Patient flow modelling based on the major trauma network in South Wales

Wang, Zihao ORCID: https://orcid.org/0009-0008-0008-5813 2024. Patient flow modelling based on the major trauma network in South Wales. PhD Thesis, Cardiff University. Item availability restricted.

	PDF - Accepted Post-Print Version Restricted to Repository staff only until 13 March 2026 due to copyright restrictions. Download (18MB)
	PDF (Cardiff University Electronic Publication Form) - Supplemental Material Restricted to Repository staff only Download (122kB)

Abstract

This thesis investigates variations in patient flow and quantifies operational performance within the newly established South Wales Trauma Network (SWTN) by focusing on the intermediate outcome of length of stay (LOS). A systematic literature review uncovered a significant research gap in evaluating LOS across trauma networks, while a scoping review of LOS modelling techniques highlighted the right-skewed distribution of LOS data, providing insights into developing appropriate analytical models. Utilising a comprehensive secondary dataset from the trauma registry, which includes 26,238 admissions from January 2012 to August 2021, this research began with a robust data quality assessment. Advanced data visualisation techniques were employed to examine variations in patient flow before and after the establishment of the SWTN and to identify critical factors influencing LOS. Three sophisticated statistical models—Lasso regression, random forest, and generalised additive model (GAM)—were meticulously developed to effectively address the skewed nature of LOS, aiming to pinpoint key predictors of LOS within the SWTN. A thorough comparative analysis demonstrated that Lasso regression most effectively captured the characteristics of the data, revealing the association between relevant factors and LOS. Notably, the Lasso model’s variable importance features highlighted the top 30 factors, such as patient age, specific ward types, and their interactions with hospitalisation frequency, transfer status, and higher Maximum Abstract Abbreviated Injury Scale (MAIS) spine scores. These insights are invaluable for trauma system managers making critical decisions about bed allocation, staffing, and discharge rehabilitation planning. This research significantly advances the theoretical understanding by identifying a comprehensive set of evaluation indicators that integrate operations management principles with healthcare performance metrics, thereby extending prevailing theories to trauma-specific contexts. Novel predictors of LOS, including specific ward types and their interactions with ward transfers, provide enhanced insights into patient flow and resource allocation, contributing to improved trauma care management and patient outcomes.

Item Type:	Thesis (PhD)
Date Type:	Completion
Status:	Unpublished
Schools:	Schools > Business (Including Economics)
Date of First Compliant Deposit:	18 March 2025
Date of Acceptance:	13 March 2025
Last Modified:	18 Mar 2025 10:06
URI:	https://orca.cardiff.ac.uk/id/eprint/176854

Actions (repository staff only)

Edit Item