Spatial correlation and hot spot analysis of dengue hemorrhagic fever in Thailand

Abstract

Background: Dengue hemorrhagic fever (DHF) is an important public health problem found in all regions of Thailand and there are many cases, including deaths, reported every year. Since several spatial characteristics are determinants of DHF epidemic, objectives of this study are to explore spatial relationship of the disease in the country and to identify areas at high risk of dengue epidemic.
Methods: Spatial autocorrelation analyses were performed using data of provincial DHF incidence of Thailand from the National Notifiable Diseases Surveillance (Report 506) during 2014-2016.
Results: Overall, DHF incidences at provincial level in Thailand had statistically significant spatial correlation in 2014, 2015 and 2016 with Moran’s I coefficients of 0.38, 0.41 and 0.51, respectively. Clusters of DHF epidemic were identified and varied by geographical regions. Among provinces in the most south region, clusters of epidemic were found in all three years including two years with “hot spot”, i.e. aggregation of high incidence provinces. In the central region, hot spot was identified in one of three years. Additionally, there were some provinces outside the hot spots but had high risk of either importing or exporting epidemics from adjacent provinces, based on their statistically significance level.
Conclusion and discussion: Findings of this study indicated that DHF epidemic in Thailand had clear pattern of hot spot in some regions while, in other areas, provinces at high risk of exporting or importing epidemic were also identified. We strongly recommend relevant organizations to consider applying spatial autocorrelation analysis, in addition to conventional methods, to improve provision of appropriate disease control measures.