Table 3.
Sensitivity analyses of methodological characteristics
| Antenatal depression | Postnatal depression | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Number of studies | Total number of subjects (HIV+/controls) | OR (95% CI) | I2 (%) | P | Number of studies | Total number of subjects (HIV+/controls) | OR (95% CI) | I2 (%) | P | |
| Meta-analysis | ||||||||||
| Random-effects model | 13 | 1520/4383 | 1.42 (1.12, 1.80) | 55 | 10 | 1645/2135 | 1.58 (1.08, 2.32) | 65 | ||
| Fixed-effects model | 13 | 1520/4383 | 1.32 (1.15, 1.52) | 55 | 10 | 1645/2135 | 1.39 (1.13, 1.71) | 65 | ||
| Sub group analysis | ||||||||||
| Methodological quality | 0.31 | 0.04 | ||||||||
| Low | 0 | 0 | ||||||||
| Medium | 6 | 343/1732 | 1.88 (0.93, 3.79) | 76 | 2 | 477/259 | 0.97 (0.64, 1.46) | 0 | ||
| High | 7 | 1177/2651 | 1.29 (1.11, 1.51) | 0 | 8 | 1168/1876 | 1.86 (1.16, 2.98) | 67 | ||
| Study design | 0.11 | 0.13 | ||||||||
| Cross-sectional | 8 | 808/2428 | 1.74 (1.19, 2.54) | 58 | 7 | 1119/1763 | 1.92 (1.11, 3.34) | 72 | ||
| Case-control | 2 | 86/200 | 1.62 (0.59, 4.45) | 73 | 1 | 49/113 | 1.55 (0.73, 3.30) | |||
| Other | 3 | 626/1755 | 1.11 (0.91, 1.36) | 0 | 2 | 477/259 | 0.97 (0.64, 1.46) | 0 | ||
| Depression screening tool | 0.71 | 0.04 | ||||||||
| EPDS | 7 | 1095/3714 | 1.58 (1.13, 2.21) | 67 | 4 | 567/661 | 2.31 (1.17, 4.56) | 69 | ||
| CES-D | 3 | 225/305 | 1.25 (0.63, 2.47) | 68 | 3 | 878/679 | 0.97 (0.72, 1.29) | 0 | ||
| Other | 3 | 200/364 | 1.30 (0.86, 1.99) | 0 | 3 | 200/795 | 1.73 (0.82, 3.65) | 32 | ||
| Geographical locations | 0.03 | 0.19 | ||||||||
| Africa | 11 | 1331/4165 | 1.55 (1.20, 2.01) | 56 | 8 | 1457/1917 | 1.78 (1.12, 2.84) | 68 | ||
| North America | 2 | 189/218 | 0.88 (0.58, 1.35) | 0 | 2 | 188/218 | 1.07 (0.58, 1.97) | |||
Abbreviations: CES-D Centre for Epidemiologic Surveys for Depression, EPDS Edinburgh Postnatal Depression Scale