Table 2 Results of the assessment of risk of bias in included non-randomized studies by using the ROBINS-I assessment tool
Study ID | 1. Bias caused by confounding | 2. Bias caused by selection of participants | 3. Bias caused by classification of interventions | 4. Bias caused by deviations from intended interventions | 5. Attrition bias caused by missing data | 6. Detection bias caused by measurement of outcomes | 7. Reporting bias caused by selection of the reported results | Overall judgement |
|---|---|---|---|---|---|---|---|---|
Assah et al. [43] | Serious | Serious | Serious | Low | No information | Serious | Serious | Serious |
Baumann et al. [44] | Serious | Serious | Serious | Low | Low (information on reasons for missing data provided) | Serious | Serious | Serious |
Eggermont [45] | Serious | Serious | Serious | Low | Low (information on reasons for missing data provided) | Serious | Moderate | Serious |
Rotheram-Borus et al. [46] | Serious | Serious | Serious | Low | Serious | Serious | Serious | Serious |
Shen [47] | Serious | Serious | Serious | Low | No information | Serious | Low | Serious |
Less et al. [48] | Serious | Serious | Serious | Low | Low (information on reasons for missing data provided) | Serious | Low | Serious |
Micikas et al. [49] | Serious | Serious | Serious | Low | No information | Serious | Serious | Serious |