The use of demographic characteristics in predicting length of stay in a state mental hospital.

Publisher Website
Google Scholar
Abstract
This study was concerned with investigating the feasibility of using routinely collected demographic data for predicting length of hospitalization of patients admitted to state mental hospitals. The study material consisted of 13,731 patients admitted to the state-supported mental hospitals of Oklahoma, USA. The model utilized was Bayes'' theorem. This technique makes possible the calculation of the probability of a patient''s remaining in a hospital a certain length of time given that he possesses certain demographic characteristics. Bayes'' theorem predicted length of hospitalization with varying degrees of accuracy depending on the construction of the various length of hospitalization categories. The method predicted with an over-all accuracy of 86% when the length of hospitalization categories were less than 30 days and 30 days or more. This was the best effort. These figures compare favorably with those of other prediction studies. The advantages of this method are the use of data already being routinely collected, and the speed with which the results can be obtained as a result of using data available at the time of the patient''s admission.