Time-series analysis under quasi-Poisson distribution was applied. Two statistical models were used to analyze the estimated flood effect on diarrhoeal disease: (1) a model adjusting for seasonality and long-term trends (the simplest) and (2) a model adding adjustment for mean temperature to the simplest model. The final model is given as follows: