Data validation for healthcare cost analysis in STROKE OWL

Abstract

The STROKE OWL project applies a new approach of cross-sector care management for strokes and determines the costs of this new approach. An evaluator determines the current costs of stroke care for the health insurance companies involved in the project as a comparative figure. Seven health insurance companies provide all necessary data in a uniform format for the health care cost analysis, despite different systems and internal formats. One important aim is to prevent the evaluator from performing analyses specific to individual health insurance companies. Therefore, a mediator accepts the data from the health insurance companies, checks the conformity of the data with the agreed format, and tries to make the re-identification of health insurance companies as difficult as possible through suitable transformations before forwarding the data to the evaluator. The conformance checking is particularly challenging due to underestimated effort and communication overhead for every participant. We propose a process for the data validation and a system assisting the validation of the mediator and describe our experience with the data validation for the health care cost analysis.

Keywords

• data science
• electronic data processing
• health information exchange
• data management