Evaluating Process-Mining Interoperability: Stroke Case Analysis On Italian Health Real-World Data

Abstract Background The Institute for Health Sciences in Aragon developed a process-mining based algorithm deployed using a container solution to describe the Stroke Care Pathway reproducible in different Health Systems. Our aim is to apply the proposed solution to administrative data of an Italian region to explore interoperability in the case of the stroke care; the study is developed in the context of the Joint Action on Health Information 'InfAct' funded by the European Commission. Methods The target population included 1538055 residents in Marche Region who received National Health System (NHS) assistance in 2017. The collected information was sex, date of birth and death, residence from NHS Beneficiaries database; admission and discharge date/time, hospital code, admission type, discharge status, up to 6 diagnoses from Hospital Discharge database; admission date/time, urgent care facility code, discharge status, diagnosis from Emergency Care database. Patients without hospital/urgent care events were excluded from NHS Beneficiaries database. Care pathway traces and map with frequency information, timing across the pathway and timeline plots with time information were the outputs obtained. Results Data included 17752 patients, 10390 hospital, and 18329 urgent care events. Results were represented graphically, easy to read and spread: 500 episodes of stroke were detected, 66% of patients was hospitalised after emergency room (ER) admissions, 19% directly hospitalised and 10% discharged after ER admission. The most frequent access and exit points of the pathway were ER admission and hospital discharge. Conclusions The application of the stroke case to Italian data demonstrates feasibility and usefulness of interoperability in analysing health care administrative data for public health purposes. A European health information system is crucial to build capacity between researchers across Europe and improve health data on Health Monitoring and Health System Performance Assessment. Key messages Sharing a process-mining methodology deployed using a container solution is a useful strategy to describe care pathways and ensures results’ reliability across European Health Systems. Developing and testing tools to encourage interoperability between researchers across Europe is essential for public health policymaker in Europe.