EURO 2025 Leeds
Abstract Submission

2476. Clinical Pathway Modelling of a Trauma & Orthopaedics Department

Invited abstract in session TD-13: Machine learning in healthcare, stream OR in Healthcare (ORAHS).

Tuesday, 14:30-16:00
Room: Clarendon SR 1.01

Authors (first author is the speaker)

1. Matthew Howells
School of Mathematics, Cardiff University
2. Paul Harper
School of Mathematics, Cardiff University
3. Daniel Gartner
School of Mathematics, Cardiff University
4. Geraint Palmer
School of Mathematics, Cardiff University
5. Antonio Riccioli
Cardiff and Vale University Health Board

Abstract

Trauma and Orthopaedic (T&O) departments face dual pressures from backlogs caused by the COVID-19 pandemic and ageing populations, leading to increased frailty and greater demand for services. Growing waitlists already negatively impact patients and will continue to do so without intervention.
Clinical pathways (CPs) provide structured processes for standardising essential steps in managing a health condition or medical procedure. This research implements the ALERGIA algorithm in Python to generalise CPs from a T&O patient dataset in the UK’s National Health Service. This allows us to infer potential future CPs not present in the dataset while filtering out statistically insignificant pathways.
We further demonstrate how these CPs inform a novel approach to parameterising a discrete-event simulation (DES) of a holistic T&O surgical pathway. This ensures the model remains adaptable rather than strictly tied to historical data while refining routing probabilities by reducing noise. The DES model is then integrated with a system dynamics model that assesses the burden of orthopaedic conditions in the population and how patients deteriorate over time while awaiting a referral from their General Practitioner.
This approach enables us to assess how changes in demand and capacity, both inside and outside the hospital, impact patient flow and health outcomes. It provides hospital staff with better decision-making tools to alleviate treatment backlogs and improve patient care.

Keywords

Status: accepted


Back to the list of papers