Table 2.
Effect sizes and publication bias across SITB outcomes.
| Binary/categorical | n | RR [95% CI] | p | I2 | Fail-safe N | Begg and Mazumdar rank correlation | Egger's test of intercept | Duval and Tweedie's Trim and Fill | ||
|---|---|---|---|---|---|---|---|---|---|---|
| Classic | Orwin's | Missing effect sizes | Adjusted RR | |||||||
| Overall | 362 | 1.06 [0.99, 1.14] | 0.09 | 7.19% | 217 | 0 | τ = − 0.01, p = 0.76 | z = 1.87, p = 0.06 | 0 | 1.06 [0.99, 1.14] |
| Suicide ideation | 126 | 1.03 [0.92, 1.14] | 0.65 | 2.70% | 0 | 0 | τ = 0.04, p = 0.46 | z = 0.79, p = 0.43 | 0 | 1.03 [0.92, 1.14] |
| Suicide attempt | 53 | 1.21 [0.95, 1.55] | 0.13 | 11.77% | 0 | 0 | τ = − 0.09, p = 0.35 | z = 0.53 p = 0.60 | 0 | 1.21 [0.95, 1.55] |
| Suicide death | 6 | 0.77 [0.47, 1.26] | 0.30 | 0.00% | 0 | 0 | τ = 0.60, p = 0.14 | z = 1.69, p = 0.09 | 0 | 0.77 [0.47, 1.26] |
| NSSI | 28 | 1.18 [0.89, 1.57] | 0.25 | 0.00% | 0 | 31 | τ = − 0.23, p = 0.09 | z = − 1.25, p = 0.21 | 4 | 1.27 [0.96, 1.68] |
| Self-harm | 30 | 0.99 [0.80, 1.21] | 0.90 | 23.03% | 0 | 0 | τ = − 0.13, p = 0.34 | z = 0.32, p = 0.75 | 0 | 0.99 [0.80, 1.21] |
| Hospitalizations | 8 | 1.11 [0.89, 1.39] | 0.33 | 0.00% | 0 | 18 | τ = 0.00, p = 1.00 | z = − 0.85, p = 0.40 | 2 | 1.14 [0.92, 1.42] |
| Other/combined SITBs | 111 | 1.16 [0.99, 1.36] | 0.49 | 11.46% | 66 | 0 | τ = − 0.03, p = 0.60 | z = 2.18, p = 0.03 | 0 | 1.16 [0.99, 1.36] |
| Continuous | n | Hedges’ g [95% CI] | p | I2 | Fail-safe N | Begg and Mazumdar rank correlation | Egger's test of intercept | Duval and Tweedie's Trim and Fill | ||
|---|---|---|---|---|---|---|---|---|---|---|
| Classic | Orwin's | Missing effect sizes | Adjusted RR | |||||||
| Overall | 50 | − 0.04 [− 0.12, 0.05] | 0.37 | 64.60% | 5 | 34 | τ = − 0.05, p = 0.58 | z = − 1.37, p = 0.17 | 0 | − 0.04 [− 0.12, 0.05] |
| Suicide ideation | 33 | − 0.03 [− 0.12, 0.06] | 0.53 | 55.92% | 0 | 0 | τ = 0.003, p = 0.99 | z = 0.32, p = 0.75 | 0 | − 0.03 [− 0.12, 0.06] |
| Suicide attempt | 3 | – | – | – | – | – | – | – | – | – |
| Suicide death | 1 | – | – | – | – | – | – | – | – | – |
| NSSI | 1 | – | – | – | – | – | – | – | – | – |
| Self-harm | 7 | 0.12 [− 0.07, 0.32] | 0.22 | 0.00% | 0 | 0 | τ = 0.05, p = 1.00 | z = 0.24, p = 0.81 | 0 | 0.12 [− 0.07, 0.32] |
| Hospitalizations | 1 | – | – | – | – | – | – | – | – | – |
| Other/combined SITBs | 4 | – | – | – | – | – | – | – | – | – |
n number of effect sizes, RR weighted mean risk ratio, CI confidence interval, dashes indicate unavailable information; I2 indicates the percentage of variances due to heterogeneity between studies. Classic Fail-safe N and Orwin’s Fail-safe N represent the number of studies needed to nullify the observed effects statistically and clinically, respectively; Begg and Mazumdar Rank Correlation Test computes the rank order correlation between effect estimates and sampling variance; Egger's Test of the Intercept uses precision (i.e., the inverse of the standard error) to predict the standardized effect (i.e., effect size divided by the standard error); Duval & Tweedie's Trim & Fill estimates effect sizes after accounting for publication bias. Missing cases are the number of cases estimated to be missing below the mean. Boldface indicates significance at p < 0.05.