OP17 Inequalities in the attendance and effectiveness of behavioural weight management interventions for adults in the UK: An individual participant data (IPD) meta-analysis*

Background

Behavioural weight management interventions require high personal agency and are, therefore, susceptible to widening health inequalities. Our previous systematic review found that most trials of these interventions did not consider inequalities, there was heterogeneity in how inequality characteristics were captured, and in the performed analyses. An individual participant data meta-analysis (IPD-MA) allows uniform re-analysis of all trials with available data to determine if there are inequalities in the attendance and effectiveness of behavioural weight management interventions.

Methods

We requested data from UK randomised controlled trials of behavioural weight loss interventions in primary care settings to conduct a two-stage IPD-MA. For each characteristic with sufficient available data (ethnicity, gender, socioeconomic deprivation, and age), we conducted multivariable regression analyses with weight at 12-month follow up as the primary outcome and an interaction term between inequality characteristics and trial arm (control or intervention). Each model adjusted for baseline weight, age, and gender. The interaction terms for each inequality characteristic were synthesised using a random-effects meta-analysis. We also examined if any characteristics were associated with intervention attendance and overall weight change across the trial cohort.

Results

Data from 12 trials were analysed (n=7221 participants; complete case data n=5064 participants). Overall, being in the intervention group was associated with greater weight loss of -2.48 kg (95% CI -3.51, -1.45) compared to the control group. The difference in weight loss between the intervention and control group was 0.98 kg (95% CI 0.15, 1.80) larger for male participants than for female participants and 2 kg (95% CI 0.08, 3.92) larger for participants of White ethnicity than those from an ethnic minority background. Age nor socioeconomic deprivation moderated intervention effectiveness. Five trials had attendance data available. There was some evidence that those of a non-White ethnicity had higher attendance, but the difference was small and inconclusive (2.79% higher attendance in those who are non-White, 95% CI -2.71%, 8.29%). Across all participants, at 12-month follow up we found that being male was associated with a higher weight of 0.62 kg (95% CI 0.10, 1.15).

Conclusion

Behavioural weight management interventions are more effective in men, despite typically greater uptake of these interventions by women and lower weight loss in men across the cohort. There was no evidence of inequalities by socioeconomic deprivation, but behavioural weight management interventions may contribute to inequalities in obesity by ethnicity. Future research should investigate the mechanisms behind this to better support people with obesity from minority ethnic groups.