Healthcare Process Analytics for Care Pathways workshop

Workshop lead

Owen Johnson, School of Computing, University of Leeds

Rationale

Care pathway improvement is a key focus of attention for clinicians, health service providers, commissioners and governments. Care pathways are a type of business process with properties that are unique to healthcare – the patient is simultaneously subject, participant and customer while care is dynamic, complex, ad hoc and multi-disciplinary. In other industries business process analytics has been well established for many decades and approaches tend to combine four distinct fields – process modelling, process simulation, workflow support and, more recently, process mining. Process mining is based on an emerging set of data mining technologies focussed on temporal data features which can be used to reconstruct traces of individual cases and generalise these to stratified populations. Healthcare has been a late adopter in all four fields but the widespread adoption of clinical information systems creates opportunities to develop new approaches to process analytics that directly address the unique nature of healthcare care pathways. Such work is novel but there is a growing body of literature demonstrating successful applications of process mining to healthcare and ongoing research within the Connected Health Cities programme which is applying process analytics to care pathways as diverse as musculoskeletal disease, diabetes, cancer, childhood obesity, A&E and frailty.

Programme

Introductions (10m) – to the workshop, the team, delegates and to the field of process analytics and challenges for healthcare. Brief demonstration of Connected Health Cities pathways and review of the method.

Modelling a NICE pathway in Gout (10m) – table-top exercise using Post-Its and guidelines from http://pathways.nice.org.uk

Creating a pathway simulation model (20m) – online exercise using the outputs from the table-top exercise and the NETIMIS simulation tool www.netimis.co.uk. Exploring pathway flow and “What If” scenarios.

Process Mining a Care Pathway (20 min) – a worked exercise with sample data and the Disco process mining tool www.fluxicon/disco. Process variants and impact on clinical outcomes.

Structured Discussion (30 min) – Examples, method, challenges, opportunities, next steps. Review of outcomes.

Outcome

This workshop is aimed at researchers and data scientists working with care pathways and linked e-health record data. It will be highly interactive with demonstrations, two short hands-on tutorials (laptops required) and a structured discussion. The objective is to create an informal network of researchers focussed on care pathway data mining with sufficient support to develop this new field within health informatics. No previous experience with process mining is required and demonstrators will be on hand to support the practical sessions.