Background: This research examined past microbicide effectiveness trials to better understand how adherence and adherence assessment could be improved in future vaginal microbicide trial design. No product is currently available despite decades of clinical trials of candidate microbicides, yet the need for women to have a method to reduce their risk of sexual transmission of HIV that does not rely on male partner agreement remains urgent. Low product adherence and inaccurate adherence reporting has inhibited the ability of trials to accurately assess the biological efficacy of candidate products. Methods: Three different studies were conducted to examine adherence as a clinical trial design issue. The comparative study examined how five trials measured, calculated, and reported microbicide adherence. The quantitative study used latent class and latent profile analysis and multinomial logistic regression to examine if patterns of adherence could be identified in four trials and, if so, what individual-level factors were associated with the patterns. The qualitative study sought opinions from former trial participants about how to improve adherence and adherence reporting in future microbicide trials through focus group discussion workshops in South Africa and Tanzania. Results: There was diversity in methods used to collect and calculate adherence among the included trials. Two methods to calculate averages of overall adherence were identified. Trial documentation and publications lacked clarity in exact methods used to calculate adherence estimates. Latent structure analysis identified different patterns of adherence in all included trials, and these patterns were similar. Multinomial logistic regression identified factors associated with adherence patterns in all trials. Women join and stay in microbicide trials for their own needs, which are not necessarily related to interest in using the investigational product. Key reasons for joining and staying in trials included access to health care and financial reimbursements. Fear of adverse effects from the investigational product was the most important reason why participants did not use the gel. Participants reported that male partners can act as barriers to gel use and the key reason for inaccurate reporting of gel use was fear of removal from the trial. This study demonstrated that trial teams and participants can work together to develop improved trial designs. Recommendations: There are improvements to be made in how trialists plan, conduct, analyse and report results of microbicide trials. Trial teams can improve the clarity of their trial materials, and use analysis methods to identify patterns of adherence. To improve adherence and trial implementation, trials can test applicators for evidence of vaginal insertion and report results to participants, better engage male partners, develop a less watery gel, and create an atmosphere of transparency and respect between research teams and participants. Identifying HIV prevention products for women requires better understanding of the lives of women asked to join these trials, and application of that understanding to collaboratively develop innovative trial designs that meet both the needs of the research and the needs of participants.