Abstract
A growing number of aging adults are living with multiple chronic conditions (MCC). Older adults living with MCC are predisposed to developing frailty, a state of decreased physiologic reserve that increases risk for geriatric syndromes and associated morbidity and mortality. The electronic frailty index (eFI) is computed passively using structured EHR data and can aid in prospective screening. Unfortunately, certain diagnoses, such as functional status assessments in unstructured documentation, are less likely captured by eFI, potentially underestimating the degree of frailty. Here, we discuss current gaps for using eFI to identify frail older adults living with MCC, and artificial intelligence (AI) approaches to enhance eFI accuracy. Accurate and routine frailty assessment can aid the generalist providing care to older adults living with MCC across multiple care settings to optimize physiologic reserve for these vulnerable patients.