Spatial Autocorrelation and High-Risk Area Identification of Food Poisoning in Thailand, 2003–2022

Abstract

Food poisoning represents a persistent public health burden in Thailand, yet provincial-level spatial clustering patterns have not been comprehensively characterized over extended time series. This retrospective analytical study utilized foodborne illness incidence data from the national disease surveillance system (Report 506) of the Department of Disease Control, Ministry of Public Health, covering all 77 provinces for the period 2003–2022 (n = 2,329,463 reported cases). Provincial incidence rates (per 100,000 population) were computed annually and linked to administrative boundary polygon data. Spatial autocorrelation was assessed using the Global Moran's I statistic, and spatial cluster analysis was performed using the Local Indicators of Spatial Association (LISA) with Queen contiguity first-order spatial weights in GeoDa (version 1.14.0). Global Moran's I ranged from 0.317 to 0.522 across all study years (all p < 0.05), indicating statistically significant positive spatial autocorrelation. High–High (H–H) clusters were consistently identified in the northeastern and northern regions, with the northeastern region demonstrating 9–14 provinces per year classified. Low–Low clusters predominated in the southern region throughout the study period. The spatial clustering patterns persisted across five-year sub-periods, suggesting stable geographically determined risk factors. Findings indicate that food poisoning in Thailand exhibits non-random, geographically structured distribution attributable to inter-related socioeconomic, dietary, and food-market environmental factors. These results provide evidence for spatially targeted surveillance and intervention strategies.