Table 4.
Effect sizes, heterogeneity and meta-regression.
| Effects/heterogeneity | n | g | 95% CI | pa | NNT | I2 |
|---|---|---|---|---|---|---|
| TOTAL | 5 | 0.41 | 0.05–0.77 | 0.025 | 4.27 | 55 |
| COMPARISON | ||||||
| Active control | 3 | 0.33 | −0.18–0.84 | 0.202 | 5.26 | 27 |
| Passive control | 4 | 0.48 | −0.39–1.35 | 0.279 | 3.6 | 79* |
| MEASURE† | ||||||
| Raw | 3 | 0.03 | −0.46–0.53 | 0.896 | 59.09 | 23 |
| Standard | 2 | 0.81 | 0.30–1.33 | 0.002 | 2.25 | 0 |
| Meta-regression | n | Beta | 95% CI | pb | R2 | pc |
| % female | 5 | −0.02 | −0.10–0.06 | 0.594 | 0.00 | 0.039 |
| Age | 5 | −0.03 | −0.06–0.01 | 0.039 | 0.80 | 0.284 |
| Number of sessions | 5 | 0.33 | 0.09–0.56 | 0.007 | 1.00 | 0.663 |
| Hours of programme | 5 | 0.06 | 0.02–0.09 | 0.005 | 1.00 | 0.834 |
n, number of included studies; g, Hedge's g effect size; 95% CI, confidence interval; pa, p-value associated with g; NNT, number needed to treat; I2, proportion of real observed dispersion.
*
p-value associated to heterogeneity test < 0.01. Beta, coefficient of meta-regression; pb, p-value associated with the Beta coefficient; R2, proportion of total between-study variance explained by model; pc, p-value associated with the goodness of fit of the model assuming that the unexplained variance is null.
†
Significant subgroup contrast at p < 0.05.