Skip to main content
. Author manuscript; available in PMC: 2024 Sep 3.
Published in final edited form as: Lancet Digit Health. 2024 Mar;6(3):e187–e200. doi: 10.1016/S2589-7500(23)00252-2

Table 2:

Regression results in the odds scale for current PrEP use over time by variable

OR estimate 95% PI 98·75% PI
Intercept 0·01 0·00–0·01 0·00–0·01
 AMMI plus coaching group 1·73 0·57–5·57 0·42–7·75
 AMMI plus peer support group 1·14 0·38–3·44 0·29–4·63
 AMMI plus peer support and coaching group 0·88 0·27–2·92 0·20–4·08
Visit 2·15 1·26–3·72 1·08–4·30
 Visit × AMMI plus coaching group 0·89 0·50–1·58 0·43–1·90
 Visit × AMMI plus peer support group 1·30 0·75–2·30 0·64–2·69
 Visit × AMMI plus peer support and coaching group 2·31 1·28–4·14 1·12–4·86
Visit2 0·72 0·59–0·88 0·56–0·92
 Visit2 × AMMI plus coaching group 1·01 0·92–1·11 0·90–1·14
 Visit2 × AMMI plus peer support group 0·96 0·87–1·04 0·85–1·07
 Visit2 × AMMI plus peer support and coaching group 0·89 0·82–0·98 0·79–1·01
Visit3 1·04 1·02–1·06 1·01–1·07

The reference group is AMMI only. Visit is the time variable of these data where time between each visit is about 4 months. Visit is linear, Visit2 is quadratic, and Visit3 is cubic. The 98·75% PI is based on conservative Bonferroni correction across the four primary outcomes. AMMI=automated text messaging and monitoring. OR=odds ratio. PI=posterior interval. PrEP=pre-exposure prophylaxis.