A Data Quality Framework for Process Mining of Electronic Health Record Data


Reliable research demands data of known quality. This can be very challenging for electronic health record (EHR) based research where data quality issues can be complex and often unknown. Emerging technologies such as process mining can reveal insights into how to improve care pathways but only if technological advances are matched by strategies and methods to improve data quality. The aim of this work was to develop a care pathway data quality framework (CP-DQF) to identify, manage and mitigate EHR data quality in the context of process mining, using dental EHRs as an example. Objectives: To: 1) Design a framework implementable within our e-health record research environments; 2) Scale it to further dimensions and sources; 3) Run code to mark the data; 4) Mitigate issues and provide an audit trail. Methods: We reviewed the existing literature covering data quality frameworks for process mining and for data mining of EHRs and constructed a unified data quality framework that met the requirements of both. We applied the framework to a practical case study mining primary care dental pathways from an EHR covering 41 dental clinics and 231,760 patients in the Republic of Ireland. Results: Applying the framework helped identify many potential data quality issues and mark-up every data point affected. This enabled systematic assessment of the data quality issues relevant to mining care pathways. Conclusion: The complexity of data quality in an EHR-data research environment was addressed through a re-usable and comprehensible framework that met the needs of our case study. This structured approach saved time and brought rigor to the management and mitigation of data quality issues. The resulting metadata is being used within cohort selection, experiment and process mining software so that our research with this data is based on data of known quality. Our framework is a useful starting point for process mining researchers to address EHR data quality concerns.



Data Mining algorithms

Software And Hardware

• Hardware: Processor: i3 ,i5 RAM: 4GB Hard disk: 16 GB • Software: operating System : Windws2000/XP/7/8/10 Anaconda,jupyter,spyder,flask Frontend :-python Backend:- MYSQL