Modeling Dengue Epidemics in Selected Areas of Thailand, 1996–2008

• Main Authors: Mathuros Tipayamongkholgul, 田蜜
• Other Authors: Chi-Tai Fang
• Format: Others
• Language: en_US
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/97160425367396403402

博士 === 臺灣大學 === 流行病學研究所 === 98 === Dengue fever (DF) and Dengue hemorrhagic fever (DHF) have become a major concern in Thailand since 1950s. Despite intensive efforts in vector control, dengue epidemics continued to occur throughout this country in multi-annual cycles. Weather is considered to be the important factor in these cycles, but the extent to which the El Niño-Southern Oscillation (ENSO) is a driving force of dengue epidemics remains unclear. As well as, an increasing trend of incidence of dengue cases in Thailand have conjectured the effect of urbanization. The impacts of urban contexts on local dengue vulnerability have not been well investigated. This dissertation examined the temporality between El Niño and the occurrence of dengue epidemics, constructed Poisson autoregressive models for incidences of dengue cases, and applied geospatial model to estimate the effect of socio-geographic factors and transmission intensity of dengue on the incidence of all reported dengue cases. Dengue surveillance data, climate data (global and local climate), and socio-geographic factors were analyzed. The strength of El Niño consistently positively correlated with the occurrence of dengue epidemics throughout time lag from 1 to 11 months in two selected regions of Thailand. Up to 22% (northern inland mountainous region) and 15% (southern tropical coastal region) of variations in the monthly incidence of dengue cases were attributable to global ENSO cycles. Province-level predictive models were fitted using 1996–2004 data and validated with out-of-fit data from 2005. Multivariate ENSO index remained an independent predictor in 10 of the 13 studied provinces. Analysis showed a significant neighborhood effect (ρ= 0.405, P<0.001), which implies that villages with geographical proximity shared a similar level of vulnerability to dengue. The two independent social factors associated with a higher incidence of dengue were a shorter distance to the nearest urban area (β = –0.133, P<0.05) and a smaller average family size (β = –0.102, P<0.05). Spatial error regression and Geographically Weight Regression (GWR) consistently presented positive effect of transmission intensity and spatial dependence upon the proximity villages. Spatial heterogeneity was clearly presented by GWR; the distinct patterns of transmission intensity across regions were relocated yearly. It also clearly expressed the higher intense transmission in north region than south region. In conclusion, El Niño was one of the important driving forces of dengue epidemics across geographically diverse regions in Thailand and indicated that the trend of increased dengue occurrence in rural Thailand arose in areas under stronger urban influence rather than in remote rural areas. Population mobility may increase transmission intensity urban area and neighboring area.