Study: EHR Data Used With Predictive Analytics Can Flag Patient Falls
June 29, 2015 in News
Combining electronic health record data with predictive analytics algorithms could help identify patients in nursing homes at risk of harmful falls, according to a study published in the Journal of the American Medical Informatics Association, Health IT Analytics reports.
According to the study, falls are the most frequently reported adverse event among nursing home patients, with rates as high as nearly four falls per bed per year in some facilities. A non-fatal event can total more than $7,300 in direct costs, taking into account expenses for:
- Long-term disability;
- Reduced functioning; and
- Complications from necessary treatments (Bresnick, Health IT Analytics, 6/26).
For the study, researchers analyzed data from both the CMS Minimum Data Set and EHRs for 5,129 residents in 13 nursing homes to determine if EHR data improved fall predictions. The facilities are part of a large California chain that uses a centralized EHR system (Marier et al., JAMIA, 6/23).
According to the researchers, the CMS Minimum Data Set is an event-based assessment for nursing homes that can help predict patients at high risk for falls. However, the data are collected infrequently and do not account for all the crucial factors that can lead to falls. EHRs, however, can add recent clinical data to the overall analysis for a more thorough and up-to-date representation of patient risk.
The researchers found that adding EHR data to the predictive analytics model increased the accuracy of the algorithm by 13%. Specifically, the algorithm with the EHR data confirmed 32.2% of observed falls were among patients in the highest 10% of risk, compared with 28.6% of observed falls identified by the algorithm when it relied solely on MDS data.
The authors predicted that if the system was 100% successful in preventing falls among residents at high risk, the incorporation of EHR data into risk-prediction algorithms could prevent six additional falls per year, saving $43,842 annually.
The authors wrote, “Because falls are so frequent and so costly, a back-of-the-envelope computation implies that even small improvements in identification of high-risk residents may result in large potential cost savings” (Health IT Analytics, 6/26).