Comparison of statistical methods to predict the time to complete a seriesof surgical cases

We present a statistical model for predicting the time to complete a series of successive, elective surgical cases. The use of sample means of case ti mes and turnover times when scheduling cases does not minimize the operatin g room labor costs associated with errors in predicting times to complete s eries of cases. The problem of minimizing associated labor costs (both unde r and over utilization) can be converted to the problem of least absolute d eviation regression. The dependent variables are the times to complete seri es of cases. The independent variables are the numbers of cases in each ser ies that are in various categories (i.e., combinations of scheduled procedu res and surgeons). Although the computational method is preferred on theore tical grounds to that involving sample means, application of both methods s hows that the more practical method is to use the sample means of previous case times and turnovers.