Quantitative methods for evaluating hospital design. A stochastic model for interdepartment traffic.

A stochastic model is developed to describe the traffic among ten hospital departments or functional centers that generate 75 percent of all personnel trips to and from two nursing units with a total of 120 beds. The effects of five factors on travel frequencies are investigated by analysis of variance, as are the interactions among the factors. Multiple regression and correlation analyses are used to test the feasibility of predicting travel frequencies from the patient census in three levels of care and from the size of the nursing unit staff, which are found not to be very useful predictors. A Poisson distribution is found to be a good descriptor of the observed frequency distributions for seven of the ten departments studied.