‘Length of stay’ (LOS) is a commonly used measure of performance of hospital in-patient services.
Most adult acute mental health admissions range from a few days to a few weeks. However, some patient groups require longer LOS in hospitals. The ability to monitor length of stay could give valuable information about the differential costs of services between client groups; differential performance of staff and teams; and differential patient need / complexity by geographical area. However, service activity data rarely distinguish between the type of admission or complexity of patient need. Comparisons between Trusts also need to take into account the differences in case-mix.
This project aims to develop a more accurate method to measure and analyse LOS trends for specialist mental health admissions using data from the South Essex Partnership University NHS Foundation Trust (SEPT). Currently a variety of methods are used to calculate LOS for mental health patients, but these are rarely specified. Methods are constrained by the structure of the dataset and comparison of measures may not be valid.
Our primary objective of the project is to compare different methods of LOS analysis for stratified care groups and to examine LOS trends. Our secondary objective is to design a standardised format to be used by the Trust as one of the clinical performance dashboards.
The patient population refers to patients admitted to specialist mental health services commissioned by the South Essex Partnership University NHS Foundation Trust. The primary analysis of the LOS data has been stratified according to mental health diagnosis. This might be further stratified by age, gender and/or bed type. We have compared the following methods to estimate LOS: mean and median, a measure based on total occupied bed days, survival function and a hospital age analysis. Studies have shown that diagnosis, even when clearly defined, does not adequately predict length of in-patient stay. Additional analyses will thus be undertaken to understand the possible predictors of length of stay.
Analysis of this large dataset using different methods shed some light on the reasons behind variation on LOS across sites and diagnostic, namely by adjusting for a different care pathway and hence case mix. The results were presented to the Executive team with recommendations that future performance data be based on this more comprehensive approach to LOS analysis. Also, a working group has been set up to establish a program to reduce LOS. We are preparing peer-review publications and monitoring the intermediate outcome of this project.
This project is a collaboration between the Public Health Theme of CLAHRC CP, Dr Mel Conway, Consultant in Public Health Medicine (South Essex Partnership Trust) and the East of England Public Health Observatory (ERPHO).
For more information contact: Dr Louise Lafortune, Cambridge Institute of Public Health, University of Cambridge, email@example.com