Proceedings of the 2010 International Conference on Industrial Engineering and Operations Management Dhaka, Bangladesh, January 9-10, 2010 242 Models for Bed Occupancy Management of a Hospital in Singapore Arun Kumar and John Mo School of Aerospace, Mechanical and Manufacturing Engineering RMIT University, Melbourne, Victoria 3000, Australia Abstract This paper describes three bed prediction models in aiding hospital planners to anticipate bed demand so as to manage resources efficiently. The Poisson bed occupancy model provides an estimation of bed occupancy and optimal bed requirements in each class based on length of stay and admissions data. The simulation model was developed to predict bed occupancy levels for every class for the following week utilizing historical previous year’s same week admissions data. Regression equations were formulated based on relationship between identified variables to aid bed managers to predict the weekly average number of occupied beds. Keywords Prediction, bed occupancy, simulation, regression, Poisson model 1. Introduction The ability to anticipate the future demand of services would enable an organization to be able to plan effectively. For the past few decades, healthcare providers worldwide have thus been under tremendous pressures and obtaining demerit for failing to plan effectively. Over-crowding is perhaps the most common scene that people see in the Emergency department of a Hospital. The random arrival pattern of patients may somehow mislead healthcare planners, and thus causing them to underestimate the resources that are required within the hospital. Misinterpretation of patients’ arrival patterns is the reason for poor bed management. Hence the ability to forecast the random arrival patterns would definitely be the key solution to over-crowding problems. Many have seen over-crowding particularly in emergency departments to be an intensifying nemesis. All the local public hospitals in Singapore have been constantly facing problems such as bed shortages during peak period of the year. Due to a rise in elderly pneumonia patients, some of the hospitals were being forced to redirect non-critical cases to public hospitals nearby. Additionally, the A&E department of some of the hospitals had been forced to deny access to ambulances during certain hours of the day [1]. This deficiency in the healthcare sector has drawn public awareness. There are rising concerns of over occupancy in various local hospitals such as Changi Hospital, Singapore General Hospital and Tan Seng Hospital. In 2007, when untimely flu season hit the region, various hospitals reached their peak occupancy. The over occupancy problem was so serious that there was not even a single bed available in any class. Bed shortages or inadequacy had an effect on the following: 1. Long waiting time for incoming patients at Emergency Department (ED). 2. Patients from Tan Seng Hospital (TSH) to transfer to another hospital when there was overflow by 85-90%. 3. Cross-discipline allocation was the process whereby a particular discipline had no unoccupied bed; hence, the patient rightfully admitted into a particular discipline was channelled to another discipline. This method of cross allocating had the disadvantage of untrained staff for that discipline. 4. Up-lodging was the last resort, when the beds in other discipline were fully occupied. This solution however posed financial deficits for the hospital. Incoming patients from ED wards that had opted for Class C were channelled to a higher Class (A, B1 or B2), when the waiting time exceeded 2 hours, and with no available beds in other discipline. 5. Public lost confidence in the ability of the hospital in aspect of bed management. Thus, effective planning and management of supply and demand for hospital beds is fundamental for hospitals to deliver high medical standard commitments, including reducing the bed wait time. The objective of this research is to compare the bed forecasts obtained from various mathematical models and recommend the efficient forecast that enable hospital administrators to plan their resources within the hospital.