Non-adherence among women enrolled in a contraceptive vaginal ring use study in Kisumu, Kenya, 2014-2015

# Background Given future potential use of vaginal rings to prevent HIV infection, we examined the association of contraceptive vaginal ring (CVR) non-adherence with user dissatisfaction, tolerability, demographic, and behavioral factors. # Methods In an open-label single-group study, sexually active women aged 18-34 years using oral or injectable hormonal contraception, conveniently sampled from general population, were assigned to 6-month use of a commercial CVR currently not licensed for use in Kenya. Non-adherence in any CVR cycle completed was assessed from: (1) self-report (not used for at least 1 day), and (2) pharmacy record (failure to timely receive a new CVR or return a used one). Additionally, non-adherence was assessed in a subset of participants by residual progestin and estrogen levels measured in returned CVRs. # Results Of 202 participants who underwent CVR insertion by a study clinician, 142 completed all 6 visits, 172 responded to questions about ring use, and 43 provided used CVRs from months 1, 3, and 6 for residual hormone analysis. Non-adherence was 14.0% (24/172) by self-report and 54.5% (110/202) by pharmacy record. Non-adherence by pharmacy record was significantly reduced among women with a salary-based income (prevalence ratio (PR) 0.71, 95% confidence interval (CI) (0.55-0.91)\] compared to women with income not salary-based or no income. Participants dissatisfied with CVR on ≥4 aspects (ambiguity of instructions, inconvenience of use, sensation, sexual discomfort, etc.) were more likely to report non-adherence (PR 2.69, 95% CI=(1.31-5.52)\] compared to those dissatisfied with ≤3 aspects. Non-adherence by residual hormone levels was identified in 46.5% (20/43) participants. Over time, this subset of participants showed increasing non-adherence (*P*=0.004). We found lack of agreement among the various measures of non-adherence. # Conclusions Economic empowerment interventions, especially those emphasizing partner-independent income options, and expanded education on CVR features may alleviate non-adherence. Addressing CVR dissatisfaction preemptively may also help mitigate non-adherence.