Extracting Echocardiography Metrics from EMR

Echocardiography reports in the VUMC EMR are in the portable document format (PDF) and have undergone three formatting iterations since 1997. Reports prior to 1997 are not in digital format and not included in the EMRs. Each report contains structured, semi-structured, and unstructured data. Structured data are generally quantitative measures such as wall thicknesses, chamber dimensions, or flow velocities. Semi-structured data fields contain subjective interpretations of parameters with a limited number of potential values. These fields frequently contain ordinal data. For example, valvular lesions and abnormalities of ventricular function are often subjectively quantified as “mild”, “moderate”, or “severe”. Unstructured fields contain unrestricted prose descriptions of clinically relevant findings as interpreted by the reader.
Fields containing structured, semi-structured, and unstructured data were identified within echocardiography reports in the EMRs. Numeric values for left ventricular septal thickness, left ventricular posterior wall thicknesses, left ventricular end systolic diameter, left ventricular end diastolic diameter, left atrial diameter, and aortic root diameter were subsequently parsed from reports using natural language processing.