This paper first reviews the forecasting literature relevant to forecasting the demand for hospital services. It then reports on a survey of forecasting practices at 40 hospitals. The survey was conducted by interviewing persons at upper levels of management in late 1987 and early 1988. Forecasting of aggregate demand was seen as a significant area of concern by 2/3 of the hospitals, and a critical area by 27.5%. However, actual practices did not reflect this high level of priority; only 12.5% of the hospitals treated it as a critical area. Subjective methods were most commonly used: 87% reported using the jury of executive opinion and 67.5% used the crystal ball. This is a much higher usage of subjective methods than was reported for industrial and consumer firms by Dalrymple (1987). Given the political nature of hospital forecasts, the use of subjective forecasts would seem to present serious problems with respect to bias.
For the quantitative methods, 35% of the hospitals used regression, 52.5% used moving averages, and only 12.5% used exponential smoothing. Of those hospitals that had a computerized forecasting system (69% of the sample), most (56%) were not satisfied. Despite numerous legal and practical reasons for collecting data, over half of the hospitals reported that they did not have adequate data to prepare their demand forecasts.