The widespread adoption of EHRs has enabled the collection of massive amounts of digital ophthalmic data which have great potential for secondary use in research, quality improvement, and clinical decision support. While the amount of digital ophthalmic data recorded in the EHR is substantial and could be analyzed using the latest techniques for big data, questions about the quality of the data are a barrier to its reuse. Now that the American Academy of Ophthalmology has aggregated digital ophthalmic data from the EHR into the IRIS Registry, data quality is even more imperative for reaching the potential of the registry. To date, there has not be a comprehensive evaluation of the data quality of digital ophthalmic data, nor have there been any solutions for improving its quality. These are important gaps that will limit the utility of EHR data as a tool for knowledge discovery in ophthalmology. The goal of this grant is to assess the quality of digital ophthalmic exam data in order to improve its ability to be reused for research. Our hypothesis is that studying the variability of data quality in large datasets will provide insights into improving its quality.
The first aim employs an established framework for data quality analysis to assess the intrinsic quality of a single institution?s EHR data as well as its fitness for use--the ability to be applied to a particular research scenario. In this proposal, we are evaluating the data?s ability to identify patient cohorts for clinical trials and to accurately calculate outcome based clinical quality measures. The variability in data?s quality and fitness among providers, subspecialties, diagnoses, and visit types will be analyzed.
The second aim v alidates the analysis of the first aim by repeating it for all of the ophthalmic data in the IRIS Registry. For this analysis, the differences in quality and fitness between institutions and EHR vendors will also be assessed, along with the barriers to data quality and reuse. For both aims, ophthalmology experts will review the results to make recommendations for improving data quality and utility for digital ophthalmic data. In the future, these recommendations will provide a direction for correcting these quality issues and for ultimately advancing knowledge discovery in ophthalmic care.
Electronic health records (EHRs) have not yet reached their potential for transforming healthcare, particularly for reusing clinical data for research. The American Academy of Ophthalmology has aggregated ophthalmic data from the EHR into the IRIS Registry, and data quality is even more imperative to achieve to reach the potential of this registry. Using methodological data quality analysis, we will analyze the quality of a single institution?s ophthalmic data and again for multiple institutions? ophthalmic data in the IRIS Registry, documenting barriers for data quality and reuse that will lead to improving knowledge discovery from this data.
