An integrated statistical model of Emergency Department length of stay informed by Resilient Health Care principles. (December 2019)

• Record Type: Journal Article
• Authors: Ross, Alastair J.; Murrells, Trevor; Kirby, Tom; Jaye, Peter; Anderson, Janet E.
• Is Part Of: Safety science. Volume 120(2019)
• Page Start: 129
• Page End: 136
• Publication Date: 2019-12
• Subjects: Resilient Health Care -- Data systems -- Emergency department -- Patient flow -- Informatics; Industrial accidents -- Periodicals; Accident Prevention -- Periodicals; Safety -- Periodicals; Travail -- Accidents -- Périodiques; 363.11
• DOI: 10.1016/j.ssci.2019.06.018
• Languages: English
• ISSNs: 0925-7535

Abstract

Highlights: Emergency departments face variable demand and capacity pressures. Resilient Health Care studies how outcomes emerge from complex interactions. An integrated dataset to support resilient performance is feasible. A core set of early indicators of length of stay should be part of designed improvement efforts. Abstract: Background: Hospital Emergency Departments (EDs) face variable demand and capacity issues affecting timely discharge of patients. This is due in part to a lack of integration of routine monitoring data, affecting anticipation and response. Methods: Patient flow was modelled (four hour target breaches; time to decision-to-admit; subsequent time to admit-to-hospital) in a busy ED. Patient and organisational data were collated, screened and conceptualised using Resilient Health Care (RHC) theory. Data were collected for all patients presenting during a 24-month period (May 2014–April 2016; n = 232, 920) and analysed via multivariable logistic regression for four hour target breaches, and ordinary least squares regression for time. A measure of effect size was calculated for each independent variable. Overall model fit was assessed using percent concordant. Results: Length of stay is related to demand, capacity and process indicators including: number of patients; night shift; first location being resuscitation or major injury area(s); urgent or very urgent triage patients; patients readmitting from up to 7 days previous; bed capacity; recent ambulance arrivals; and patients where the primary presenting complaint (PPC) is related to mental health or difficult to ascertain. Conclusions: Understanding variation in performance through RHC theory can support staff and organisations in monitoring, anticipating and responding. A set of reliable core predictors has been identified to help design future ways to facilitate resilient performance through early indicators of pressure.