Singapore: Hospital capacity prediction by fuzzy linear regression

Strategic planning for hospital inpatient capacity is important in providing quality healthcare service at a reasonable cost. This study estimates annual average hospital length of stay (LOS) as an aggregate quantity for all private and public hospitals in Singapore. Because LOS depends on a patient's medical history as well as on the severity and type of illness, it is a random variable that involves a wide spread. Hence, providing an accurate estimate for LOS poses a challenge but is very valuable. In this article, a multivariate fuzzy linear regression (FLR) model is developed for predicting aggregate annual LOS. The results show that FLR can provide accurate point and interval forecasts, although the amount of training data is small and the spread is wide.