Raw and processed ecological momentary assessment data on emotion-eating behavior associations in eating disorder patients for systematic review and meta-analysis
Description
The core hypothesis is that emotional states and disordered eating behaviors have a bidirectional dynamic association in clinically diagnosed eating disorder patients: elevated negative affect predicts abnormal eating behaviors (e.g., binge eating, purging), which in turn influence subsequent emotional states (e.g., immediate negative affect relief via negative reinforcement), with this association moderated by diagnostic subtype, eating behavior type, emotional trajectory, and EMA methodological parameters. Data were aggregated from 36 observational studies (2007–2026) across 10 databases and hand-searched references (retrieved by January 19, 2026), including 2,550 participants (93.5% female, mean age ~28 years) with clinical eating disorder diagnoses, from clinical (n=11), community (n=5), and mixed (n=20) samples. EMA protocols involved 2–231 days of monitoring (mostly 7–14 days), 3–8 daily signals (usually 6), and ~85% compliance, measuring emotional states (negative/positive affect, dynamic indicators), disordered eating behaviors, and quantitative indicators (β coefficients, subgroup results, etc.). Key findings include a significant positive emotion-eating association (β=0.24, 95% CI: 0.22–0.27, p<0.001, I²=99.5%), bidirectional links (negative affect predicts eating behaviors [β=0.24]; eating behaviors reduce negative affect [β=-0.86]), and significant moderators (stronger associations in clinical samples, largest effect for loss of control eating, stronger links with negative affect trajectories, optimal EMA parameters of 7-day monitoring and 3 daily signals); Egger’s test indicated publication bias (p=0.001), with corrected β=0.044. The dataset confirms the emotion regulation model’s core proposition, though true effect size may be smaller due to heterogeneity and bias, supporting individualized interventions, and it can be used to replicate meta-analyses, explore moderating/mediating mechanisms, and develop standardized EMA protocols or prediction models.
Files
Institutions
- Zhejiang Chinese Medical UniversityZhejiang, Hangzhou