A proof-of-concept study applying machine learning methods to putative risk factors for eating disorders: results from the multi-centre European project on healthy eating.

BACKGROUND: Despite a wide range of proposed risk factors and theoretical models, prediction of eating disorder (ED) onset remains poor. This study undertook the first comparison of two machine learning (ML) approaches [penalised logistic regression (LASSO), and prediction rule ensembles (PREs)] to conventional logistic regression (LR) models to enhance prediction of ED onset and differential ED diagnoses from a range of putative risk factors. METHOD: Data were part of a European Project and comprised 1402 participants, 642 ED patients [52% with anorexia nervosa (AN) and 40% with bulimia nervosa (BN)] and 760 controls. The Cross-Cultural Risk Factor Questionnaire, which assesses retrospectively a range of sociocultural and psychological ED risk factors occurring before the age of 12 years (46 predictors in total), was used. RESULTS: All three statistical approaches had satisfactory model accuracy, with an average area under the curve (AUC) of 86% for predicting ED onset and 70% for predicting AN v. BN. Predictive performance was greatest for the two regression methods (LR and LASSO), although the PRE technique relied on fewer predictors with comparable accuracy. The individual risk factors differed depending on the outcome classification (EDs v. non-EDs and AN v. BN). CONCLUSIONS: Even though the conventional LR performed comparably to the ML approaches in terms of predictive accuracy, the ML methods produced more parsimonious predictive models. ML approaches offer a viable way to modify screening practices for ED risk that balance accuracy against participant burden.

Comparative efficacy and acceptability of psychological interventions for the treatment of adult outpatients with anorexia nervosa: a systematic review and network meta-analysis.

BACKGROUND: No consistent first-option psychological interventions for adult outpatients with anorexia nervosa emerges from guidelines. We aimed to compare stand-alone psychological interventions for adult outpatients with anorexia nervosa with a specific focus on body-mass index, eating disorder symptoms, and all-cause dropout rate. METHODS: In this systematic review and network meta-analysis, we assessed randomised controlled trials about stand-alone pharmacological or non-pharmacological treatments of adult outpatients with anorexia nervosa, defined according to standardised criteria, with data for at least two timepoints relating to either body-mass index or global eating disorder psychopathology. We searched Cochrane CENTRAL, CINAHL, MEDLINE, and PsychINFO for published and unpublished literature from inception until March 20, 2020. The primary outcomes were the change in body mass index and clinical symptoms, and the secondary outcome was all-cause dropout rate, which were all assessed for treatment as usual, cognitive behavioural therapy (CBT), Maudsley anorexia treatment for adults, family-based treatment, psychodynamic-oriented psychotherapies, a form of CBT targeting compulsive exercise, and cognitive remediation therapy followed by CBT. Global and local inconsistencies for the network meta-analysis were measured, and CINeMA was used to assess the confidence in evidence for primary outcomes. The protocol is registered in PROSPERO (CRD42017064429). FINDINGS: Of 14 003 studies assessed for their title and abstract, 16 (0·1%) randomised controlled trials for psychological treatments were included in the systematic review, of which 13 (0·1%) contributed to the network meta-analysis, with 1047 patients in total (of whom 1020 [97·4%] were female). None of the interventions outperformed treatment as usual in our primary outcomes, but the all-cause dropout rate was lower for CBT than for psychodynamic-oriented psychotherapies (OR 0·54, 95% CI 0·31-0·93). Heterogeneity or inconsistency emerged only for a few comparisons. Confidence in the evidence was low to very low. INTERPRETATION: Compared with treatment as usual, specific psychological treatments for adult outpatients with anorexia nervosa can be associated with modest improvements in terms of clinical course and quality of life, but no reliable evidence supports clear superiority or inferiority of the specific treatments that are recommended by clinical guidelines internationally. Our analysis is based on the best data from existing clinical studies, but these findings should not be seen as definitive or universally applicable. There is an urgent need to fund new research to develop and improve therapies for adults with anorexia nervosa. Meanwhile, to better understand the effects of available treatments, participant-level data should be made freely accessible to researchers to eventually identify whether specific subgroups of patients are more likely to respond to specific treatments. FUNDING: Flinders University, National Institute for Health Research Oxford Health Biomedical Research Centre.

Multidisciplinary healthcare teams' challenges and strategies in supporting people with type 1 diabetes to recover from disordered eating.

