Scoring domain | Criteria | Studies receiving 1 full point (n) | Studies receiving 0.5 points (n) | Studies receiving 0 points (n) | Total Number of Studies |
---|---|---|---|---|---|
Data harmonization | measurement & definition of variables | 11 (37.9%) | 10 (34.5%) | 8 (27.6%) | N = 29 |
differences in measurement & definitions | 10 (34.5%) | 11 (37.9%) | 8 (27.6%) | N = 29 | |
harmonization/standardization efforts | 20 (68.9%) | 8 (27.6%) | 1 (3.4%) | N = 29 | |
Accounting for missing data | missing data within and across studies | 11 (37.9%) | 14 (48.3%) | 4 (13.8%) | N = 29 |
reasons/mechanisms of missingness | 1 (3.4%) | 0 (0%) | 28 (96.6%) | N = 29 | |
how accounted for missing data | 19 (65.5%) | 3 (10.3%) | 7 (24.1%) | N = 29 | |
Imputation Model | Variables included in imputation model | 5 (50%) | 1 (10%) | 4 (40%) | N = 10 |
rationale for variables in imputation model | 0 (0%) | 0 (0%) | 10 (100%) | N = 10 | |
accounted for heterogeneity in imputation model | 4 (40%) | 0 (0%) | 6 (60%) | N = 10 | |
Pooling | assumptions to pool data | 0 (0%) | 0 (0%) | 29 (100%) | N = 29 |
testing any assumptions to pool data? | 0 (0%) | 0 (0%) | 29 (100%) | N = 29 | |
one-step or two-step | 8 (27.6%) | 0 (0%) | 21 (72.4%) | N = 29 | |
Causality | causal methods | 3 (10.3%) a | 0 (0%) | 26 (89.7%) | N = 29 |
justification of methods | 2 (6.9%) | 3 (10.3%) | 24 (82.8%) | N = 29 | |
state assumptions of analysis methods | 4 (13.8%) | 0 (0%) | 25 (86.2%) | N = 29 | |
tested testable assumptions | 3 (10.3%) | 0 (0%) | 26 (89.7%) | N = 29 | |
evaluation of untestable assumptions | 2 (6.9%) | 0 (0%) | 27 (93.1%) | N = 29 | |
investigated the heterogeneity of results | 13 (44.8%) | 5 (17.2%) | 11 (37.9%) | N = 29 | |
the generalizability of results | 10 (34.5%) | 3 (10.3%) | 16 (55.2%) | N = 29 | |
Sensitivity analyses | 23 (79.3%) | 0 (0%) | 6 (20.7%) | N = 29 | |
Confounder Control | how they controlled for clustering | 24 (82.8%) | 0 (0%) | 5 (17.2%) | N = 29 |
labelling of covariates as confounders or mediators | 15 (51.7%) | 5 (17.2%) | 9 (31%) | N = 29 | |
how covariates were selected | 15 (51.7%) | 1 (3.4%) | 13 (44.8%) | N = 29 |