorder symptoms or behaviours at baseline and post-intervention or follow-up using a validated self-report questionnaire (e.g. Eating Disorder Examination Questionnaire, Binge Eating Scale) and/or clinical assessment or diagnostic interview (e.g. Eating Disorder Examination). The protocol for this review is registered with PROSPERO (CRD42021265340), accessible from https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42021265340.
Representatives from each identified trial will be invited to join the collaboration and share IPD, that is, line-by-line data for each individual participant data. The corresponding authors of identified trials are invited to join the EDIT Collaboration via email. If, after two attempts, no response has been received, other authors on the paper or listed on a registration record will be emailed. Finally, we will attempt to contact trialists via telephone, using our networks, via their institutions or at conferences. If trialists are unable to be contacted after multiple attempts, the trial will be excluded, since our analyses are not possible using published summary data alone. Trial representatives will have the opportunity to provide input to all major stages of the project including protocol development, analysis, results interpretation and translation.
Study 1: consultation
We will identify individual participant characteristics and intervention strategies which may contribute to an increase or decrease in eating disorder risk during weight management interventions through broad stakeholder consultation. Using an online survey, we will canvass diverse opinions on possible causes of eating disorder development in weight management interventions. Participants will be asked to rate the relevance of individual characteristics (e.g. body dissatisfaction, history of self-directed dieting, disinhibition related to eating) and intervention strategies (e.g. dietary monitoring, dietary behaviour change strategies, informed by a psychological framework or theory) and identify any not listed in the survey. Individual characteristics and intervention strategies listed in the survey are informed by relevant literature as well as expert and consumer consultation (Scientific and Stakeholder Advisory Panels). The outcomes of this survey will be used to inform the analyses in the studies below.
Studies 2 and 3: individual participant data meta-analyses
Meta-analysis of IPD is considered the ‘gold standard’(Reference Stewart, Altman and Askie108) for meta-analysis, in part due to the opportunity to explore differences in treatment effects across subgroups. These subgroups might include subsets of participants, such as those with higher eating disorder risk at baseline, or subsets of studies, such as those with particular intervention strategies(Reference Sutton, Kendrick and Coupland109).
We will collate all available data from all eligible studies to examine the individual risk of eating disorder development during weight management trials. IPD from collaborating trials will be collated into a central database. Trials will be identified through systematic searches, investigator networks and study branding (editcollaboration.com). The specific variables to be included in the analysis will be informed by stakeholder consultation (study 1) and data availability, whereby any suggested predictor will be considered. We will conduct two IPD meta-analyses, with detailed methodology published elsewhere a priori. Study 2 aims to identify baseline participant risk factors which predict an increase or decrease in eating disorder risk, or related symptoms, during and following a weight management intervention. This will be a pre–post IPD meta-analysis. Study 3 aims to determine whether there are baseline participant risk factors which predict change in eating disorder risk or related symptoms, and if they receive any behavioural weight management intervention compared with no intervention (i.e. no treatment controls). This will be an IPD meta-analysis maintaining randomisation.
Study 4: intervention deconstruction
Interventions included in the EDIT Collaboration will be deconstructed into their discrete components to improve understanding of what they involve. Intervention components will include the intervention strategies to drive weight management (e.g. dietary or behaviour change strategies) and the features of how such interventions are delivered. Using a systematic coding framework developed for this study, we will then compare and synthesise the components of interventions targeting adolescent and adult populations. This project will allow for future quantitative analysis of intervention components and individual participant eating disorder risk.
Study 5: predictive modelling
Data from studies 2, 3 and 4 will be combined to identify any interactions between individual characteristics and intervention strategies which may increase or decrease the risk of eating disorders during weight management. Detailed methodology informed by the findings of studies 1–4 will be published separately.
Translation plan and recommendations
Knowledge gains (Box 1) will inform the translation plan and recommendations. Workshops with the Stakeholder Advisory Panel (which includes stakeholders who are consumers with lived experience of obesity and eating disorders), Scientific Advisory Panel and trial representatives will inform our translation action plan. A working group will be formed to develop plans for five key areas for translation: (i) models of care for obesity management that consider eating disorder risk; (ii) health professional education (e.g. training webinars, scientific publications, conferences, recommendations for screening and monitoring protocols); (iii) community dissemination (e.g. website, newsletter, community seminars, decision aids for informing treatment consent); (iv) identification of strategies to support further implementation; and (v) policy briefing documents summarising key evidence that emerges from the research.
Expected knowledge gains
Diverse community perspectives on the role of individual characteristics and intervention strategies that may increase or decrease eating disorder risk
Comprehensive understanding of how weight management interventions impact eating disorder risk at an individual level
Identify participant level predictors of change in eating disorder risk in the context of weight management
Understand intervention strategies used during weight management interventions and how they may contribute to eating disorder risk
Find interactions between individual characteristics and intervention components that contribute to an individual’s increase or decrease in eating disorder risk
Development of an international database of weight management interventions measuring eating disorder risk
Development of recommendations for tailored models of care for obesity treatment
Strengths and limitations
The strengths of this research are first the use of robust statistical methods to quantitatively examine individual-level and intervention-level eating disorder risk during behavioural weight management interventions. We will use statistical methods recommended by the Cochrane Collaboration(Reference Stewart, Altman and Askie108,Reference Stewart, Tierney and Clarke110) , led by a team of biostatisticians with previous IPD meta-analyses experience(Reference Askie, Espinoza and Martin111,Reference Seidler, Hunter and Baur112) . Secondly, this project incorporates consumers’ views and broader stakeholder engagement to set research priorities and to translate the outcomes of the study. Thirdly, trials on adolescents are included as an important life stage when the trajectories of both obesity and eating disorders become firmly established.
