Talk to the Veterans Crisis Line now
U.S. flag
An official website of the United States government

VA Health Systems Research

Go to the VA ORD website
Go to the QUERI website
2015 Conference Logo

2015 HSR&D/QUERI National Conference Abstract

3113 — Development of an automated tool to identify and extract functional status data from national VA data sources

Brown RT, San Francisco VA Medical Center; Komaiko KD, San Francisco VA Medical Center; Fung KZ, San Francisco VA Medical Center; Steinman MA, San Francisco VA Medical Center;

The ability to perform basic daily activities ("functional status") is central to older adults' quality of life and strongly predicts adverse health outcomes. However, data on functional status have seldom been systematically collected during clinical care or captured in administrative data sources, which has limited the ability of clinicians and researchers to use these data to improve care for older adults. With support from the VA Office of Geriatrics and Extended Care, a number of VA medical centers have started assessing functional status during primary care appointments for Veterans age 75 and older. We are currently conducting a study to validate these measures. In this abstract, we present results from phase 1 of the project, in which we developed automated tools to identify and extract functional data from national VA data sources.

To identify functional status data, we first extracted all CDW Health Factors data for individuals age 65 and older in 2009-2013. Because templates for recording functional data vary across VA health systems, we used a broad range of keywords to identify measures of activities of daily living (ADL) and instrumental activities of daily living (IADL). Using encounter codes, we restricted our list to measures collected in primary care appointments.

From 238,150 unique Health Factors, we identified 2178 that included 1 or more keywords and were collected in primary care appointments. We narrowed this list to include 442 Health Factors from 17 stations that collected complete data on each of 5 basic ADLs and 8 IADLs. We automated this extraction process to allow real-time identification of patients who had functional data collected within the past week. This process identifies approximately 2460 patients per week, of whom about 8% are dependent in 1 or more ADLs.

We developed an automated approach to identify and extract functional data from national VA data sources. Next steps include completing an ongoing study to validate these measures compared to a gold standard of structured self-report.

If our study confirms the validity of VA functional measures, VA should consider implementing standardized functional assessment nationwide so that these data can be used to improve the wellbeing of older Veterans.