Spatial clustering and risk factors for late HIV diagnosis in the Yinzhou, China: a population-based retrospective cohort study

Background Late HIV diagnosis is a critical challenge for controlling the transmission; if not addressed, it could inevitably slow or reverse current progress toward the UNAIDS 95-95-95 targets by 2030. Methods We conduct a retrospective study of a full registry of 1404 HIV cases in Yinzhou District (2001–2025), China. Adjusted relative ratio (aRR), posterior relative ratio (RR), Global Moran's I index, and adjusted hazard ratios (aHR) of late HIV diagnosis were calculated using Poisson's regression analysis, subgroup analysis, spatial autocorrelation analysis, and Cox regression, respectively. Results The proportion of late HIV diagnosis was 76.4% in Yinzhou. Older age (aRR = 1.003, 95%CI: 1.001–1.006) at HIV diagnosis was significantly associated with a higher risk of late diagnosis. Patients who tested voluntarily (aRR = 0.918, 95%CI: 0.849–0.993) and living in the northwest (aRR = 0.933, 95%CI: 0.876–0.994) were associated with a lower risk. The northwest was a cold spot, and the central and southeast areas were hotspots for late HIV diagnosis (global Moran's I = 0.65, p < 0.001). Patients aged 60 and over and tested non-voluntarily was identified as a priority subgroup (posterior RR = 1.097, 95%PI: 1.013–1.174). Late presenters had a higher hazard of death than non-late presenters (aHR = 3.03, 95%CI: 1.04–8.84). Conclusion We identified a priority subgroup at high absolute risk and a hotspot for late HIV diagnosis in Yinzhou. This finding provides evidence for developing targeted intervention strategies.