Back to 2015 Annual Meeting
Redesigned Electronic Medical Notes Allow Automated Clinical Data Extraction and Decrease Provider Documentation Time
Jose G. Christiano, MD.
University of Rochester, Rochester, NY, USA.
BACKGROUND:
Over the last few years, widespread implementation of electronic medical records (EMR) has increased the burden of documentation on providers. Nevertheless, retrieval of meaningful data from EMR notes for research or quality control purposes continues to be mostly renegaded to individual chart review and/or manual entry into databanks. We hypothesized that standard EMR notes could be redesigned to provide customary documentation and allow automated clinical data retrieval by commercially available text data extraction software (TDES), with minimal disruption to provider workflow (documentation time).
METHODS:
Twenty fictitious patients undergoing reduction mammoplasty were created. Fictitious encounters included initial consultation, preoperative visit, surgery, and postoperative visits at 1, 8, and 25 weeks. Each encounter was documented with our previous standard note (SN) and a redesigned "data-friendly" note (DFN). All notes were entered by the author. Documentation time was measured and compared between the two note groups (SN and DFN) using Student's t-test. All DFNs were then exported in PDF format into a TDES for data accrual. Seventy-six variables were assigned for data monitoring and retrieval, spanning from patient's demographics (medical record number, name, age, gender, date of birth, etc), to elements of the history and physical (complaints of neck/shoulder/back pain, allergies, comorbidities, vital signs, breast measurements, etc), to operative details (skin pattern, nipple pedicle design, resected breast tissue weight, blood loss, duration of procedure, etc), to clinical outcome (pathology result, postoperative pain, nipple sensation, fat necrosis, viability of grafted nipple/areola, etc).
RESULTS:
The TDES successfully analyzed all 120 DFNs (300 pages) in less than 10 seconds, generating a database containing 4,850 clinical data points. Total documentation time per patient was actually found to be less in the DFN group (mean 20.5 min, standard deviation 1.5 min) than in the SN group (mean 21.3 min, standard deviation 1.6 min), reaching statistical significance (p<0.01).
CONCLUSIONS:
Redesigned EMR "data-friendly" notes allow automated clinical data retrieval by commercially available text data extraction software and decrease provider documentation time.
Back to 2015 Annual Meeting
|
