On the Computational Bayesian Survival Spatial DHF Modelling with CAR Frailty

Dengue Hemorrhagic Fever (DHF) is one of the most common diseases experienced in Indonesia, especially in the rainy season. DHF is caused by the dengue virus wlijich is transmitted by female Aedes aegypti mosquitoes. The recovery rate of DHF patients is influenced by several factors. In this study, we are modeling how these factors can influence the recovery rate for DHF patients. Thus, we can provide advice to the hospital. Survival analysis is a method used to model survival data. In this case, the survival model will be used to find the relationship between the recovery rate of DHF patients and factors that can affect it. The environment of patient residence is one of the factors that need to be considered. Tins case required analysis by considermg the spatial effect of the proximity between one sub-district and another. The addition of a random effect can be done to avoid bias on estimation by using conditionally autoregressive (CAR) prior. The result showed that DHF in eastern Surabaya mostly occurred in Tambaksari and Gubeng sub-districts. The mean of survival time of patients is 4 days. In this study, the Weibull and Generalized Gamma distributions were used to modelmg the patient recovery∗ rate. In this study, there are spatial autocorrelations cases which are measured with the Moran's I index. The Weibull distribution with addition frailty using Normal CAR is the best model that can be used to modeling the recovery rate of DHF patients.