Table 5.
. Effects of intervention compared with control conditions and subgroup meta-analyses
| Study or subgroup | Hedges’ g* (95% CI) | P |
|---|---|---|
| Effect of intervention vs control conditions | ||
| All included trials (58–63,71–75)† | 1.01 (0.48 to 1.54) | <.001 |
| Included trials (58–62,71–75)‡; outlier trial (63) removed§ | 0.43 (0.30 to 0.56) | <.001 |
| Included trials (58–62,71–75) using lowest reported effect size | 0.36 (0.22 to 0.49) | <.001 |
| Included trials (58–62,71–75) using highest reported effect size | 0.50 (0.31 to 0.70) | <.001 |
| Effect of intervention compared with control conditions postintervention and at follow-up assessments | ||
| Included trials with 6–8 mos postrandomization follow-up assessments (59,72,74) | ||
| Postintervention | 0.54 (0.27 to 0.82) | <.001 |
| Follow-up | 0.56 (0.34 to 0.78) | <.001 |
| Included trials with 12–18 mos postrandomization follow-up assessments (73–75) | ||
| Postintervention | 0.34 (0.16 to 0.53) | <.001 |
| Follow-up | 0.37 (0.18 to 0.55) | <.001 |
| Included trial with 24 mos postrandomization follow-up assessments (75) | ||
| Postintervention | 0.26 (0.02 to 0.51) | .040 |
| Follow-up | 0.19 (-0.08 to 0.46) | .173 |
| Subgroup analyses | ||
| Effect of CBT vs PST vs pharmacologic interventions | ||
| CBT(59,61) | 0.83 (0.48 to 1.18) | <.001 |
| PST (72–75) | 0.33 (0.15 to 0.50) | <.001 |
| Pharmacologic (58,60,62,71) | 0.44 (0.19 to 0.68) | <.001 |
| Effect of intervention in psychotherapeutic trials meeting PEDro criterion: Whether at least 85% of study participants had postintervention data available for analysis | ||
| Criterion met (59,74) | 0.58 (0.34 to 0.81) | <.001 |
| Criterion not met (61,72,73) | 0.31 (0.10 to 0.52) | .004 |
* Hedges’ g was calculated on the basis of the standardized mean difference effect size which uses the pooled within-groups SD. Subgroup analyses used a mixed effects model, which pools studies within subgroups with a random effects model, but tested for statistically significant between-groups differences with a fixed effects model. All P values are two-tailed. CBT = cognitive behavioral therapy; PST = problem-solving therapy.
† Heterogeneity: I 2 = 93.4%, 95% CI = 90.0 to 95.0; Q = 181.1, degrees of freedom (df) = 12, P < .001.
‡ Heterogeneity: I 2 = 0%, 95% CI = 0.0 to 60.0; Q = 8.67, df = 10, P = .56.
§ Comparisons from Nezu et al. (63) included examination of effects of PST (Hedges’ g = 3.57, 95% CI = 2.90 to 4.23, P < .001) and PST with involvement of a significant other (Hedges’ g = 4.32, 95% CI = 3.55 to 5.08, P < .001).