Table 3.
Any AI | Condomless AI | ||||||||
---|---|---|---|---|---|---|---|---|---|
Univariate analysis | Multivariable analysis | Univariate analysis | Multivariable analysis | ||||||
OR | 95% CI | aOR | 95% CI | OR | 95% CI | aOR | 95% CI | ||
HIV status | Seronegative | Ref | Ref | Ref | Ref | ||||
Seropositive | 0.67 | 0.54–0.83 | 1.04 | 0.89–1.28 | 0.45 | 0.35–0.57 | 0.61 | 0.46–0.74 | |
Demographic determinants | |||||||||
Age | Years, continuous | 0.94 | 0.94–0.95 | 0.97 | 0.96–0.98 | 0.94 | 0.93–0.95 | 0.96 | 0.95–0.97 |
Race and ethnicitya | Non-Hispanic Black | Ref | Ref | Ref | Ref | ||||
Hispanic/Latina | 1.50 | 1.17–1.92 | 1.88 | 1.47–2.41 | 1.45 | 1.08–1.94 | 1.65 | 1.25–2.20 | |
Non-Hispanic White | 1.40 | 0.98–1.99 | 1.62 | 1.15–2.30 | 0.98 | 0.63–1.39 | 1.05 | 0.73–1.55 | |
Other | 1.21 | 0.71–2.03 | 1.29 | 0.79–2.03 | 1.23 | 0.66–2.30 | 1.21 | 0.82–2.18 | |
Educationa | < High school | Ref | Ref | Ref | Ref | ||||
≥ High school | 1.20 | 0.98–1.48 | 1.33 | 1.08–1.65 | 1.09 | 0.85–1.40 | 1.25 | 0.97–1.61 | |
Married or living with a partner | No | Ref | Ref | Ref | Ref | ||||
Yes | 1.12 | 0.94–1.33 | 0.88 | 0.76–1.06 | 1.24 | 1.03–1.51 | 1.10 | 0.93–1.32 | |
Household annual income | < $12,000 | Ref | Ref | Ref | Ref | ||||
≥ $12,000 | 1.19 | 1.02–1.39 | 1.14 | 0.98–1.33 | 1.06 | 0.86–1.26 | 0.98 | 0.81–1.15 | |
Violence victimization | None | Ref | Ref | Ref | Ref | ||||
Either physical or sexual | 2.64 | 2.18–3.18 | 1.35 | 1.10–1.62 | 2.94 | 2.35–3.56 | 1.62 | 1.28–2.08 | |
Both physical and sexual | 4.50 | 2.95–7.12 | 1.77 | 1.13–2.82 | 5.82 | 3.37–9.06 | 2.50 | 1.44–4.12 | |
Behavioral determinants | |||||||||
Alcohol use | < 8 drinks/week | Ref | Ref | Ref | Ref | ||||
≥ 8 drinks/week | 1.91 | 1.55–2.31 | 1.27 | 1.05–1.51 | 1.96 | 1.56–2.34 | 1.31 | 1.04–1.62 | |
Crack, cocaine or heroin use | No | Ref | Ref | Ref | Ref | ||||
Yes | 2.08 | 1.69–2.52 | 1.34 | 1.09–1.66 | 2.05 | 1.61–2.62 | 1.35 | 1.02–1.78 | |
Number of male sex partners and exchange sexb | 1 partner, no exchange sex | Ref | Ref | Ref | Ref | ||||
≥ 2 partners, no exchange sex | 2.64 | 2.23–3.11 | 2.10 | 1.79–2.50 | 2.49 | 2.00–3.03 | 1.91 | 1.52–2.33 | |
Any exchange sex | 3.45 | 2.53–4.71 | 2.20 | 1.57–3.13 | 3.21 | 2.13–4.68 | 1.83 | 1.16–2.81 | |
Number of female sex partners | 0 | Ref | Ref | Ref | Ref | ||||
≥ 1 | 1.11 | 0.84–1.47 | 1.32 | 1.01–1.75 | 1.21 | 0.87–1.71 | 1.32 | 0.89–1.92 | |
Condomless VI | No | Ref | Ref | – | – | – | – | ||
Yes | 3.48 | 2.95–4.02 | 1.80 | 1.53–2.09 | – | – | – | – | |
Psycho-social determinants | |||||||||
High depressive symptomsc | No | Ref | Ref | Ref | Ref | ||||
Yes | 1.36 | 1.19–1.53 | 1.23 | 1.08–1.39 | 1.42 | 1.20–1.67 | 1.25 | 1.03–1.48 |
Visits for which AI data were available and for which the prior visit was no longer than 12 months ago (i.e., maximum one visit skipped) were included in the analysis. Most follow-up visits (93.1%) were six months apart, with 5.3% of visits occurring after one skipped visit; applying this criterion therefore removed only 1.6% of visits. Missing values for covariates were imputed (see Methods for details). The covariates HIV status, age and race and ethnicity had no missing values. Education status was missing for 1 woman (measured at baseline). Data on marital status were missing at 4.9% of visits, household income at 5.7%, violence victimization at 39.2%, alcohol use at 3.6%, crack, cocaine, or heroin use at 3.5%, number of male partners and exchange sex at 6.0%, number of female partners at 3.5%, condomless VI at 5.2% and high depressive symptoms at 11.3% of visits
OR odds ratio, aOR adjusted odds ratio, CI confidence interval, VI vaginal intercourse. All covariates were collected over follow-up and represent since the last visit except race and ethnicity and education level, which were measured at baseline. Results in bold indicate that the 95% CI does not include the null value
aVariables measured at baseline. All other variables were measured over follow-up. bWomen reported only one male sex partner at 197 of 1487 visits (13.2%) where exchange sex was reported. cScores of > 15 in the Center for Epidemiologic Studies Depression Scale were defined as high depressive symptoms(36)