Electronic frailty index predicts risk of adverse outcomes

Mass General Brigham researchers have created an electronic frailty index that can identify older adults at increased risk of adverse outcomes, as outlined in a Journal of the American Geriatric Society study.

The risk prediction tool measures patient frailty, an aging-related syndrome, by integrating EHR data of over 500,000 patients collected across multiple Mass General Brigham hospitals.

The study included patients aged 60 and above who had experienced one to two outpatient visits within a two- or three-year period prior to 2017. The analysis also included a sub-cohort that received primary care within the same time window.

The researchers developed a frailty index that classified patients as robust, pre-frail, frail or very frail based on the presence or absence of 31 aging-related health deficits in their EHR data.

Across 518,449 patients with a mean age of 71.94 years, the study identified 72.9% as robust, 15.8% as pre-frail, 6.9% as frail and 2.8% as very frail.

Relative to robust individuals, very frail patients had increased rates of death and hospital readmission within a 90-day window. Rates of worse outcomes grew from pre-frail to very frail individuals, compared to robust individuals.

The authors noted that while the tool was created through cross-system collaboration at Mass General Brigham, the model could also help other health systems identify older adults at risk of adverse outcomes.

“Our study shows that frailty can be measured in EHRs, even when data is incomplete,” Ariela R. Orkaby, MD, MPH, senior author of the study, said in a press release.

“For hospitals caring for an older and sicker population, an automated frailty tool, such as the Mass General Brigham-Electronic Frailty Index, is a rapid way to identify patients at highest risk of an adverse health outcome,” Orkaby added. “We hope that this tool will help improve care of all older adults entering the healthcare system.”

Hannah Nelson has been covering news related to health information technology and health data interoperability since 2020.