Prediction of Discharge of Inpatients

This paper describes a survey in which 273 surgical patients at St. Thomas' Hospital, London, were studied to ascertain the extent to which their discharge could be predicted. Initial length-of-stay estimates were recorded for 57 percent of the cases; the excess of actual over predicted length...

Full description

Saved in:
Bibliographic Details
Published inHealth services research Vol. 3; no. 3; pp. 174 - 184
Main Authors Bithell, John F., Devlin, H. Brendan
Format Journal Article
LanguageEnglish
Published 01.01.1968
Online AccessGet full text

Cover

More Information
Summary:This paper describes a survey in which 273 surgical patients at St. Thomas' Hospital, London, were studied to ascertain the extent to which their discharge could be predicted. Initial length-of-stay estimates were recorded for 57 percent of the cases; the excess of actual over predicted length of stay had a mean of 2.71 and a standard deviation of 10.2 days. The length of stay of the other 43 percent had greater variability. Continuous predictions of discharge were made, in which discharge dates were regularly estimated and assigned degrees of certainty, which were also recorded. The information obtained represented a 31.5 percent reduction in the squared prediction error of the number of discharges within three days over an estimate employing only a count of patients under review. The implications of the findings are discussed.
ISSN:0017-9124