Skip to main content
. 2023 Dec 17;13(1):5–18. doi: 10.1002/pchj.715

TABLE 5.

Meta‐analyses: statistically significant effect sizes.

Outcome Effect size SMD = g (95% CI) I 2% GRADE rating
Prevention studies: At‐risk patients
EDE‐Q
Eating disorder
Cognitive symptoms
Restraint −0.41 (−0.80, −0.02) 75 Very low
Weight concern −0.31 (−0.57, −0.06) 50 Very low
Global score −0.46 (−0.82, −0.09) 79 Very low
EDI/EDI (2)
Drive for thinness −0.45 (−0.61, −0.28) 26 Low
WCS −0.47 (−0.82, −0.11) 80 Very low
Treatment studies: (Bulimia nervosa [BN] and binge‐eating disorder [BED] patients)
EDE‐Q
Eating disorder
Cognitive symptoms
Eating concern −0.99 (−1.34, −0.63) NA Very low
Weight concern −1.11 (−1.47, −0.75) NA Very low
Shape concern −0.65 (−1.29, −0.01) 80 Very low
Global score −0.70 (−1.22, −0.17) 77 Very low
EDE‐Q
Behavioural items
Binge rate −0.66 (−1.11, −0.22) 69 Very low
EDI/EDI (2)
Bulimia −0.85 (−1.33, −0.37) NA Very low
Interoceptive awareness −0.51 (−0.97, −0.05) NA Very low
BDI/CES‐D −0.47 (−0.74, −0.20) 0 Very low
SCL‐90R −0.30 (−0.57, −0.03) 0 Very low
Relapse prevention study: (Anorexia nervosa [AN] and related eating disorder not otherwise specified [EDNOS] patients)
EDI/EDI (2)
Maturity fears −0.43 (−0.70, −0.16) NA Low
Social insecurity −0.29 (−0.56, −0.03) NA Low

Note: The bold values indicates the effect sizes (SMD with 95% CI).

Abbreviations: 95% CI, 95% confidence interval; BDI/BDI‐2, Beck depression inventory (Beck et al., 1988); CES‐D, Centre for Epidemiological Studies Depression scale (Orme et al., 1986); EDE‐Q, eating disorder examination questionnaire (Fairburn & Beglin, 1994); EDI/EDI‐2, eating disorder inventory (Garner & Olmsted, 1984); GRADE, grades of recommendation, assessment, development and evaluation ratings; I 2, extent of heterogeneity (inconsistency between studies); NA, not applicable; SMD, standardised mean difference, in this case, Hedges' (adjusted) g; SCL‐90R (GSI), Symptom checklist 90R, general symptom index score–general psychopathology/anxiety (Derogatis, 1975); WCS, weight concern scale (Killen et al., 1994).