Moderation models testing associations between unprotected anal intercourse, drug use and location of sex during sex between new male partners
| Variable | Any drug use | Poppers use† | Alcohol use† |
|---|---|---|---|
| OR (95% CI) | OR (95% CI) | OR (95% CI) | |
| Home | Reference | Reference | Reference |
| Drug use‡ | 1.47*** (1.21 to 1.79) | 1.61*** (1.28 to 2.03) | 0.98 (0.79 to 1.21) |
| Sex-on-premises venues | 0.43** (0.26 to 0.72) | 0.54** (0.37 to 0.79) | 0.60** (0.42 to 0.85) |
| Sex-on-premises venues×drug use | 2.23* (1.21 to 4.09) | 2.13* (1.15 to 3.93) | 1.81 (0.97 to 3.40) |
| Cruising locations | 1.21 (0.74 to 1.97) | 1.03 (0.64 to 1.65) | 1.18 (0.77 to 1.81) |
| Cruising locations×drug use | 0.73 (0.30 to 1.75) | 1.44 (0.55 to 3.78) | 0.64 (0.17 to 2.43) |
| Wald test (χ2, df, p value) | 33.52, 5, <0.0001 | 78.60, 10, <0.0001 | 77.54, 10, <0.0001 |
*p<0.05, **p<0.01, ***p<0.001.
†These models included additional specific drug use variables.
‡Because moderation models present ‘conditional’ estimates of regression coefficients, the strict interpretation of this coefficient is the increment in sexual risk associated with drug use in sexual encounters at home.