Prediction of dengue outbreaks in Kerala state using disease surveillance and meteorological data

Abstract

Background: Dengue is one of the most serious and fast emerging tropical diseases. Its incidence of is influenced by many meteorological factors such as rain fall in mm, temperature, humidity etc. Information about these factors can be used to forecast the incidence of dengue fever cases in the next coming months.Methods: The current study was an analytical study using retrospective secondary data from Kerala state. The annual integrated disease surveillance reports of dengue fever cases. Rain fall data and mean monthly temperatures for a period of twelve years from 2006 to 2017 were used. Best fitted model was developed and accuracy of the prediction was tested. All analyses were performed in R software using the mgcv package and nlme package.Results: A total of 144 months study period from January 2006 to December 2017 was used for analysis. Five different models developed for prediction of dengue cases among them, best fitted model including optimal combination of meteorological variables and recent and long term transition of dengue was selected. Out of 84 months predictions in the training period, 68 months prediction was correctly negative, 5 months prediction was correctly positive, 2 months prediction was incorrectly negative and 9 months prediction was incorrectly positive.Conclusions: A better predictive generalized additive model can be developed using the optimal combination of meteorological predictors and dengue fever counts. It will enable the health care administrators to forecast future out breaks and to take better precautionary measures.