Table 3:
Regression results in odds scale for current PrEP use over time controlling for covariates by variable
| OR estimate | 95% PI | 98·75% PI | |
|---|---|---|---|
| Intercept | 0·00 | 0·00–0·00 | 0·00–0·00 |
| AMMI plus coaching group | 1·73 | 0·49–6·31 | 0·33–9·15 |
| AMMI plus peer support group | 0·90 | 0·26–3·00 | 0·19–4·25 |
| AMMI plus peer support and coaching group | 0·69 | 0·20–2·39 | 0·14–3·45 |
| Visit | 1·89 | 1·01–3·56 | 0·86–4·30 |
| Visit × AMMI plus coaching group | 0·99 | 0·51–1·91 | 0·42–2·31 |
| Visit × AMMI plus peer support group | 1·59 | 0·84–2·99 | 0·71–3·50 |
| Visit × AMMI plus peer support and coaching group | 2·75 | 1·42–5·34 | 1·22–6·45 |
| Visit2 | 0·76 | 0·60–0·95 | 0·56–1·01 |
| Visit2 × AMMI plus coaching group | 0·99 | 0·89–1·09 | 0·86–1·13 |
| Visit2 × AMMI plus peer support group | 0·93 | 0·84–1·03 | 0·82–1·05 |
| Visit2 × AMMI plus peer support and coaching group | 0·87 | 0·79–0·97 | 0·76–0·99 |
| Visit3 | 1·03 | 1·01–1·06 | 1·00–1·07 |
| Age | 1·26 | 1·06–1·51 | 1·01–1·58 |
| Gender* | |||
| Cisgender | 1 (ref) | ·· | ·· |
| Transgender or gender diverse | 2·50 | 0·83–7·67 | 0·61–10·28 |
| Race or ethnicity | |||
| White (non-Hispanic) | 1 (ref) | ·· | ·· |
| Black or African American | 1·37 | 0·51–4·08 | 0·39–5·73 |
| Latinx or Hispanic | 1·27 | 0·47–3·60 | 0·35–4·98 |
| Other† | 1·36 | 0·36–5·27 | 0·23–7·82 |
| Sexual orientation* | |||
| Gay | 1 (ref) | ·· | ·· |
| Bisexual | 0·22 | 0·09–0·56 | 0·07–0·72 |
| Other | 0·53 | 0·16–1·72 | 0·11–2·35 |
| City | |||
| Los Angeles, CA, USA | 1 (ref) | ·· | ·· |
| New Orleans, LA, USA | 1·81 | 0·75–4·23 | 0·59–5·33 |
| Insured | 4·06 | 2·47–6·79 | 2·19–7·90 |
| Income above federal poverty level (2021) | 1·90 | 0·89–3·92 | 0·72–4·94 |
| Condomless anal sex in the past year | 1·20 | 0·55–2·63 | 0·44–3·30 |
| Homelessness (in lifetime) | 0·37 | 0·14–0·93 | 0·11–1·18 |
| Incarceration (in lifetime) | 0·20 | 0·06–0·64 | 0·04–0·89 |
| Popper use (in lifetime) | 2·50 | 1·10–5·78 | 0·82–7·42 |
| Sex exchange (in lifetime) | 0·91 | 0·35–2·27 | 0·28–2·84 |
| Condomless sex with partner with HIV (in lifetime) | 17·43 | 5·74–55·24 | 4·23–75·44 |
| Number of recent sexual partners (in past 4 months)‡ | 3·03 | 2·32–3·98 | 2·16–4·30 |
| Intimate partner violence (in lifetime) | 0·87 | 0·39–1·90 | 0·31–2·48 |
| Used support services in past 4 months | 0·94 | 0·62–1·44 | 0·56–1·60 |
| After COVID-19§ | 0·90 | 0·54–1·48 | 0·47–1·71 |
| Enrolment date¶ | 0·93 | 0·75–1·15 | 0·71–1·22 |
The 98·75% PI is based on conservative Bonferroni correction across the four primary outcomes. Visit is linear, Visit2 is quadratic, and Visit3 is cubic. AMMI=automated text messaging and monitoring. OR=odds ratio. PI=posterior interval. PrEP=pre-exposure prophylaxis.
Some gender and sexual orientation response categories shown in table 1 were combined for longitudinal analyses due the small proportion of people in some categories.
Asian, Hawaiian or Pacific Islander, native American, Alaskan native, and other are combined due to the small proportion of people in some categories.
The cubed root of this variable was taken to ameliorate positive skew.
Follow-up visit occurred after March 16, 2020.
Centred at midpoint enrolment time and scaled by 100 days.