EHR Data Valuable to Study of Cardiac Structure, Function
Researchers from Vanderbilt University have discovered that electrocardiograph data harvested from electronic health records can be efficiently extracted as a rapid and efficient method for the study of cardiac structure and function.
Researchers from Vanderbilt University have discovered that electrocardiograph data harvested from electronic health records can be efficiently extracted as a rapid and efficient method for the study of cardiac structure and function.
The study, published in the Journal of Clinical Bioinformatics, suggests that EHR-based cohorts have the potential to make major contributions toward the study of epidemiologic and genotype-phenotype associations for cardiac structure and function in diverse populations. Cardiac structure, such as wall thickness and left ventricular dilation, can predict cardiovascular disease events and heart failure.
The authors studied 6,076 echocardiography reports for 2,834 unique adult subjects. Vanderbilt University Medical Center's EHR-based ECG report archives go back to 1997 and are stored in the PDF format. Missing data were uncommon with over 90 percent of data points present.
Data irregularities were primarily related to inconsistent use of measurement units and transcriptional errors. The researchers filtering method required manual review of very few data points (fewer than 1 percent), and filtered echocardiographic parameters were similar to published data from epidemiologic populations of similar ethnicity. Moreover, the cohort was comparable in size, and in some cases larger than community-based cohorts of similar race/ethnicity.
While the authors noted several limitations to their study, such as the exclusion of clinically relevant semi-structured and unstructured data, they also concluded the benefits of extracting the available data efficiently offered promising signals were sufficient to warrant further consideration.
Future directions include refinement of methods for extraction and filtering of semi-structured and unstructured echocardiographic date from EHRs and leveraging of the Vanderbilt echocardiography cohort for study of cardiac structure and function genotype-phenotype associations.
The study, published in the Journal of Clinical Bioinformatics, suggests that EHR-based cohorts have the potential to make major contributions toward the study of epidemiologic and genotype-phenotype associations for cardiac structure and function in diverse populations. Cardiac structure, such as wall thickness and left ventricular dilation, can predict cardiovascular disease events and heart failure.
The authors studied 6,076 echocardiography reports for 2,834 unique adult subjects. Vanderbilt University Medical Center's EHR-based ECG report archives go back to 1997 and are stored in the PDF format. Missing data were uncommon with over 90 percent of data points present.
Data irregularities were primarily related to inconsistent use of measurement units and transcriptional errors. The researchers filtering method required manual review of very few data points (fewer than 1 percent), and filtered echocardiographic parameters were similar to published data from epidemiologic populations of similar ethnicity. Moreover, the cohort was comparable in size, and in some cases larger than community-based cohorts of similar race/ethnicity.
While the authors noted several limitations to their study, such as the exclusion of clinically relevant semi-structured and unstructured data, they also concluded the benefits of extracting the available data efficiently offered promising signals were sufficient to warrant further consideration.
Future directions include refinement of methods for extraction and filtering of semi-structured and unstructured echocardiographic date from EHRs and leveraging of the Vanderbilt echocardiography cohort for study of cardiac structure and function genotype-phenotype associations.
More for you
Loading data for hdm_tax_topic #better-outcomes...