Validation of patient identification in an HL7 messages integrator for health data monitoring and portability

Abstract

Abstract Introduction: In the healthcare sector, data quality is a critical aspect with high impact in the clinical care process. The quality of patient identification prevents misidentifications and lack of information. Methods: In this work we used an integration engine to receive, process and route HL7 messages. By analysing these messages, we can provide the means to overcome the heterogeneity present in existing health information systems data models and architecture. The aim of this work was to create a validation solution for patient identification using HL7 messages as source. The solution accepts a patient name and returns information about their quality. Results: A total of 1.048.576 messages were gathered and processed by the solution. The performed tests identified erroneous patient names (n=40.699) and also systematic errors caused by some health information systems. It also provided a method to increase the visibility of these problems, and act accordingly to correct them. Discussion: In a production environment, the tests performed confirmed the solution’s ability to identify common errors that happen across communications in a health institution network. Most common errors detected were related to the patient name field being used for other functions than those for which it was designed.