Electronic Health Record Data Quality and Performance Assessments: A Scoping Review (Preprint)

Abstract

<sec> <title>BACKGROUND</title> Electronic Health Records (EHRs) have an enormous potential to advance medical research and practice through easily accessible and interpretable EHR-derived databases. Attainability of this potential is limited by issues with data quality and performance assessment. </sec> <sec> <title>OBJECTIVE</title> This review aims to streamline the current best practices on EHR Data Quality and Performance assessments as a replicable standard for researchers in the field. </sec> <sec> <title>METHODS</title> PubMed was systematically searched for original research articles assessing EHR data quality and/or performance from inception until May 7, 2023. </sec> <sec> <title>RESULTS</title> Our search yielded 26 original research articles. Most articles suffered from one or more significant limitations, including incomplete or inconsistent reporting (30%), poor replicability (25%), and lacking generalizability of results (25%). Completeness (81%), Conformance (69%), and Plausibility (62%) were the most cited indicators of Data Quality, while Correctness/Accuracy (54%) was most cited for Data Performance, with context-specific supplementation by Recency (27%), Fairness (23%), Stability (15%), and Shareability (8%) assessments. Artificial Intelligence (AI)-based techniques including natural language data extraction, data imputation, and fairness algorithms were demonstrated to play a rising role in improving both dataset quality and performance. </sec> <sec> <title>CONCLUSIONS</title> This review highlights the need for incentivizing data quality and performance assessments and their standardization. The results suggest utility of the adoption of AI-based techniques for enhancing data quality and performance to unlock the full potential of EHRs to improve medical research and practice. </sec>

Ähnliche Arbeiten