Evaluation of differentiated service delivery models on HIV treatment retention among key populations in Nigeria: a prospective cohort analysis

IntroductionDifferentiated Service Delivery models (DSD) that cater to the needs of key populations (KP) have shown promise in providing KP-sensitive and KP-tailored care and treatment services. We evaluated the effect of two DSD models on linkage to and retention in treatment in Nigeria.MethodsBetween December 2017 and June 2018, newly-diagnosed men who have sex with men (MSM) and female sex workers (FSWs) were enrolled into treatment through two DSD models and followed prospectively for one year. Model 1 was a fully integrated one-stop-shop (OSS) while model 2 was a hybrid HIV prevention and treatment site (DIC). Retention was estimated from drug pick-up records and was defined as being on treatment within 90 days one-year post ART initiation. Cox regression was used to identify the independent effect of the DSD models on retention while probability of being retained in treatment at 1-year was estimated with Kaplan-Meier product limit.ResultsA total of 605 newly-diagnosed clients were enrolled into the study (340 in OSS and 265 in DIC; 342 were FSWs, while 263 were MSM). Median age was 26 years for MSM and 30 years for FSWs. Among those linked to treatment, retention was higher in the OSS than in DIC (63% vs. 48%; p = 0.002). Among those not retained, mean days to be lost-to-follow up (LFTU) was 60 days. Controlling for educational level, population type and age, clients who received treatment in M2 were 6 times more likely not to be retained in treatment at the end of 1 year (Hazard ratio 5.89; 95% CI: 1.04-33.16). The Kaplan Meier estimates of the probability of being retained in 6 months, 9 months and 12 months was 0.97, 0.92, 0.80 for the DIC and 1.00, 0.96, 0.91 for the OSS.ConclusionLinkage to treatment was suboptimal across both models with less than 90% of newly-diagnosed positives initiated on treatment and suggests that facilitated linkage is required to achieve at least 95%. Though retention was higher in the OSS model, both models had an alarmingly short time frame for LTFU and thus intensive monitoring during this phase of treatment is recommended irrespective of the type of DSD. Addressing these gaps will improve service delivery by one stop shops for HIV care and treatment.