Table 3: Bivariate associations between hypothesized predictor variables and highest access to HIV services.

CondomsLubricantsHIV testing*HIV treatment*
ORCI ORCI ORCI ORCI

Homophobia0.530.48–0.590.0000.390.35–0.440.0000.380.33–0.420.0000.380.29–0.490.000
Violence-MSM0.840.76–0.930.0010.670.58–0.780.0000.730.65–0.830.0000.640.51–0.800.000
Violence-HIV*0.600.43–0.840.003
Provider stigma0.570.46–0.700.0000.690.53–0.910.0070.610.47–0.800.0000.770.54–1.100.152
PrEP stigma
 Outness1.201.14–1.260.0001.221.15–1.300.0001.371.29–1.460.0001.281.08–1.510.005
 Community  Engagement1.591.40–1.790.0001.531.34–1.750.0001.721.49–1.980.0001.331.03–1.720.032
 Connection to gay  community1.411.29–1.530.0001.411.27–1.560.0001.541.40–1.690.0001.261.05–1.530.014
 Comfort with provider1.721.59–1.860.0001.971.78–2.190.0002.402.17–2.660.0001.741.45–2.090.000
PrEP knowledge
Age (measured in decades)1.081.03–1.140.0041.231.16–1.300.0001.411.32–1.500.0001.181.02–1.360.027
Country income 0.0000.0000.0000.000
 High income (referent)
 Low income0.600.38–0.930.150.07–0.340.430.27–0.700.190.06–0.63
 Lower-middle income0.550.45–0.680.300.23–0.390.260.20–0.320.380.21–0.67
 Upper-middle income0.490.42–0.570.320.27–0.390.290.24–0.350.560.38–0.83
HIV status 0.5650.542
 No HIV positive
 (referent)
 HIV positive1.050.88–1.251.070.87–1.31
Education 0.5140.0780.9850.381
 No postsecondary
 (referent)
 Postsecondary0.940.79–1.130.830.67–1.021.000.82–1.220.830.55–1.25
Personal income 0.0260.0200.0000.246
 None (referent)
 Low income/impoverished1.220.88–1.711.170.79–1.721.400.95–2.061.130.40–3.16
 Low-middle 0.930.72–1.210.970.71–1.311.060.79–1.410.960.37–2.44
 Middle 1.190.92–1.531.240.92–1.671.321.01–1.730.830.33–2.13
 High 1.240.87–1.771.541.03–2.292.001.37–2.931.930.59–6.26
Housing status 0.0000.0000.0000.078
 Stable place to live
 (referent)
 Unstable or no place0.670.560.800.600.470.750.560.460.690.700.47–1.04
Minority status 0.0010.0250.0300.127
 Not minority (referent)
 Minority1.301.111.521.221.021.451.211.021.440.780.57–1.07

Violence-HIV and HIV treatment access were only measured among participants who reported being HIV positive. Bivariate associations with HIV testing access were calculated among HIV negative participants.