Abstract

The government hospital in Malaysia is prominent for the low cost of health care and medical treatment, and it had been reported that the hospital admission is increasing annually. This has led to the widespread problem of overcrowding. In order to assist the government to plan and manage demands for health services and health care needs, the prediction for future admission is outlined. The model is fitted to the matrix of admission rates, spanning the period of historical data from 2001 to 2011 for estimation. In particular, there are three types of estimation approaches used to generate the parameters of the Lee-Carter model which are singular value decomposition, iterative Newton–Raphson method and Poisson maximum likelihood estimation. The resulting estimation generated by these three estimation approaches is then being compared and evaluated by two types of error measures which are mean squared error and mean absolute percentage error. The smallest values of MSE and MAPE indicate the better performance of the estimated parameters. The best estimation approach is subsequently used to forecast the admission rates of HRPZ II, Kota Bharu, for ten years ahead; using the out samples data from period 2012–2015. At the end of the study, it can be generally concluded that the occurrence of admission to HRPZ II, Kota Bharu, for both genders in each of the all broad age groups will slowly decrease along with the increment of years.