Individual participant data meta-analysis of eating behaviour traits as effect modifiers in acceptance and commitment therapy-based weight management interventions

Abstract

Background: Obesity care may benefit from precision approaches, matching patients to treatment types based on their individual characteristics, including eating behaviour traits (EBTs) like emotional eating, uncontrolled eating, external eating, internal disinhibition and restraint. Initial evidence suggests that Acceptance and Commitment Therapy (ACT)-based interventions might address dysregulated EBTs more effectively than standard behavioural treatments. However, it is unclear if ACT is more effective for certain EBT levels. Methods and analysis: This pre-registered (CRD42022359691) one-stage Individual Participant Data (IPD) meta-analysis explored the moderating effects of baseline EBTs on weight outcomes in trials of ACT-based interventions for adults with a BMI ≥ 25 kg/m2. Unlike traditional meta-analyses, IPD meta-analyses re-analyse existing data to answer novel research questions. We identified 16 eligible trials through a systematic search of eight databases until June 20, 2022. We obtained, checked, and harmonised data from 15 trials (N = 2535). We used mixed regression models to investigate both continuous and categorical interaction effects. Results: We found no evidence of interactions between ACT vs. control and baseline EBTs as continuous variables on percentage weight change. However, we found evidence to suggest an added difference in weight change of −4.47% (95%CI −1.15, −7.73) from baseline to 12-months after intervention end in participants with medium levels of internal disinhibition compared to those with high levels. Sensitivity analyses similarly indicated a greater intervention benefit for participants with medium, rather than high, emotional eating levels (in trials that reduced experiential avoidance and in trials using the three-factor eating questionnaire) and internal disinhibition (in analyses of participants with at least 60% attendance). Given the exploratory nature of analyses, results should be interpreted with caution. Conclusion: Findings suggest potential non-linear interaction effects of ACT with internal disinhibition but require replication in confirmatory trials. These results may help guide further research on precision approaches based on EBTs.