Statistical approaches for identifying heavy users of inpatient mental health services

Abstract

Background: A lack of consensus exists concerning how to identify “heavy users” of inpatient mental health services.

Aim: To identify a statistical approach that captures, in a clinically meaningful way, “heavy” users of inpatient services using number of admissions and total time spent in hospital.

Methods: “Simple” statistical methods (e.g. top 2%) and data driven methods (e.g. the Poisson mixture distribution) were applied to admissions made to adult acute services of a London mental health trust.

Results: The Poisson mixture distribution distinguished “frequent users” of inpatient services, defined as having 4 + admissions in the study period. It also distinguished “high users” of inpatient services, defined as having 52 + occupied bed days. Together “frequent” and “high” users were classified as “heavy users”.

Conclusions: Data driven criteria such as the Poisson mixture distribution can identify “heavy” users of inpatient services. The needs of this group require particular attention.

Keywords:

• Inpatient services
• statistical methodology
• heavy use
• length of stay
• readmission

Acknowledgements

The authors would like to thank the staff at the BRC Nucleus for their contribution to the data extract. Thanks also to Professor Til Wykes, Vice-Dean of Psychology and Systems Sciences and Professor of Clinical Psychology and Rehabilitation.

Declaration of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this article.

CRIS is supported by the NIHR Biomedical Research Centre for Mental Health BRC Nucleus at the South London and Maudsley NHS Foundation Trust and Institute of Psychiatry, King's College London jointly funded by the Guy's and St Thomas' Trustees and the South London and Maudsley Trustees.