Contraceptive method preferences and use among Kenyan women living with HIV

Women living with HIV (WLWH) face unique challenges relating to contraception, and understanding how preferences influence method selection can improve contraceptive counseling. Data are lacking on how contraceptive attribute preferences relate to attributes of methods. We determined relationships between contraceptive attribute preferences and attributes of methods used, identified correlates of contraceptive satisfaction, and identified preference clusters. Methods: We used baseline data from 2479 Kenyan WLWH enrolled in a cluster randomized control trial of a reproductive health counseling intervention. Participants for the current study included women who used contraception based on baseline assessment. We used self-administered electronic survey data to measure relationships between contraceptive attribute preferences (including long-acting and effectiveness) and attributes of methods used. We examined correlates of contraceptive satisfaction using Poisson regression. Principal components analysis (PCA) was used to identify dimensions of contraceptive preferences. Results: Median age was 33 years (interquartile range: 28-38). Most (66%) of women were married/cohabitating. The most common contraceptive methods used were implants (36%), injectables (35%), and condoms (34%), followed by oral contraceptive pills (OCPs) (8%), intrauterine devices (4%), and tubal ligation (2%). Most participants preferred methods that avoid intermittent bleeding (85%), are effective (73%), and avoid daily dosing (71%). Women who preferred methods that were long-acting (Prevalence Ratio [PR] 1.36, 95% Confidence Interval [CI]: 1.20-1.55), avoid daily dosing (PR 1.10, 95% CI: 1.00-1.21), permit self-discontinuation (PR 1.34, 95% CI: 1.22-1.48), avoid intermittent bleeding (PR 1.43, 95% CI: 1.20-1.74), and are non-hormonal (PR 1.19, 95% CI: 1.05-1.36) were significantly more likely to use methods with these attributes compared to women who did not hold these preferences. Relationships remained significant in adjusted models. Participants with agreement between preference and use of methods that are long-acting or avoid intermittent bleeding were more likely to have high contraceptive method satisfaction. In PCA, preferences for attributes were found to cluster on three dimensions: (1) avoid heavy bleeding, weight changes, libido changes, and non-hormonal; (2) long-acting, avoid daily dosing, permit self-discontinuation, and avoid intermittent bleeding; and (3) concealable without an effectiveness preference. Immediate return to fertility was independent and found to not cluster with other variables, and was therefore modeled separately. The first and second dimensions were correlated with condom use. The second dimension was correlated with having a tubal ligation but inversely correlated with OCP use while the third was correlated with injectable use, and preferences for immediate return to fertility were correlated with condom use but inversely correlated with having a tubal ligation.