Table 2.
Pooled estimates for all comparisons, modality, mode of delivery and format across outcomes
| Outcome | K | Effect size [95% CI] | p | I 2* | [95% PI] | NNTa [95% CI] | Effect sizeadj [95% CI] | Kimp |
| Gambling severity | Hedges' g | Hedges' g adj | ||||||
| All comparisons | 28 | −0.71 [−1.03; −0.39] | <0.0001 | 89.6 | [−2.34; 0.92] | 2.6 [1.9, 4.6] | – | – |
| PFI control conditions excluded | 28 | −0.72 [−1.03; −0.40] | <0.0001 | 90.3 | [−2.32; 0.89] | 2.6 [1.9; 4.5] | – | – |
| Outliers excludedb | 26 | −0.59 [−0.87; −0.30] | 0.0001 | 86.8 | [−1.97; 0.80] | 3.1 [2.2; 5.9] | – | – |
| Modality | ||||||||
| CBT | 18 | −0.85 [−1.36; −0.34] | 0.0012 | 94.5 | [−3.11; 1.41] | 2.2 [1.5; 5.3] | – | – |
| MI | 3 | 0.08 [−0.15; 0.32] | 0.4850 | 0.0 | [−1.44; 1.61] | – | – | – |
| CBT/MI | 7 | −0.53 [−0.94; −0.12] | 0.0108 | 78.2 | [−1.88, 0.82] | 3.4 [2.0; 14.5] | – | – |
| Other interventions | 5 | −0.78 [−1.03; −0.52] | <0.0001 | 0.0 | [−1.18; −0.37] | 2.4 [1.9; 3.5] | – | – |
| Mode of delivery and format | ||||||||
| Face-to-face | 16 | −1.03 [−1.54; −0.53] | 0.0001 | 87.8 | [−3.08; 1.01] | 1.9 [1.4; 3.4] | – | – |
| Individually | 10 | −0.89 [−1.53; −0.25] | 0.0062 | 91.1 | [−3.25; 1.46] | 2.1 [1.4; 7.0] | – | – |
| In groups | 6 | −1.33 [−2.18; −0.47] | 0.0023 | 74.4 | [−4.01; 1.35] | 1.5 [1.1; 3.8] | −0.86 [−1.74; 0.02] | 2 |
| Remote | 13 | −0.36 [−0.65; −0.06] | 0.0184 | 84.7 | [−1.47; 0.76] | 5.0 [2.8; 29.6] | – | – |
| Therapist guided | 6 | −0.42 [−0.97; 0.14] | 0.1421 | 83.3 | [−2.33; 1.49] | – | – | – |
| Unguided | 8 | −0.22 [−0.55; 0.12] | 0.2057 | 85.4 | [−1.37: 0.93] | – | – | – |
| Internet program | 6 | −0.42 [−0.92; 0.09] | 0.1080 | 91.2 | [−2.23; 1.40] | – | – | – |
| Remission | Odds Ratio | Odds Ratio adj | ||||||
| All comparisons | 12 | 0.47 [0.25; 0.88] | 0.0178 | 80.3 | [0.05; 4.27] | 5.3 [3.0; 32.4] | – | – |
| Outliers excludedb | 11 | 0.56 [0.32; 0.97] | 0.0376 | 74.1 | [0.09; 3.47] | 7.1 [3.6; 129.5] | – | – |
| Modality | ||||||||
| CBT | 9 | 0.45 [0.19; 1.04] | 0.0616 | 81.6 | [0.03; 6.97] | – | – | – |
| Mode of delivery and format | ||||||||
| Face-to-face | 7 | 0.24 [0.08; 0.73] | 0.0116 | 78.9 | [0.01; 8.63] | 2.9 [1.9; 13.1] | – | – |
| Individually | 3 | 0.54 [0.27; 1.09] | 0.0845 | 34.3 | [0.00; 346.26] | – | – | – |
| In groups | 4 | 0.10 [0.02; 0.61] | 0.0128 | 73.0 | [0.00; 234.95] | 2.0 [1.6; 8.5] | – | – |
| Remote | 6 | 0.87 [0.63; 1.20] | 0.3850 | 17.3 | [0.45; 1.67] | – | 0.94 [0.65; 1.37] | 1 |
| Therapist guided | 3 | 0.82 [0.38; 1.76] | 0.6019 | 64.0 | [0.00; 3,838.27] | – | – | – |
| Unguided | 3 | 0.80 [0.52; 1.22] | 0.2906 | 0.0 | [0.05; 12.38] | – | – | – |
| Depressive symptoms | Hedges' g | Hedges' g adj | ||||||
| All comparisons | 8 | −0.46 [−0.77; −0.15] | 0.0034 | 55.4 | [−1.34; 0.42] | 3.9 [2.4; 11.7] | – | – |
| Modality | ||||||||
| CBT | 6 | −0.40 [−0.74; −0.06] | 0.0200 | 48.4 | [−1.33; 0.53] | 4.5 [2.5; 28.1] | – | – |
| Mode of delivery | ||||||||
| Face-to-face | 4 | −0.63 [−0.97; −0.30] | 0.0002 | 0.0 | [−1.37; 0.10] | 2.9 [2.0; 6.0] | – | – |
| Remote | 4 | −0.32 [−0.77; 0.12] | 0.1509 | 68.6 | [−2.2; 1.55] | – | – | – |
| Anxiety | Hedges' g | Hedges' g adj | ||||||
| All comparisons | 7 | −0.56 [−0.78; −0.35] | <0.0000 | 0.0 | [−0.85; −0.28] | 3.2 [2.4; 5.2] | −0.54 [−0.75; −0.33] | 1 |
| Modality | ||||||||
| CBT | 5 | −0.60 [−0.86; −0.35] | <0.0000 | 0.0 | [−1.02; −0.19] | 3.0 [2.2; 5.1] | – | – |
| Mode of delivery | ||||||||
| Face-to-face | 4 | −0.68 [−1.02; −0.35] | 0.0001 | 0.0 | [−1.42; 0.06] | 2.7 [1.9; 5.2] | – | – |
| Remote | 3 | −0.48 [−0.77; −0.20] | 0.0010 | 0.0 | [−2.33; 1.37] | 3.8 [2.4; 9.1] | −0.34 [−0.56; −0.11] | 2 |
Bold p-values are statistically significant (<0.05), abbreviations: K, number of comparisons; CBT, cognitive behavioral therapy; MI, motivational intervention; PI, prediction interval; Effect sizeadj, Effect size adjusted for publication bias, Kimp, number of studies imputed to adjust for publication bias using trim and fill; NNT, number-needed-to-treat, *) for estimates generated on only few comparisons, heterogeneity and random variance is not possible to outline, leading to large confidence intervals on tau2. Thus, I2 becomes biased towards homogeneity (I2 = 0) (Higgins & Thompson, 2002), a) NNT was calculated only for estimates of effect that were statistically significant, b) in sensitivity analysis studies identified as 2 standard deviations above/below the pooled estimate were excluded.