Background: In 2014, PEPFAR changed its investment strategy within priority countries. This policy shift increased PEPFAR funds and support to some subnational units (SNU), while leaving investment unchanged?or virtually eliminated?in others, according to HIV disease burden. We aimed to identify shifts in HIV and non-HIV service delivery outcomes associated with changes in PEPFAR investment at the SNU level between 2015-2017 in Kenya and Uganda.
Methods: MEASURE Evaluation extracted quantitative data from relevant national HIV health information system databases (e.g., DHIS 2, TiBU [Kenya], iHRIS, etc.) between 2015-2017. Outcomes examined included HIV testing, initiation on ART, and adherence, ANC 4, or 4+ visits, confirmed malaria cases, and TB case detection rate. Qualitative interviews were conducted with SNU health teams to better understand the trends observed. Longitudinal multivariate analyses were conducted to determine the level of statistically significant changes in study outcomes by year, by change in PEPFAR investment, and for an interaction effect between year and PEPFAR investment level.
Results: Preliminary analyses of data between 2015-2016 suggest changes in PEPFAR investment did not immediately affect the direction of trends observed in HIV or non-HIV outcomes at SNU level. However, the magnitude of changes in outcomes observed in the post-PEPFAR prioritization period did significantly vary for some health outcomes by level of PEPFAR investment. For example, the proportion of newly identified HIV-positive individuals initiated on ART in Ugandan districts that lost PEPFAR funding increased significantly less than the same proportion in districts with maintained or increased PEPFAR support.
Conclusions: Because there were some significant differences in the magnitude of changes in health delivery outcomes (both HIV and non-HIV) associated with PEPFAR investment status at the SNU, findings suggests that shifts in PEPFAR investment (and by extension other large shifts in donor investment) may contribute to or worsen pre-existing health outcome disparities among subnational units within countries. However, as this study used aggregate SNU data, there could be a lag time between the implementation of investment shifts within countries, implementation of new interventions, and when changes in service delivery outcomes can be observed.