Improved Identification of Venous Thromboembolism From Electronic Medical Records Using a Novel Information...

Area of Science:

• Medical Informatics
• Health Services Research
• Radiology

Background:

• Federally mandated reporting of venous thromboembolism (VTE) relies on administrative data with unproven accuracy.
• The Agency for Healthcare Research & Quality Patient Safety Indicator 12 (AHRQ PSI-12) is a key metric for VTE surveillance.
• Accurate VTE identification is crucial for patient safety and quality improvement initiatives.

Purpose of the Study:

• To evaluate the accuracy of IDEAL-X, a novel information extraction software, in identifying VTE from electronic medical records.
• To compare the performance of IDEAL-X with the established AHRQ PSI-12 method for VTE detection.
• To assess the potential of IDEAL-X for improving VTE surveillance.

Main Methods:

• A retrospective review of 13,248 patient records from an orthopedic specialty hospital (2009-2014).
• Patient encounters were defined by surgery (spine, hip, knee) and VTE-detecting radiology studies.
• Radiology reports were analyzed by both manual physician review and the IDEAL-X software.

Main Results:

• IDEAL-X achieved high accuracy in identifying VTE from radiology reports (sensitivity 97.2%, specificity 99.3%).
• Performance in surgical encounters was also high (sensitivity 97.6%, specificity 99.8%).
• IDEAL-X demonstrated superior specificity compared to AHRQ PSI-12 (99.3% vs. 92.9%, P<0.01).

Conclusions:

• IDEAL-X accurately identifies VTE from radiology reports, outperforming AHRQ PSI-12 in specificity.
• This software has the potential to enhance the detection and surveillance of various medical conditions using electronic health record data.
• IDEAL-X offers a promising tool for improving the accuracy of VTE reporting and patient care.