Top ↑

Dictation Mining

Parsing and Full Text Searching of Physician Dictation

Despite the addition of EMR, physicians continue to rely on the speed and ease of oral dictation to update patient charts, a habit reinforced by payers who require dictated notes for reimbursement. Because of this, key elements of the patient record may be contained only in physician dictations.

Data aggregation that relies solely on EMR can miss important facts, such as metastases, cancer stage, patient ethnicity and race, or specific type of cancer.

EMOL Health solutions include powerful software capabilities to find and mine information from oral dictation, giving our aggregated data more detail and accuracy.

We use two techniques to uncover patient information in physician dictations:

Full-Text Searching

After indexing all the words in a patient dictation, our software searches for significant matches using an algorithm to increase the accuracy of a match. For example, we know that physicians generally mention a patient’s race in the first sentence of dictation. Any subsequent uses of words that may seem to be a description of race are less likely to be a description of race, reducing the accuracy of that match.

Data Parsing

Our system also creates value through data parsing – the ability to dig though data to find important patient information that is more significant to researchers than to payers, such as co-morbidities, cancer stage, or histology code. EMOL Health can parse medications and diagnoses out of dictations with more than 90 percent accuracy.