Evaluation of person-centered interventions to eliminate perinatal HIV transmission in Kisumu County, Kenya: A repeated cross-sectional study using aggregated registry data.

Abstract

Background: Following a decline in perinatal HIV transmission from 20% to 10% between 2010 and 2017 in Kenya, rates have since plateaued with an estimated 8% transmission rate in 2021. Between October 2016 and September 2021, Family AIDS Care & Education Services (FACES) supported HIV care and treatment services across 61 facilities in Kisumu County, Kenya with an emphasis on service strengthening for pregnant and postpartum women living with HIV to reduce perinatal HIV transmission. This included rigorous implementation of national HIV guidelines and implementation of 3 locally adapted evidence-based interventions targeted to the unique needs of women and their infants. We examined whether these person-centered program enhancements were associated with changes in perinatal HIV transmission at FACES-supported sites over time.

Methods and findings: We conducted a repeated cross-sectional study of annually aggregated routinely collected documentation of perinatal HIV transmission risk through the end of breastfeeding at FACES-supported facilities between October 2016 and September 2021. Data included 12,599 women living with HIV with baseline antenatal care metrics, and, a separate data set of 11,879 mother-infant pairs who were followed from birth through the end of breastfeeding (overlapping with those in antenatal care 2 years prior). FACES implemented 3 interventions for pregnant and postpartum women living with HIV in 2019: (1) high-risk clinics; (2) case management; and (3) a mobile app to support treatment engagement. Our primary outcome was infant HIV acquisition by the end of breastfeeding (18 to 24 months). We compared infant HIV acquisition risk in the final year of the FACES program (2021) to the year before intervention scale-up and following implementation of the "Treat All" policy (2018). Mother-infant pair loss to follow-up was a secondary outcome. Program data were aggregated by year and site, thus in multivariable regression, we adjusted for site-level characteristics, including facility type, urban versus rural, number of women with HIV in antenatal care each year, and the proportion among them under 25 years of age. Between October 2016 and September 2021, 81,172 pregnant women received HIV testing at the initiation of antenatal care, among whom 12,599 (15.5%) were living with HIV, with little variation in HIV prevalence over time. The risk of infant HIV acquisition by 24 months of age declined from 4.9% (101/2,072) in 2018 to 2.2% (48/2,156) in 2021 (adjusted risk difference -2.6% [95% confidence interval (CI): -3.7, -1.6]; p < 0.001). Loss to follow-up declined from 9.9% (253/2,556) in 2018 to 2.5% (59/2,393) in 2021 (risk difference -7.5% [95% CI: -8.8, -6.2]; p < 0.001). During the same period, UNAIDS estimated rates of perinatal transmission in the broader Nyanza region and in Kenya as a whole did not decline. The main limitation of this study is that we lacked a comparable control group.

Conclusions: These findings suggest that implementation of person-centered interventions was associated with significant declines in perinatal HIV transmission and loss to follow-up of pregnant and postpartum women.