Spatio-temporal climate-based model of dengue infection in Southern, Thailand

This study explored the spatio-temporal patterns of dengue infection in southern Thailand. Data on monthly-notified cases of dengue fever, over the period of January 1981- May 2014 were collected from the Bureau of Epidemiology, Department of Disease Control, Ministry of Public Health. Weather data over the period of January 2007-May 2014 were obtained from the Thai Meteorological Department. Box and whisker plots were used to study the spatial temporal patterns of dengue incidence. Spearman correlation analysis and time-series adjusted Poisson regression analysis were performed to quantify the relationship between weather and the number of dengue cases. The results show that the highest dengue cases occurred in July in the Gulf of Thailand. Conversely, for the Andaman Sea, the highest dengue cases occurred in June in Phang-Nga, Phuket and Ranong provinces. Only Krabi province had the most dengue cases occurring in July. When we compared dengue transmission duration between the Andaman Sea and the Gulf of Thailand, we found that for the Andaman Sea, Trang province had the longest dengue transmission duration (i.e. June-September) and for the Gulf of Thailand, Nakhon Si Thammarat had the longest dengue transmission duration (i.e. June-September). The number of rainy days and relative humidity were the main predictors of dengue incidence by the Gulf of Thailand, while the amount of rainfall and the temperature were the main predictors of dengue incidence by the Andaman Sea. The time series Poisson regression models provided such goodness-of-fit that the correlation between observed and predicted numbers of dengue incidence exceeded 80%. These models could be used to optimise dengue prevention by predicting trends in dengue incidence. Accurate predictions, for even a few months, provide an invaluable opportunity to mount a vector control intervention or to prepare for hospital demand in the community.