AIM: To examine the challenges healthcare teams face when treating people with type 1 diabetes and disordered eating and the strategies these teams have developed to facilitate effective treatment. METHODS: Four semi-structured focus groups were conducted including two tertiary diabetes specialist teams and three tertiary eating disorders specialist teams between July and December 2018. Thematic analysis of the transcripts followed a six-phase process. RESULTS: Twenty-nine experienced healthcare professionals (16 diabetes and 13 eating disorder specialists, 16±12 years' professional experience) were interviewed. The challenges identified in treating people with type 1 diabetes and disordered eating included subthemes the 'challenges specific to the healthcare professional' (feeling not competent enough and perceived emotional burden), 'challenges pertaining to patient factors' (e.g. difficulties with engaging in therapy) and 'challenges created by the healthcare system' (time pressure and staff shortage). Healthcare professionals expressed the need for a consensus on diagnosis and the definition of disordered eating in type 1 diabetes, as well as the need for training and educational resources specific to type 1 diabetes and disordered eating. Healthcare professionals gave practical examples of strategies of communication for better patient engagement and felt that multidisciplinary working in joint clinics with the other specialty were facilitators for recovery from disordered eating. CONCLUSIONS: Healthcare professionals require multidisciplinary team support when treating people with type 1 diabetes and to improve their own competencies. The development of effective screening and assessment tools, educational resources and training for healthcare professionals, and developing multidisciplinary treatment pathways will be key to improving outcomes for their service users with type 1 diabetes and disordered eating.

25 Years of psychological research investigating disordered eating in people with diabetes: what have we learnt?

Disordered eating is a serious and under-recognized problem in people with diabetes. This narrative review summarizes the research contributions made by psychological science over the past 25 years to the study of disordered eating in people with type 1 or type 2 diabetes, and identifies gaps and future directions relevant to both healthcare professionals and researchers. Key focus areas of psychological research investigating disordered eating in people with diabetes have been: (1) defining and classifying types of disordered eating; (2) identifying demographic, diabetes-specific and psychosocial correlates of disordered eating, and developing theoretical models of disordered eating in people with type 1 diabetes; (3) identifying the physical and psychosocial consequences of disordered eating; and (4) developing screening measures to identify disordered eating in people with type 1 diabetes. Psychological science has made significant contributions over the past 25 years to our understanding of the nature of this problem and the multiple factors which may interrelate with disordered eating in people with diabetes. Key areas for further attention include: (1) a better definition of disordered eating subtypes in people with type 1 diabetes; (2) characterizing disordered eating in people with type 2 diabetes; and (3) developing multidisciplinary, evidence-based prevention and treatment interventions for comorbid disordered eating and diabetes.

Correction: Investigation of common, low-frequency and rare genome-wide variation in anorexia nervosa.

The fortieth author's name was listed incorrectly. The correct presentation is A Keski-Rahkonen.

'Diabulima' through the lens of social media: a qualitative review and analysis of online blogs by people with Type 1 diabetes mellitus and eating disorders.

AIMS: To perform a qualitative review of online blogs authored by people self-identifying as having Type 1 diabetes and an eating disorder or 'diabulimia', a term used by people with Type 1 diabetes to describe an eating disorder that is characterized by deliberate restriction of insulin to control weight. METHODS: We conducted a structured qualitative review of online blogs published between 2012 and 2017 authored by people who report having Type 1 diabetes and an eating disorder or diabulimia. The subsequent thematic analysis followed a six-phase process and was conducted by two independent researchers. RESULTS: From 147 000 search results, 11 blogs (304 posts) matched criteria for further analyses. Three key themes and 18 subthemes emerged: 1) different aspects of bloggers' relationship with insulin, including motives for omitting insulin, secrecy of insulin omission and perception of control; 2) bloggers' experiences of diabetes complications, and diabetes ketoacidosis in particular, as well as their worries about future complications; 3) strategies for recovery and triggers for relapse, which involved diabetes self-management and setting up a support system. CONCLUSIONS: Qualitative analyses of blogs authored by people with Type 1 diabetes and an eating disorder or diabulimia have identified high levels of diabetes distress and provided insight into different motives for insulin omission and strategies for recovery. Considering the limited evidence for effective interventions, these findings may help the development of complex interventions to improve biomedical and psychological outcomes in this group.

Investigation of common, low-frequency and rare genome-wide variation in anorexia nervosa.

Anorexia nervosa (AN) is a complex neuropsychiatric disorder presenting with dangerously low body weight, and a deep and persistent fear of gaining weight. To date, only one genome-wide significant locus associated with AN has been identified. We performed an exome-chip based genome-wide association studies (GWAS) in 2158 cases from nine populations of European origin and 15 485 ancestrally matched controls. Unlike previous studies, this GWAS also probed association in low-frequency and rare variants. Sixteen independent variants were taken forward for in silico and de novo replication (11 common and 5 rare). No findings reached genome-wide significance. Two notable common variants were identified: rs10791286, an intronic variant in OPCML (P=9.89 × 10-6), and rs7700147, an intergenic variant (P=2.93 × 10-5). No low-frequency variant associations were identified at genome-wide significance, although the study was well-powered to detect low-frequency variants with large effect sizes, suggesting that there may be no AN loci in this genomic search space with large effect sizes.

