Prediction of Discharge from a Psychogeriatric Unit: (Development and Evaluation of the LPRS Prognosis Index)

Abstract

Data based on the 36-item London Psychogeriatric Rating Scale was statistically analysed in order to select a small number of questions which would discriminate well between those psychogeriatric inpatients who were likely to be discharged in the following six months and those who were not. A very brief Prognosis Index was developed utilizing the five best discriminating items of the LPRS. The content of these indicated that patients who could communicate coherently, made friends and did not threaten others were those most likely to leave hospital. The predictive accuracy of the index was assessed for the inpatient populations of a 136-bed psychogeriatric unit at three different points in time. The index was found to correctly predict whether or not patients would be discharged within six months in 78–85% of all cases. Correct prediction was generally seen to increase with time. Predictive accuracy reached as high as 96% (25/26) in the sample of acute ward patients examined after eighteen months. Since discharge was seen to be dependent on the availability of appropriate placements, the predictive accuracy of the Prognosis Index might be seen as surprisingly high. The simplicity and predictive accuracy of the items utilized and the fact that administration and scoring of the index can be done in less than five minutes without cooperation on the part of the patient suggest that the Prognosis Index is a useful measure. The results also emphasize the importance of personal relationships in determining discharge from hospital.