Risk Factors for Dengue Fever in Thailand Using Mathematical Generalized Estimating Equation (GEE) - ปัจจัยเสี่ยงของกาารเกิดไข้เลือดออกในประเทศไทยโดยใช้ตัวแบบเชิงคณิตศาสตร์ Generalized Estimating Equation (GEE)

Abstract

The objective of this research is to investigate risk factors for Dengue fever in Thailand. The
mathematical generalized estimating equation (GEE) whose dependent variables have Poisson
distributions and negative binomial distributions are studied. The secondary province-level data
consisting of the number of Dengue fever patients, rainfall, average temperature, forest area, household
income are collected. The regions and seasons are also considered. The research finds that the
GEE whose dependent variables have negative binomial distributions are more appropriate, hence it
is used for the data analysis. The central region and the season during February-April are assigned to
be reference groups. At the significance of 0.05, the risk factors for Dengue fever are rainfall (Relative
Risk (RR) 1.0009), average temperature (RR 1.1736), forest area (RR 1.0482), southern region
(RR 2.7390), eastern region (RR 2.0489), western region (RR 1.4051), the season during May-July
(RR 3.0526) and August-October (RR 3.2677).
Key words: Dengue fever, Generalized Estimating Equation (GEE), Poisson, Negative binomial