Poor participant adherence to treatment may contribute to lack of impact in some biomedical HIV prevention trials. This qualitative study explored adherence in a randomized controlled trial of herpes suppressive therapy to reduce HIV acquisition and infectivity among 1305 Tanzanian women. The trial found participants completed 72% of visits on treatment; 52-56% of women on treatment had > or = 90% adherence by pill count estimate; and between six and nine months 30/86 (35%) of urine samples from acyclovir recipients tested acyclovir negative, and 7/86 (8%) from placebo recipients tested acyclovir positive. Twenty in-depth interviews (IDIs) were conducted after 30 months with respondents randomly selected from "acyclovir negative" acyclovir recipients and "acyclovir positive" placebo recipients, or by preliminary pill count adherence categories ("under users," "good users," and "over users"). Almost all respondents reported appropriate adherence and positive trial attitudes, e.g., trusting staff, appreciating services, perceiving pills as beneficial. Fourteen understood placebo use, and six understood the trial purpose. Notably, 5/9 acyclovir recipients and 1/11 placebo recipients believed their pills had treated pre-existing sexually transmitted infections. Limited understanding did not negatively affect reported adherence. Reported adherence problems usually related to illness, travel, and/or family obligations (e.g., husband's disapproval). "Acyclovir positive" placebo recipients denied taking other participants' pills. The IDIs also did not resolve discrepant reports of pill loss or theft. Biomedical HIV interventions often have strong behavioral components that require close attention during intervention development, trial design, and process and impact evaluation. This study identified topics which warrant further consideration, including: information reinforcement and comprehension assessment throughout a trial for long-term participant understanding; involving partners in adherence promotion activities; strategizing with participants to maintain adherence during familial illnesses or other crises; and close monitoring, identification, and follow-up of (1) individuals with discrepant biological tests, and (2) other sources of the treatment in the trial area. Methodological research is also needed to improve adherence measures.