Discrete conditional survival models for trolley waiting times in accident and emergency

Adele H. Marshall, Louise Burns

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Health care managers have a huge responsibility to appropriately allocate the limited resources available. As a result focus is on intelligent patient management models using approaches from operational research and statistics. This paper wishes to consider one such approach for modeling an Accident and Emergency (A&E) department in a UK hospital. The main objective is to develop a model that can accurately predict which patients will experience a 'trolley wait' from the time the doctor decides they should be admitted to hospital until the time they are allocated a hospital bed. The Discrete Conditional Survival Model is developed using Naïve Bayes Classification to predict the patient admission and a survival model represented by a lognormal distribution to model the trolley waiting time. Based on patient information available on arrival to hospital, the model can predict the patients likely to experience trolley waits and plan ahead to prevent such cases happening.

Original languageEnglish
Title of host publication2010 IEEE Workshop on Health Care Management, WHCM 2010
DOIs
Publication statusPublished - 24 May 2010
Event2010 IEEE Workshop on Health Care Management, WHCM 2010 - Venice, Italy
Duration: 18 Feb 201020 Feb 2010

Publication series

Name2010 IEEE Workshop on Health Care Management, WHCM 2010

Conference

Conference2010 IEEE Workshop on Health Care Management, WHCM 2010
CountryItaly
CityVenice
Period18/02/201020/02/2010

Keywords

  • Discrete conditional survival models
  • Trolley waits

ASJC Scopus subject areas

  • Public Health, Environmental and Occupational Health
  • Health(social science)

Fingerprint Dive into the research topics of 'Discrete conditional survival models for trolley waiting times in accident and emergency'. Together they form a unique fingerprint.

Cite this