Table 4. Multivariable model* of screening factors associated with incident pregnancy.
| Unadjusted Rate Ratio | AdjustedΛ Rate Ratio (95% CI) | p-value | |
| Contraception supplied at site | |||
| Yes | 0.68 | 0.43 (0.22–0.86) | 0.02 |
| No | 1 | 1 | |
| Consistent condom use (in 6 months prior to screening) | |||
| Yes | 0.78 | 0.54 (0.28–1.04) | 0.07 |
| No | 1 | 1 | |
| Using injectable contraception at screening | |||
| Yes | 0.46 | 0.37 (0.21–0.67) | 0.0009 |
| No | 1 | 1 | |
| Marijuana use, heavy drinking ** and/or use during sex (in 6 months prior to screening) | |||
| ≥ 2 such activities | 2.02 | 2.66 (1.24–5.72) | 0.05 |
| 1 such activity | 0.80 | 0.85 (0.36–2.02) | |
| None | 1 | 1 | |
| Partner type by knowledge of HIV status (in 6 months. prior to screening) | |||
| >2 partners mixed knowledge of status | 1.04 | 1.01 (0.27–3.74) | 0.01 |
| 2 partners both unknown status | 3.42 | 4.42 (1.59–12.29) | |
| 1 partner known negative | 2.03 | 2.34 (1.16–4.73) | |
| 1 partner unknown status | 1 | 1 | |
*
Final model does not include 3 women with known HIV-positive partners as this stratum was too thin for modeling.
Λ
Adjusted for all other variables in the model; for variables with >2 levels the overall p value shown.
**
Heavy drinking defined as 5 or more alcoholic drinks in one day.