Associations Between Baseline Severity and Treatment Effects in Digital Interventions for Eating Disorders: A Meta-Regression.
Claudia Liu, Cleo Anderson, Mariel Messer et al.
✦ AI-curated · Sources linked
In 30 seconds
This meta-analysis used study-level meta-regressions of 31 randomised controlled trials to examine whether baseline eating-disorder severity was associated with the effects and dropout of digital interventions compared with controls. Higher average objective binge eating was associated with larger treatment effects and lower dropout in bivariable analyses; these associations were not retained in multivariable models. Baseline global eating-disorder psychopathology was not associated with effects in bivariable analyses, but was positively associated with effects after adjustment for other study characteristics. The findings suggest that greater baseline severity does not necessarily preclude benefit.
Key findings
- Higher average objective binge eating was associated with larger treatment effects and lower dropout in bivariable meta-regressions.
- The objective binge-eating associations were not significant in multivariable meta-regressions.
- Baseline global psychopathology was unrelated to effects in bivariable analyses but positively associated with effects in the adjusted model.
Why it matters
The findings challenge the assumption that digital interventions are mainly suitable for milder eating-disorder presentations. They suggest that baseline severity may not reliably identify people less likely to benefit or remain engaged, although the differing bivariable and adjusted results warrant cautious interpretation.
What to keep in mind
The analyses were study-level meta-regressions, so associations between average trial characteristics and outcomes may not reflect individual-level relationships. Results also differed between bivariable and multivariable models.