Table 3. Meta-analytic Comparison of More Comparable vs Less Comparable Recruitment for Maternal and Infant Confounding Variables.
| Factor | More comparablea | Less comparableb |
|---|---|---|
| Maternal factors | ||
| Race/ethnicity | ||
| SMD (95% CI) | −0.63 (−1.31 to 0.05) | −0.45 (−2.22 to 1.33) |
| I2, % | 87 | 79 |
| Individuals, No.c | 578 | 156 |
| Source | Hans cohort,27,28 Marcus cohort,31,32 Rosen cohort,34,35,36,37 Strauss cohort,39,40 Whitham,44 Wilson cohort,45,46,47 Woodward cohort48,49,50 | Bakhireva et al,23 Chasnoff et al,26 Rosen and Johnson,33 |
| Education | ||
| SMD (95% CI) | −0.47 (−1.59 to 0.65) | −0.56 (−1.64 to 0.51) |
| I2, % | 82 | 88 |
| Individuals, No.c | 251 | 361 |
| Source | Chasnoff et al,26 Hans cohort,27,28 Strauss cohort,39,40 Wilson cohort,45,46,47 | Bakhireva et al,23 van Baar cohort,41,42,43 Whitham,44 Woodward cohort48,49,50 |
| Socioeconomic status | ||
| SMD (95% CI) | −0.60 (−1.60 to 0.41) | −0.56 (−1.64 to 0.51) |
| I2, % | 88 | 88 |
| Individuals, No.c | 339 | 361 |
| Source | Hans cohort,27,28 Kaltenbach and Finnegan,30 Johnson,33 Rosen cohort,34,35,36,37 Strauss cohort,39,40 Wilson cohort,45,46,47 | Bakhireva et al,23 van Baar cohort,41,42,43 Whitham,44 Woodward cohort48,49,50 |
| Child factors | ||
| Infant sex | ||
| SMD (95% CI) | −0.40 (−1.35 to 0.55) | −0.84 (−1.40 to −0.28) |
| I2, % | 87 | 67 |
| Individuals, No.c | 400 | 427 |
| Source | Bakhireva et al,23 Bunikowski et al,25 Rosen cohort,34,35,36,37 Strauss cohort,39,40 Whitham,44 Wilson cohort,45,46,47 | Hans cohort27,28 Hunt et al,29 Rosen and Johnson,33 van Baar cohort,41,42,43 Woodward cohort48,49,50 |
| Prenatal tobacco exposure | ||
| SMD (95% CI) | −0.11 (−0.42 to 0.20) | −1.19 (−2.00 to −0.39) |
| I2, % | 0 | 87 |
| Individuals, No.c | 246 | 448 |
| Source | Bunikowski et al,25 Chasnoff et al,26 Whitham,44 Wilson cohort,45,46,47 25,26,44,45,46,47 | Bakhireva et al,23 Johnson,33 Rosen cohort,34,35,36,37 Salo et al,38 van Baar cohort,41,42,43 Woodward cohort48,49,50 |
Abbreviation: SMD, standardized mean difference.
a
More comparable was defined as proportions in the exposed and unexposed groups that were similar within a study or cohort. Within 10% of exposed group value was used as threshold.
b
Less comparable was defined as greater than 10% and had to be explicitly reported by the authors. Studies that did not report the data for both exposed and unexposed groups were excluded.
c
Total number of individuals (exposed and unexposed combined).