Table 4. Summary of findings.
| Outcomes | Social support intervention(s) | Relative risk (CI) | Number of participants (studies) | Quality of evidence* | Risk of bias | Inconsistency | Imprecision | Indirectness |
|---|---|---|---|---|---|---|---|---|
| Treatment success | Social support interventions (overall) | 1.17 (1.09–1.25) | 6547, 10 studies | Very low | Serious risk of bias | Serious inconsistency, downgraded with one level due to high heterogeneity (I2 of 72.8%, P = <0.001). | No serious imprecision, adequate sample size (n = 345). | Very serious indirectness |
| Treatment success | Psycho-emotional support | 1.37 (1.08–1.73) | 400, 3 studies | Very low | Serious risk of bias, downgraded with one level for high risk of bias on two domains for one study. | Serious inconsistency, downgraded with one level due to high heterogeneity (I2 of 78%, P = 0.011) and large variation in point estimates. | No serious imprecision, adequate sample size (n = 44) | Very serious indirectness, downgraded with two levels. The studies provided different PE interventions (counseling, psychotherapy and self-help groups). One study provided the intervention to a different population (MDR-TB patients). In addition, mostly indirect comparisons are made. |
| Treatment success | Socio-economic support | 1.08 (1.03–1.13) | 4324, 4 studies | Very low | Serious risk of bias, downgraded with one level on high risk of bias on one domain in three studies. | No downgrading for inconsistency | No serious imprecision, adequate sample size (n = 748). | Very serious indirectness, downgraded with two levels. Three included studies provided food supplementation; one study provided indirect economic support. In addition, mostly indirect comparisons are made. |
| Treatment success | Combined support | 1.17 (1.12–1.22) | 1823, 3 studies | Very low | Serious risk of bias, downgraded with one level for high risk of bias on two domains in one study. | No downgrading for inconsistency | No serious imprecision, adequate sample size (n = 133). | Very serious indirectness, downgraded with two levels. All studies provided counseling and one or more PE and/or SE interventions. One study provided the intervention to a different population (MDR-TB patients). In addition, mostly indirect comparisons are made. |
| Unsuccessful treatment outcomes | Social support interventions (overall). | 0.53 (0.41–0.70) | 7301, 10 studies | Very low | Serious risk of bias | Serious inconsistency, downgraded with one level due to high heterogeneity (I2 of 80.2%, P = <0.001) and large variation in point estimates. | No serious imprecision, adequate sample size (n = 358) | Very serious indirectness |
| Unsuccessful treatment outcomes | Psycho-emotional support | 0.46 (0.22–0.96) | 1419, 4 studies | Very low | Very serious risk of bias, downgraded with two levels for high risk of bias in two studies with high risk of bias on two domains. | Serious inconsistency, downgraded with one level due to high heterogeneity (I2 of 85.5%, P = <0.001) and large variation in point estimates. | No serious imprecision, adequate sample size (n = 267). | Very serious indirectness, downgraded with two levels. The studies provided different PE interventions (counseling, psychotherapy and self-help groups). One study provided the intervention to a different population (MDR-TB patients). In addition, mostly indirect comparisons are made. |
| Unsuccessful treatment outcomes | Socio-economic support | 0.78 (0.69–0.88) | 3967, 2 studies | Very low | Serious risk of bias, downgraded by one level for high risk of bias on one domain in 2 studies. | Serious inconsistency, downgraded with one level due to large variation in point estimates (RR = 0.2 and 0.78). | No serious imprecision, adequate sample size (n = 1059). | Very serious indirectness, downgraded with two levels. The studies provided different SE interventions (food supplementation and indirect economic support). In addition, mostly indirect comparisons are made. |
| Unsuccessful treatment outcomes | Combined support | 0.42 (0.23–0.75) | 1915, 4 studies | Very low | Serious risk of bias, downgraded by one level for high risk of bias on two domains in one study and one study with high risk of bias on one domain. | Serious inconsistency, downgraded with one level due to high heterogeneity (I2 of 64.2%, P = 0.039) and large variation in point estimates. | No serious imprecision, adequate sample size (n = 127). | Very serious indirectness, downgraded with two levels. All studies provided counseling and one or more PE and/or SE interventions. One study provided the intervention to a different population (MDR-TB patients). In addition, mostly indirect comparisons are made. |
* GRADE Working Group levels of evidence.