Eating disorders in people with Type 1 diabetes: experiential perspectives of both clients and healthcare professionals.

AIMS: To explore the experiential perspective of people with Type 1 diabetes mellitus and eating disorders and that of the healthcare professionals treating them, and to understand the experience of both sides to inform future development of healthcare services. METHODS: Participants were recruited from Diabetics with Eating Disorders (a national UK charity), and through professional networks. Nine partially/fully recovered individuals with Type 1 diabetes and eating disorders and eight healthcare professionals participated in semi-structured interviews carried out by medically trained researchers. Data were transcribed and coded using a six-stage framework of thematic analysis. RESULTS: Four superordinate themes and several subordinate themes emerged from the Type 1 diabetes and eating disorders dataset: (1) perceptions surrounding service provision; (2) reflections on the recovery process; (3) the experiential perspective of living with Type 1 diabetes and an eating disorder; and (4) support mechanisms. Healthcare professional data elicited three superordinate themes and several subordinate themes: (1) service provision; (2) personal insight and reflection of professional role; and (3) challenges of working with dual diagnoses. CONCLUSION: People with Type 1 diabetes and eating disorders and their healthcare professionals provided insight into healthcare services from the patient and care delivery perspectives. There was general agreement from both groups that a multidisciplinary, collaborative (family inclusive), clinical approach to treatment is important, as well as adequate training opportunities for service providers. These findings may help to inform development strategies for multidisciplinary care approaches to Type 1 diabetes complicated by eating disorders.

Systematic review and meta-analysis of the efficacy of interventions for people with Type 1 diabetes mellitus and disordered eating.

AIM: To examine the types of interventions currently available for people with Type 1 diabetes mellitus and their effectiveness. BACKGROUND: The prevalence of disordered eating in people with Type 1 diabetes mellitus is twice that in their counterparts without diabetes, and is associated with worse biomedical outcomes and greater mortality. METHODS: Medline, Embase, PsycINFO, the Cochrane Library, PubMed and OpenGrey databases were searched up to August 2016 to identify studies on interventions in people with Type 1 diabetes-associated disordered eating. For the systematic review, intervention components were identified and their effectiveness was examined. For the meta-analysis, the pooled effect sizes of glycaemic control (HbA1c ) between pre- and post-treatment in treatment and comparison groups were calculated using a random effects model. RESULTS: Of 91 abstracts reviewed, six studies met the inclusion criteria, of which three had appropriate data for the meta-analysis (n = 118). The pooled effect size was -0.21 95% CI (-0.58 to 0.16; where negative values represent an improvement in HbA1c levels), indicating no statistically significant improvement in the treatment group compared with comparison group. Inpatient therapy appeared to be the most effective treatment, and this had multiple components including cognitive behavioural therapy, psychoeducation and family therapy. CONCLUSION: Limited or no improvement in glycaemic control and disordered eating symptoms was observed in people with Type 1 diabetes-associated disordered eating who were receiving currently available interventions. The present review suggests that developing an intensive intervention with a joint focus on both disordered eating and diabetes management is needed for this complex patient group.

A randomised controlled trial of three psychological treatments for anorexia nervosa.

BACKGROUND: There is a lack of evidence pointing to the efficacy of any specific psychotherapy for adults with anorexia nervosa (AN). The aim of this study was to compare three psychological treatments for AN: Specialist Supportive Clinical Management, Maudsley Model Anorexia Nervosa Treatment for Adults and Enhanced Cognitive Behavioural Therapy. METHOD: A multi-centre randomised controlled trial was conducted with outcomes assessed at pre-, mid- and post-treatment, and 6- and 12-month follow-up by researchers blind to treatment allocation. All analyses were intention-to-treat. One hundred and twenty individuals meeting diagnostic criteria for AN were recruited from outpatient treatment settings in three Australian cities and offered 25-40 sessions over a 10-month period. Primary outcomes were body mass index (BMI) and eating disorder psychopathology. Secondary outcomes included depression, anxiety, stress and psychosocial impairment. RESULTS: Treatment was completed by 60% of participants and 52.5% of the total sample completed 12-month follow-up. Completion rates did not differ between treatments. There were no significant differences between treatments on continuous outcomes; all resulted in clinically significant improvements in BMI, eating disorder psychopathology, general psychopathology and psychosocial impairment that were maintained over follow-up. There were no significant differences between treatments with regard to the achievement of a healthy weight (mean = 50%) or remission (mean = 28.3%) at 12-month follow-up. CONCLUSION: The findings add to the evidence base for these three psychological treatments for adults with AN, but the results underscore the need for continued efforts to improve outpatient treatments for this disorder. Trial Registration Australian New Zealand Clinical Trials Registry (ACTRN 12611000725965) http://www.anzctr.org.au/.