Geostatistical point process modelling of HIV prevalence using disease progression and CD4 cell count dynamics

KwaZulu-Natal, South Africa, remains the epicentre of the HIV epidemic, characterized by pronounced spatial and demographic heterogeneity. Capturing geographic variation in HIV burden and disease progression is critical for informing targeted public health responses. We used Bayesian geostatistical models to examine spatial patterns of HIV prevalence and immunologic decline, with CD4 cell count dynamics used as temporal indicators of disease progression. Data were obtained from two population-based cross-sectional surveys conducted in uMgungundlovu Municipality under the HIV Incidence Provincial Surveillance System. Spatial variation in HIV prevalence was modelled using geostatistical point-process methods with enumeration area structure and spatially continuous random effects. CD4 outcomes were analyzed in relation to demographic, socio-economic, and clinical factors, with spatial clustering assessed using posterior exceedance probabilities. Distinct hotspot areas were identified, reflecting local variation in HIV risk and progression. These findings highlight the value of spatial modeling for guiding geographically targeted HIV prevention and treatment strategies.