Comparison of different methods to forecast hospital bed needs

Objectives: This study compares three methods to forecast the number of acute somatic hospital beds needed in a Swiss academic hospital over the period 2010-2030. Design: Information about inpatient stays is provided through a yearly mandatory reporting of Swiss hospitals, containing anonymized data. Forecast of the numbers of beds needed compares a basic scenario relying on population projections with two other methods in use in our country that integrate additional hypotheses on future trends in admission rates and length of stay (LOS). Results: The total number of beds in the hospital under study is projected to increase from 960 beds in 2010 to a range of 1188 to 1332 in 2030. This estimated growth equals to 24% to 39% of the hospital size at baseline, and these different figures are largely due to hypotheses about future reduction in LOS. Conversely, as the models' hypotheses regarding future demand have balanced effects on global admission rates, the three methods lead to similar projections for the number of stays, which is estimated to increase by 33% between 2010 and 2030. Conclusion: In the context of population aging, forecasting the number of hospital beds should not rely on a unique scenario. Using different hypotheses about admission rates and LOS, and comparing projected versus real use are important to avoid shortage of beds. (C) 2015 Published by Elsevier Masson SAS.