State-level clustering in PrEP implementation factors among family planning clinics in the Southern United States

BACKGROUND: Availability of PrEP-providing clinics is low in the Southern US. Little is known about factors influencing PrEP implementation in Southern states, which vary in state-level policies. We explored state-level clustering of organizational constructs relevant to PrEP implementation in family planning (FP) clinics in the Southern US.
METHODS: We surveyed providers and administrators of publicly-funded FP clinics not providing PrEP in 18 Southern states (Feb-Jun 2018, N=414 respondents from 224 clinics). The Consolidated Framework for Implementation Research (CFIR) informed construct selection, including readiness to implement PrEP and others previously associated with PrEP readiness: PrEP knowledge/attitudes, implementation climate, leadership engagement, and available resources. We analyzed each construct using linear mixed models with fixed effects for state, provider, and clinic-level covariates, and a random effect for clinic. A principal component analysis of the resulting construct-specific, state-level fixed effects (8 states excluded due to insufficient data) identified 6 principal components, which were inputted into a K-means clustering analysis to examine state-level clustering.


RESULTS: Three clusters (C1-3) were identified with 5, 3, and 4 states, respectively (Figure 1). Canonical variable 1 (CV1; x-axis) separated C1 and C2 from C3 and was primarily driven by PrEP readiness, HIV-specific implementation climate, PrEP-specific leadership engagement, PrEP attitudes, PrEP knowledge, and general resource availability. Canonical variable 2 (CV2; y-axis), which distinguished C2 from C1, was primarily driven by PrEP-specific resource availability, PrEP attitudes, and general implementation climate. All C3 states had expanded Medicaid, compared to 1 C1 state (none in C2).
CONCLUSIONS: CFIR constructs relevant for PrEP implementation exhibited state-level clustering, suggesting that tailored strategies could be used by clustered states to improve PrEP provision in FP clinics. Medicaid expansion was a common feature in states within C3, which could explain the similarity of their implementation constructs. The role of Medicaid expansion and state-level policies on PrEP implementation warrants further exploration.