The objective of this study was to find the appropriate forecasting model for monthly
patient admissions and discharges. A data set on monthly patient admissions and discharges was
collected from the Rajavithi Hospital, consisting of 72 monthly observations from October 1998 to
September 2004. The analyses were done in four steps. First, the intervention analysis was conducted
to model the historical data taking into account a new intervention policy of the universal of
healthcare project (UHP), which started on October 1, 2001. Second, the Box-Jenkins (BJ) method
was applied to develop the forecasting models using the past data of 36, 48, 60 and 72 monthly series
that ran backward from September 2004 for each individual series of monthly patient admissions and
discharges. Third, The forecasting performances of the BJ models were evaluated using the mean
absolute deviation (MAD), the mean square error (MSE) and the mean absolute percentage Error
(MAPE). The appropriate forecasting model was considered from the minimum values of these
indices. Fourth, the chosen model was used to make forecast for six months ahead of monthly patient
admissions and discharges. Then, the forecast values were compared with the actual values for the
period of October 2004 to March 2005. Results showed that the BJ approach was a more appropriate
way for both monthly patient admissions and discharges using the 60 monthly historical data.
However, the accuracy of the forecasting depended upon the variation within the data. . . .