TABLE 4.
Estimates of the temporal relationship between HIV risk indicators at baseline and subsequent perpetration of IPV at follow up (H2) after propensity score matching: adjusted ORs and 95% CI
| HIV risk indicators at baseline | Perpetration of physical/injurious IPV at follow-up | |
|---|---|---|
| N | OR (95% CI)a | |
| Always condom use during vaginal sex (vs. at least one or more unprotected act) | 201 | 0.5 (0.3, 1.1)* |
| Sometimes condom use during vaginal sex (vs. never) | 92 | 0.3 (0.1, 0.8)** |
| Any STI (vs. no) | 34 | 5.6 (0.98, 31.6)* |
| HIV positive (vs. negative/unknown) | 98 | 0.6 (0.2, 1.8) |
| Having > 1 intimate partner (vs. no) | 144 | 1.5 (0.6, 3.8) |
| Buying sex (vs. no) | 52 | – |
| Injected drugs (vs. no) | 129 | 3.7 (1.1, 11.9)** |
| Partner HIV positive (vs. negative/unknown) | 50 | 1.5 (0.2, 13.1) |
| Partner having >1 partner (vs. no) | 36 | – |
| Partner injected drugs (vs. no) | 52 | 1.5 (0.1, 26.0) |
| 100% protected vaginal sex with other nonmain partners (vs. any unprotected)b | 349 | 2.1 (0.5, 8.6) |
| Any sexual coercion (vs. no) | 110 | 2.6 (1.2, 5.4)** |
*p < 0.1
**p < 0.05
aThe confounders for propensity score matching are participant’s characteristics (age, ethnicity, education, unemployment, length on methadone), relationship with partner (length of relationship, type of relationship, contribution to household expense), participant’s substance use (binge drinking and illicit drug use), and partner’s substance use (binge drinking and illicit drug use). Baseline measurement of the outcome was included in the logistic regression analysis.
bFor “100% protected vaginal sex with other nonmain partners,” those who did not have any outside partner were coded as 100% protected.