Multiple PrEP modalities—including daily oral tenofovir disoproxil fumarate/emtricitabine (TDF/FTC), topical rings, and long-acting injectables—have demonstrated efficacy for cisgender women in randomized trials. Within oral options, on-demand regimens such as 2-1-1 PrEP (two pills before sex and one pill on each of the two days after sex) and other event-driven schedules have proven effective in cisgender men who have sex with men and transgender women, but only daily oral PrEP is currently recommended for cisgender women. Barriers to high daily adherence among women include intimate partner violence, childcare and domestic responsibilities, stigma, misinformation, and clinic uptake delays. Some evidence also suggests that missed doses may have greater consequences for protection in women than in MSM/TGW, prompting debate about whether on-demand oral PrEP should be offered to cisgender women.
The HPTN 067 trial examined non-daily PrEP strategies and collected detailed pill-taking and sexual behavior data. These data provide an opportunity to model potential effectiveness of non-daily regimens for women and to explore whether offering on-demand options could improve individual-level acceptability, lower pill burden, and increase effective protection for particular subgroups.
The authors used adherence and sexual behavior data from the HPTN 067 Cape Town site to construct a synthetic cohort of PrEP users. In HPTN 067, women had comparatively high adherence to daily PrEP, with 75% of reported sex acts covered by PrEP and drug detected in 68% of samples when sex had been reported in the previous week. Adherence to non-daily regimens in the trial was lower: the time-driven arm had 56% of reported sex acts covered and drug detected in 58% of weekly reported-sex samples, and the event-driven arm had 52% coverage and drug detected in 41% of such samples.
Using these observed patterns of sex-act frequency and pill-taking (from electronic pill-bottle monitoring, weekly interviews, and pharmacokinetics), the investigators simulated individual-level pill use around sex acts across alternative regimens and estimated expected protection.
Four oral PrEP regimens were modeled for six months each in the simulated cohort:
Effectiveness was estimated using published adherence–efficacy relationships for oral TDF/FTC and by counting the number and timing of pills taken in relation to sex acts in the simulation. The model incorporated the observed relationship between regularly scheduled pill taking and pills taken around expected sex, as well as individual variation in ability to predict sex acts.
Median estimated effectiveness across the synthetic cohort was:
These estimates reflect the combination of sex-act coverage and adherence patterns observed or inferred from HPTN 067. The model showed that, on average, daily PrEP provided the highest estimated protection in this cohort, consistent with observed high daily adherence at the Cape Town site.
A subgroup of simulated users with low adherence to daily PrEP—defined as taking fewer than 2.8 pills per week—comprised about 8% of the cohort. In this low-adherence group, switching from daily to 2-1-1 PrEP produced a modest increase in estimated effectiveness (from 41% to 44%) while reducing weekly pill taking (from a median of 2.2 to 1.4 pills per week).
At the population level, however, overall effectiveness did not change when this low-adherence subgroup switched to on-demand PrEP. The model also evaluated person-centered assignment strategies and found no advantage to allocating 2-1-1 PrEP based solely on sex-act frequency.
The authors interpret their findings to indicate that offering on-demand PrEP could be a useful additional option for cisgender women—particularly for those with low adherence to daily pills—because it can modestly increase individual-level effectiveness while lowering pill burden. Where daily oral PrEP is already available, providing 2-1-1 or other on-demand options may be straightforward to implement and could improve acceptability and persistence for some women.
The modeled results also caution that the greatest population-level gains depend on the proportion of women who would both benefit and choose to switch; in this simulation the small size of the low-adherence subgroup limited effects on overall effectiveness. The analysis further suggests that selecting candidates for on-demand PrEP based on sex-act frequency alone may not improve population impact.
The HPTN 067 dataset used for this analysis is publicly available, and the modeling code is provided in a GitHub repository. Funding was provided by the U.S. National Institute of Allergy and Infectious Diseases; the funders did not influence study design, analysis, or reporting. The authors declared no competing interests.