Revisiting physician impact analyses: predicted versus observed hospital resource use

In order to determine the accuracy of predictions made using Physician Impact Analysis (PIA), a comparison was undertaken of predicted versus actual resource use for 10 new physician specialists at a southern Ontario community hospital. The predictions were done from 1987 to 1989 using methodology recommended by the Ontario Hospital Association (OHA) and available at that time. This included (1) Hospital Medical Records Institute data and (2) a hospital service department survey. A comparison was made between PIA predictions and actual physician resource utilization data gathered in 1991. There was little agreement between the two. The usefulness of PIA rests largely on its ability to generate accurate predictions about a new physician recruit's number of cases and hospital resource use. However, because this research was undertaken in only one community hospital, further evaluation of the PIA process is recommended.