Principal Investigator: Jennifer Garvin, PhD Project Title: Automated Data Acquisition for Heart Failure: Performance Measures and Treatment ANTICIPATED IMPACT ON VETERANS HEALTHCARE: The VA redesigned its health care system, the Veterans Health Administration (VHA), by including better use of information technology, measurement and reporting of performance, reorganization of health care delivery and realignment of payment policies. However, the collection of data for quality measures through retrospective chart review is time consuming and expensive. Using information extraction (IE) techniques to capture data for quality measurement would be beneficial to the health of veterans and would improve health care delivery in the VA. PROJECT BACKGROUND: There is an association of heart failure (HF) with high mortality and poor quality of life. Heart failure currently affects nearly 5 million Americans, and hospital admissions for this condition have increased six-fold in the United States since 1970 (from 80,000 per year to more than 500,000) due, in part, to an aging population. Heart failure is also the number one cause of hospitalizations of veterans treated within the VA health care system. Because of routine use of the electronic health record and the increasing availability of IE techniques, the potential of capturing the performance measurement data in an automated way is a realistic pursuit. OBJECTIVES: The objectives of the study are to: a. use information extraction techniques to identify and extract the required elements of the treatment dimension of the performance measure for inpatients with congestive heart failure, b. to compare these techniques to existing manual abstraction methods, and c. to examine performance deviations. METHODS: This is a descriptive study that will be undertaken in concert with the Consortium for Healthcare Informatics Research (CHIR) to provide accuracy measures of sensitivity, specificity, and accuracy of information extraction methodology to identify the presence of performance measurement data elements of: evaluation of left ventricular systolic function, ejection fraction, and appropriate medication prescription for inpatients with heart failure. The criteria used to determine the presence of the data elements will be the same as the EPRP required elements. This study will provide a comparison of the accuracy of the existing manual method of data abstraction with a new method of automated data abstraction, evaluate providers acceptance of these new methods, and develop collaborate with the VA Office of Performance and Quality and other constituent groups to disseminate the findings of the research.
The VA uses information technology to measure and report performance to better its service to veterans. This research will examine the use of information extraction (IE) techniques to capture relevant metrics for inpatients with chronic heart failure. The findings of the research will inform how IE techniques can be used to improve performance data acquisition and the how metrics extracted by these techniques will be perceived by providers.
| Meystre, Stéphane M; Kim, Youngjun; Heavirland, Julia et al. (2015) Heart Failure Medications Detection and Prescription Status Classification in Clinical Narrative Documents. Stud Health Technol Inform 216:609-13 |
| Kim, Youngjun; Garvin, Jennifer; Goldstein, Mary K et al. (2015) Classification of Contextual Use of Left Ventricular Ejection Fraction Assessments. Stud Health Technol Inform 216:599-603 |
