Using Text Data Extraction Software to Automate Capture of Adverse Event Data from Electronic Medical Notes
Jose G. Christiano, MD, FACS.
University of Rochester, Rochester, NY, USA.
PURPOSE: Tracking of adverse events (AEs) in clinical practice is usually overseen by quality-control officers. This process can be time-consuming, labor-intensive, and inaccurate, as it may depend on remote event recollection by providers and/or chart data interpretation by the officers. In this study, we aimed to (1) prove that a commercially available text data extraction software (TDES) could be used to retrieve patient and AE-specific data from clinical notes entered on our electronic medical record system (EMRS) during standard patient encounters, and (2) measure the added provider documentation time (PDT) associated with this "on-the-fly" AE documentation.
METHODS: A 9-line text template was created on our institution’s EMRS for documentation of each AE. One hundred AEs were documented within progress notes of 50 fictitious patients. PDT for each AE was measured. To simulate a full month of clinical data, notes without AEs were added, reaching a total of 1,000 notes from 100 patients. All notes were then exported (PDF format) and fed into a TDES, which retrieved provider-entered AE data, along with embedded patient demographics and encounter data.
RESULTS: The TDES successfully analyzed the fictitious monthly report containing 1,000 encounter notes (3,000 pages), generating a database containing 50 patients, 100 AEs, and 2,000 data points. AE PDT ranged from 55 to 119 seconds, averaging 82 seconds.
CONCLUSION: A commercially available TDES can be used to retrieve patient and AE-specific data from EMRS clinical notes. "On-the-fly" documentation added, on average, less than 1.5 minute of PDT per AE.
Back to 2016 Joint Meeting Abstracts