Healthcare Pathway Discovery, Conformance, and Enrichment
Michael O’Sullivan1, Andreas W. Kempa-Liehr1, Christina Lin1, Delwyn Armstrong2, and Randall Britten3
1. Department of Engineering Science, University of Auckland
2. Waitemata District Health Board, Auckland
3. Orion Health, Auckland
Healthcare pathways are critical for reducing clinical variability and maximizing health outcomes. This project will investigate the utilization of Business Process Modelling (BPM) to provide a scaffold for healthcare pathway discovery, conformance analysis, and enrichment.
The efficacy of the BPM approach is demonstrated via a case study that applies the process mining pipeline to discover appendicitis and cholecystitis healthcare pathways from hospital records. Two years’ worth of data from 2015 to 2017 on appendicitis and cholecystitis pathways are used. The resulting pathways have been reviewed by clinical experts and this review confirmed that the two discovered healthcare pathways comply with their knowledge of real clinical cases.
The healthcare pathways are subsequently enriched with demographic data and machine learning methods are utilised to explore factors that influence patient recovery time. A partial least squares regression model which estimates patient recovery time based on the available information at the end of surgery is developed. The machine learning models presented here have the potential to be very useful for hospital scheduling purposes.
The designed process mining pipeline is effective for analysis of simple healthcare pathways. Future work would apply the same pipeline to complex healthcare pathways.
This presentation is eligible for the ORSNZ Young Practitioners Prize.