Improving A Large Healthcare System Research Data Warehouse Using OHDSI's Data Quality Dashboard

In this study, we have focused on AI Implementation Science research to explore how multimodal medical data from multiple hospitals can be harmonized in the real world clinical setting to improve patient care. IN implementing AI for health-care, a major challenge is the lack of clear identification of AI Implementation Science methods that are systematic from the ones that are case-specific minor fixes. We used the state-of-the-art HL7 Standard Fast Health Interoperability Resource (FHIR) to upgrade the SC clinical research informatics infrastructure; we conducted a comprehensive study of the SC Research Data Warehouse (RDW) using the Observational Health Data Sciences and Informatics (OHDSI) Data Quality Dashboard (DQD); we reported the completeness and conformity of the data to the Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM), and the data content plausibility; and we identified the failure. We also built a visual dashboard as a detailed feedback mechanism to help Shriners' researchers evaluate data quality. With these major implementation science issues discovered and solved, we not only improved the access and quality of SC RDW data to enable clinicians and researchers to more trustworthy decision support for better patient care outcome, but more importantly, we have developed an entire workflow and pipeline that can be extended and utilized for other healthcare systems.