Dataset for "<b>Spatial analysis of HIV incidence and relationships with determinants in Ghana, 2018 to 2022"</b>

Background: Sustainable Development Goal (SDG) 3.3 aims to end the HIV/AIDS epidemic by 2030. Identifying areas with high HIV burden is essential for targeted interventions, but spatial evidence in Ghana remains limited. This study examines the spatial distribution of HIV burden and its association with socioeconomic and health-related factors to guide targeted interventions. Methods: We analyzed HIV incidence in Ghana (2018-2022) using data from the Ghana AIDS Commission and the 2022 Ghana Demographic Health Survey (GDHS). Crude incidence rates (per 1,000 population) were calculated and smoothed with Global and Local Empirical Bayesian methods. Spatial dependence and clustering were evaluated using Global Moran’s I, Local Indicators of Spatial Association (LISA), and Getis-Ord statistics, while space-time clusters were identified with SaTScan. Spatial associations between HIV incidence and its determinants were assessed using bivariate LISA. Predictor importance was quantified using Gradient Boosting, a nonparametric ensemble learning approach, with interpretability enhanced by SHapley Additive exPlanations (SHAP), including feature importance rankings, partial dependence, and waterfall plots. Results: From 2018 to 2022, Ghana reported 1,727,897 new HIV cases, with crude incidence rates (CIR) highest in Greater Accra (15.57/1,000), Bono (15.36/1,000), and Eastern (14.42/1,000), and lowest in Northern (2.69/1,000) and North East (2.59/1,000). Empirical Bayesian smoothing confirmed elevated risk in Greater Accra, Ahafo, Ashanti, Eastern, and Bono. Spatial analysis indicated significant clustering (Global Moran’s I, p<0.01), with high-risk areas concentrated in the southern and middle belt and low-risk areas in the north. Local Indicators of Spatial Association (LISA) identified a High–High hotspot in Ashanti and Low–Low coldspots in Northern, North East, Upper East, and Upper West, while Getis–Ord Gi* confirmed these patterns (p<0.05). Space–time scan statistics detected three significant clusters (p<0.05): Greater Accra (2018–2019, RR = 1.66), Eastern–Ashanti–Ahafo–Bono East (2020–2021, RR = 1.28), and Central–Bono (2018, RR = 1.21). Bivariate analysis showed a significant negative spatial autocorrelation between Gross National Income (GNI) and HIV incidence (Moran’s I = –0.341, p=0.034) but no significant associations with tuberculosis prevalence or the Consumer Price Index (CPI). BiLISA revealed high HIV burden in southern regions. Gradient Boosting identified GNI as the most influential predictor of HIV risk, followed by tuberculosis prevalence, while CPI had minimal influence. Conclusion: This study highlights significant spatial and temporal heterogeneity in HIV incidence across Ghana, with hotspots concentrated in the southern and middle belt regions. Gross National Income emerged as a key socioeconomic determinant negatively associated with HIV burden, underscoring economic disparities as critical drivers of regional risk. The identification of distinct space–time clusters and spatial mismatches with other determinants like tuberculosis prevalence and inflation suggests the need for tailored, region-specific interventions. Integrating spatial epidemiology with advanced machine learning offers valuable insights for targeted public health strategies to effectively reduce HIV transmission in Ghana.