However, the studies outlined in this current research plan have limitations. This programme will not report on qualitative experiences of participants who have undertaken weight management interventions. The IPD relies on retrospective analysis of data from clinical trials. Thus, there is an inherent risk that adverse eating disorder outcomes are not captured due to missing data, higher participant attrition among those at risk, or insufficient follow-up of the included studies. However, IPD allow us to include data from excluded participants (e.g. outliers), more variables and timepoints from datasets that may not be included in a traditional aggregate data meta-analysis, thus increasing power to conduct subgroup analyses and detect adverse events. Further, where data are available, we will include in our analysis known psychosocial predictors of eating disorder development (e.g. self-esteem, depression, anxiety, bulimic behaviours, body dissatisfaction and drive for thinness) to identify potential changes in an individual’s risk profile that may precede changes in global eating disorder risk. Future research addressing qualitative experiences and a prospective data analysis will be important to complement the current research plan.
There is also the possibility that the clinical trials/interventions eligible for the EDIT Collaboration (i.e. including a validated comprehensive measure eating disorder risk) are not representative of broader weight management interventions and findings will not be generalisable to all weight management. Moreover, the withdrawal of interventions and support may also influence eating disorder risk, and long-term data may not capture this changing risk profile. All included trials are providing a weight management intervention; thus, whether eating disorder risk would differ for those not referred or enrolled in an intervention will not be determined. Instead, this project will identify adolescents and adults presenting for weight management: (i) for whom weight management will likely improve physical and mental health; (ii) for whom behavioural weight management is not recommended; and (iii) whether intervention components can improve outcomes for different individuals.
Summary and conclusion
The EDIT Collaboration will combine IPD meta-analysis and intervention coding to quantitatively explore the underlying pathways that increase or decrease eating disorder risk during behavioural weight management interventions. By understanding how individual participant characteristics may interact with intervention components to influence eating disorder risk, we have the potential to create an innovative toolbox for clinicians to build the safest interventions for each individual. Future combined research between obesity and eating disorder fields has the potential to lead to a tailored precision therapeutic response, improving both obesity and eating disorders care.
Acknowledgements
The authors would like to thank members of the TOPCHILD (Transforming Obesity Prevention for CHILDren) Collaboration for project support and feedback on study design.
N.B.L., H.J., L.A.B., S.J.P. and A.L.S. conceived the idea for the study. N.B.L., H.J., L.A.B., S.J.P., S.P.G., A.L.A., D.W., S.M., A.S., K.S., B.J.J. and A.L.S. developed the rationale and research question. N.B.L. wrote the first draft of the manuscript. H.J., L.A.B., S.J.P., S.P.G., A.L.A., D.W., S.M., A.S., K.S., C.B., A.J.H., D.N., R.A.J., G.D., A.M.G., K.C., T.K.K., F.A.H., F.Q., K.E.H., C.M., B.J.J. and A.L.S. provided critical review and feedback at each stage of the process. All authors critically revised the manuscript for intellectual content, and agreed and approved the final manuscript. N.B.L. is the guarantor of the review.
The EDIT collaboration is funded by the Australian National Health and Medical Research Council Ideas Grant (GTN2002310). N.B.L. is a recipient of a National Health and Medical Research Council Peter Doherty Early Career Fellowship (GTN1145748). A.L.A. and R.A.J. are supported by the Medical Research Council (MRC) (Grant MC_UU_00006/6). A.S. is funded by a Senior Research Fellowship (#1135897) from the NHMRC of Australia. D.N. is supported by the National Institute for Health Research (NIHR) under the Applied Health Research (ARC) programme for Northwest London. The views expressed in this publication are those of the author(s) and not necessarily those of the National Health Service, the NIHR or the Department of Health in England. A.L.S. is funded by an NHMRC Emerging Leadership Investigator Grant (#2009432). D.W. is supported by the Scott Rudolph University Endowed Professorship at Washington University in St. Louis School of Medicine. H.J. is supported by the Sydney Medical School Foundation (University of Sydney).
Funders had no role in the design, analysis or writing of this article.
A.S. owns 50% of the shares in Zuman International, which receives royalties for books AS has written and payments for presentations. A.S. additionally reports receiving presentation fees and travel reimbursements from Eli Lilly and Co, the Pharmacy Guild of Australia, Novo Nordisk, the Dietitians Association of Australia, Shoalhaven Family Medical Centres, the Pharmaceutical Society of Australia, and Metagenics, and serving on the Nestlé Health Science Optifast VLCD advisory board from 2016 to 2018. A.L.A. is Principal Investigator on two publicly funded trials where the intervention is provided by WW (formerly Weight Watchers) at no cost and is a member of the WW Scientific Advisory Board. K.S. has received in kind support as meals from ‘Lite and Easy’ for a clinical trial of weight stigma in young women in the last 5 years. A.J.H. reports receiving payment for advice given to Slimming World (UK). T.K.K. reports professional fees from Emerald Lake Safety, Gelesis, Johnson & Johnson, Novo Nordisk and Nutrisystem. F.A.H. reports personal fees from Novo Nordisk, outside the submitted work. L.A.B. reports receiving honoaria for speaking in forums organized by Novo Nordisk in relation to management of adolescent obesity and ACTION-Teens study. This study is sponsored by Novo Nordisk. It is a multi-country on-line study of attitudes towards and perceptions of obesity held by adolescents living with obesity, their parents and health care professionals. LAB is the Australian lead of the study.