Background: Lifelong antiretroviral therapy for HIV infected pregnant and lactating women (Option B+) has been rapidly scaled up but there are concerns about poor retention of women initiating treatment. However, facility-based data could underestimate retention in the absence of measures to account for self-transfers to other facilities. We assessed retention-in-care among women on Option B+ in Uganda, using facility data and follow-up to ascertain transfers to other facilities. Methods: In a 25-month retrospective cohort analysis of routine program data, women who initiated Option B+ between March 2013 and March 2015 were tracked and interviewed quantitatively and qualitatively (in-depth interviews). Kaplan Meier survival analysis was used to estimate time to loss-to-follow-up (LTFU) while multivariable Cox proportional hazards regression was applied to estimate the adjusted predictors of LTFU, based on facility data. Thematic analysis was done for qualitative data, using MAXQDA 12. Quantitative data were analyzed with STATA® 13. Results: A total of 518 records were reviewed. The mean (SD) age was 26.4 (5.5) years, 289 women (55.6%) attended primary school, and 53% (276/518) had not disclosed their HIV status to their partners. At 25 months post-ART initiation, 278 (53.7%) were LTFU based on routine facility data, with mean time to LTFU of 15.6 months. Retention was 60.2 per 1000 months of observation (pmo) (95% CI: 55.9-64.3) at 12, and 46.3/1000pmo (95% CI: 42.0-50.5) at 25 months. Overall, 237 (55%) women were successfully tracked and interviewed and 43/118 (36.4%) of those who were classified as LTFU at facility level had self-transferred to another facility. The true 25 months post-ART initiation retention after tracking was 71.3% (169/237). Women < 25 years, aHR = 1.71 (95% CI: 1.28-2.30); those with no education, aHR = 5.55 (95% CI: 3.11-9.92), and those who had not disclosed their status to their partners, aHR = 1.59 (95% CI: 1.16-2.19) were more likely to be LTFU. Facilitators for Option B+ retention based on qualitative findings were adequate counselling, disclosure, and the desire to stay alive and raise HIV-free children. Drug side effects, inadequate counselling, stigma, and unsupportive spouses, were barriers to retention in care. Conclusions: Retention under Option B+ is suboptimal and is under-estimated at health facility level. There is need to institute mechanisms for tracking of women across facilities. Retention could be enhanced through strategies to enhance disclosure to partners, targeting the uneducated, and those < 25 years.
