Association between an electronic health record (EHR)–embedded frailty index and survival among older adults receiving cancer chemotherapy.

Abstract

12009 Background: Innovative strategies to facilitate rapid frailty screening are critically needed to enhance personalized oncology care among vulnerable older patients. An electronic frailty index (eFI) holds promise as a passive measure of frailty. The eFI was initially designed for primary care. The objective of this study was to establish feasibility, estimate prevalence of frailty, and evaluate the predictive value of the eFI for overall survival (OS) among older adults with cancer planned to receive chemotherapy. Methods: Consecutive patients (N = 509) aged 65+ with newly diagnosed lung, colorectal or breast cancer treated with chemotherapy were identified from our cancer registry between 2017 and 2020. Calculation of the eFI requires at least two ambulatory visits over a 2-year period and utilizes demographic information, vital signs, smoking status, ICD-10 diagnosis codes, select outpatient laboratory measurements, and functional information (if available from Medicare Annual Wellness Visits) during the 2 years prior to diagnosis. Frailty status is categorized as fit (eFI ≤0.10), pre-frail (0.10 < eFI≤0.21), and frail (eFI > 0.21) based on the proportion of deficits present over the total number evaluated (score range 0-1). Non-calculable scores indicate insufficient historical primary care data within two years prior to cancer diagnosis. OS was estimated by eFI category using Kaplan Meier curves and compared using log rank testing. Cox proportional hazards models evaluated the adjusted association between eFI categories and mortality. Results: The cohort included 509 adults (median age 72.2 yrs, 55% female, 83.5% white, 13.4% black, 45.8% stage 4) with lung (N = 312), colorectal (N = 111) and breast (N = 86) cancer. Distribution of eFI categories at diagnosis were fit (25.9%), pre-frail (41.1%), frail (17.3 %) and not calculable (15.7%). The proportion of patients categorized as “fit” differed by disease type (19.9% lung, 30.6% colorectal, 41.9% breast, p < 0.0001). In univariate analyses, eFI frailty category was associated with OS (median OS for fit, pre-frail, frail and not calculable were > 54, 25, 19 and 10 months respectively, p < 0.0001.). Adjusting for age, gender, race, stage, and cancer type, the hazard of death was higher for pre-frail (Hazard Ratio, [HR] 1.7, 95% Confidence Interval [CI] 1.2-2.4), frail (HR 2.3, 95% CI 1.5-3.4) and not calculable eFI categories (HR 2.8, 95% CI 1.8-4.2) compared to fit patients. Conclusions: Calculating an EHR embedded eFI at the time of diagnosis among older adults treated with chemotherapy was feasible and identified nearly one-fifth of patients as frail. eFI-defined frailty status was associated with survival, with poorer survival among the most frail supporting eFI validity. The eFI is a promising scalable tool to efficiently conduct frailty screening in oncology clinical practice.