Objectives: In VOICE, a phase IIB trial of daily oral and vaginal tenofovir for HIV prevention, at least 50% of women receiving active products had undetectable tenofovir in all plasma samples tested. MTN-003D, an ancillary study using in-depth interviews (IDIs) and focus group discussions (FGDs), together with retrospective disclosure of plasma tenofovir pharmacokinetic results, explored adherence challenges during VOICE. Methods: We systematically recruited participants with pharmacokinetic data (median six plasma samples), categorized as low (0%, N = 79), inconsistent (1-74%, N = 28) or high (≥75%; N = 20) on the basis of frequency of tenofovir detection. Following disclosure of pharmacokinetic results, reactions were captured and adherence challenges systematically elicited; IDIs and FGDs were audio-recorded, transcribed, coded and thematically analysed. Results: We interviewed 127 participants from South Africa, Uganda and Zimbabwe. The most common reactions to pharmacokinetic results included surprise (41%; low pharmacokinetic), acceptance (39%; inconsistent pharmacokinetic) and happiness (65%; high pharmacokinetic). On the basis of participants' explanations, we developed a typology of adherence patterns: noninitiation, discontinuation, misimplementation (resulting from visit-driven use, variable taking, modified dosing or regimen) and adherence. Fear of product side effects/harm was a frequent concern, fuelled by stories shared among participants. Although women with high pharmacokinetic levels reported similar concerns, several described strategies to overcome challenges. Women at all pharmacokinetic levels suggested real-time drug monitoring and feedback to improve adherence and reporting. Conclusion: Retrospective provision of pharmacokinetic results seemingly promoted candid discussions around nonadherence and study participation. The effect of real-time drug monitoring and feedback on adherence and accuracy of reporting should be evaluated in trials.