Date of Graduation
School of Public Health
Unintentional falls among older adults are a complex public health problem both nationally and in West Virginia. Nationally, nearly 40% of community-dwelling adults age 65 and older fall at least once a year, making unintentional falls the leading cause of both fatal and non-fatal injuries among this age group. This problem is especially relevant to West Virginia, which has a population ageing faster on average than the rest of the nation. Identifying falls risk in the primary care setting poses a serious challenge. Currently, the Timed Get-Up-and-Go test is the only recommended screening tool for determining risk. However, nationally this test is completed only 30-37% of the time. Use of electronic health record data as clinical decision support in identifying at-risk patients may help alleviate this problem. However, to date there have been no published studies on using electronic health record data as clinical decision support in the identification of this particular population. This presents opportunity to contribute to the fields of falls prevention and health informatics through novel use of electronic health record data. That stated, this research is designed to: 1) develop an understanding of current falls risk screening practices, facilitators, and barriers to screening in select West Virginia primary care centers; 2) assess the capture of falls risk data and the quality of those data to help facilitate identification of at-risk patients; and 3) build an internally validated model for using electronic health record data for identification of at-risk patients. Through focus group discussions with primary care partners, we find a significant lack of readiness to innovatively use routinely collected data for population health management for falls prevention. The topic of falls risk identification is a rarely discussed topic across these sites, with accompanying low rates of screening and ad-hoc documentation. The need for enhanced team-based care, policy, and procedure surrounding falls is evident. Using de-identified electronic health record data from a sample of West Virginia primary care centers, we find that it is both feasible and worthwhile to repurpose routinely collected data to identify older adult patients at-risk for falls. Among 3,933 patients 65 and older, only 133 patients (3.4%) have an indication in their medical records of falling. Searching the free text data was vital to finding even this low number of patients, as 33.8% were identified using free text searches. Given the focus group findings, underreporting of falls on the part of the patients and missed opportunities to learn of falls due to lack of information sharing across health care service sites are also contributing factors. Similarly, documentation of falls risk assessments were sparse with only 23 patients (0.6%) having documentation of a falls risk assessment in their medical records at some point in the past. As with falls, locating documentation of falls risk assessments was largely dependent on semi-structured and free text data. Current Procedural Terminology coding alone missed 26.1% of all falls risk assessments. Repurposing electronic health record data in a population health framework allows for concurrent examination of primary and secondary falls risk factors in a way which is sensitive to time constraints of the routine office visit, complementary to the movement toward Meaningful Use, while providing opportunity to bolster low screening rates.
Baus, Adam, "Understanding Falls Risk Screening Practices and Potential for Electronic Health Record Data-Driven Falls Risk Identification in Select West Virginia Primary Care Centers" (2015). Graduate Theses, Dissertations, and Problem Reports. 5166.