Changes in disordered eating behaviors over 10 or more years: A meta-analysis

Abstract

Objective: The present meta-analysis evaluated changes in individuals' risk of engaging in distinct disordered eating behaviors (DEBs) in the long-term.

Method: Longitudinal studies assessing changes in DEBs via ≥2 assessments with a time lag of ≥10 years were included. Risk ratios were calculated for baseline to 10-14.9-year (M = 11.04) follow-up and baseline to ≥15-year (M = 18.62) follow-up changes in the use of binge eating, multiple purging, self-induced vomiting, laxatives, diuretics, diet pills, compensatory exercising, fasting/dieting, and multiple DEBs; Cohen's d was used for continuous binge-eating plus purging variable changes. Subgroup and meta-regression analyses tested whether eating disorder (ED) clinical sample versus nonclinical sample status, female versus male sex/gender, higher versus lower study bias, and baseline mean age and body mass index influenced overall effect magnitude for analyses with sufficient data.

Results: Seventeen studies (26 [sub]samples) were included. Overall, individuals' risk of engaging in various restrictive eating and other compensatory behaviors decreased over time and the magnitudes of risk reductions for the use of certain compensatory DEBs were larger over longer follow-up durations. Specifically, for significant DEB change models, risk reductions spanned from 20.0-39.8% for 10-year follow-up and 24.7-74.8% for ≥15-year follow-up. However, nuances were found in the nature of these DEB changes as a function of DEB type, follow-up length, ED versus nonclinical sample composition, and baseline mean age.

Conclusions: These findings provide important information that can help identify treatment priorities and suggest that targeted and tailored preventative ED treatments warrant consistent implementation at the community-level, particularly for youth.

Keywords: anorexia nervosa; binge eating; bulimia nervosa; eating disorders; meta-analysis.