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.

Online psychological interventions to reduce symptoms of depression, anxiety, and general distress in those with chronic health conditions: a systematic review and meta-analysis of randomized controlled trials.

BACKGROUND: Over the past 15 years, there has been substantial growth in web-based psychological interventions. We summarize evidence regarding the efficacy of web-based self-directed psychological interventions on depressive, anxiety and distress symptoms in people living with a chronic health condition. METHOD: We searched Medline, PsycINFO, CINAHL, EMBASE databases and Cochrane Database from 1990 to 1 May 2019. English language papers of randomized controlled trials (usual care or waitlist control) of web-based psychological interventions with a primary or secondary aim to reduce anxiety, depression or distress in adults with a chronic health condition were eligible. Results were assessed using narrative synthases and random-effects meta-analyses. RESULTS: In total 70 eligible studies across 17 health conditions [most commonly: cancer (k = 20), chronic pain (k = 9), arthritis (k = 6) and multiple sclerosis (k = 5), diabetes (k = 4), fibromyalgia (k = 4)] were identified. Interventions were based on CBT principles in 46 (66%) studies and 42 (60%) included a facilitator. When combining all chronic health conditions, web-based interventions were more efficacious than control conditions in reducing symptoms of depression g = 0.30 (95% CI 0.22-0.39), anxiety g = 0.19 (95% CI 0.12-0.27), and distress g = 0.36 (95% CI 0.23-0.49). CONCLUSION: Evidence regarding effectiveness for specific chronic health conditions was inconsistent. While self-guided online psychological interventions may help to reduce symptoms of anxiety, depression and distress in people with chronic health conditions in general, it is unclear if these interventions are effective for specific health conditions. More high-quality evidence is needed before definite conclusions can be made.

Improving body image at scale among Brazilian adolescents: study protocol for the co-creation and randomised trial evaluation of a chatbot intervention.

BACKGROUND: Body image concerns are prevalent among Brazilian adolescents and can lead to poor psychological and physical health. Yet, there is a scarcity of culturally-appropriate, evidence-based interventions that have been evaluated and made widely available. Chatbot technology (i.e., software that mimics written or spoken human speech) offers an innovative method to increase the scalability of mental health interventions for adolescents. The present protocol outlines the co-creation and evaluation of a body image chatbot for Brazilian adolescents via a partnership between academics, industry organisations and the United Nations Children's Fund (UNICEF). METHODS: A two-armed fully remote randomised controlled trial will evaluate the chatbot's effectiveness at improving body image and well-being. Adolescent girls and boys (N = 2800) aged 13-18 years recruited online will be randomly allocated (1:1) into either: 1) a body image chatbot or 2) an assessment-only control condition. Adolescents will engage with the chatbot over a 72-hour period on Facebook Messenger. Primary outcomes will assess the immediate and short-term impact of the chatbot on state- and trait-based body image, respectively. Secondary outcomes will include state- and trait-based affect, trait self-efficacy and treatment adherence. DISCUSSION: This research is the first to develop an evidence-informed body image chatbot for Brazilian adolescents, with the proposed efficacy trial aiming to provide support for accessible, scalable and cost-effective interventions that address disparities in body image prevalence and readily available resources. TRIAL REGISTRATION NUMBER: NCT04825184 , registered 30th March 2021.