Applicability of the electronic frailty index in younger and older adults in England: a population-based cohort study
Daniel R. Morales, Bruce Guthrie, Thomas J Downes, David McAllister, Peter Hanlon
Abstract
BACKGROUND: The electronic frailty index (eFI) was developed in older adults (aged ≥65 years). There are currently no validated frailty scores in clinical practice for younger adults (aged 18-64 years). The aim of this study was to examine whether the eFI score in younger adults had similar or different associations with adverse health outcomes compared with older adults. METHODS: In this population-based cohort study, electronic health records from the UK Clinical Practice Research Datalink GOLD database were analysed. We used a cross-section of patients on Nov 30, 2015, who were alive and had been registered with a general practice for at least 2 years before data capture. Patients were stratified into younger adults (aged 18-64 years, n=708 235; 49·4% female) and older adults (aged 65-95 years, n=231 819; 54·3% female). For all included patients, eFI score, prevalence of individual eFI deficits, and eFI frailty category were calculated. For the main outcomes, crude and age-sex adjusted hazard ratios (HRs) were calculated for 1-year and 3-year mortality and emergency hospitalisation for each group compared with adults defined by the eFI as fit. FINDINGS: The prevalence of eFI-defined frailty was higher in older adults than younger adults. Specifically, in older adults, 77 290 (33·3%) of 231 819 had mild frailty, 44 523 (19·2%) had moderate frailty, and 22 572 (9·7%) had severe frailty. For younger adults, 76 991 (10·9%) of 708 235 had mild frailty, 12 552 (1·8%) had moderate frailty, and 2088 (0·3%) had severe frailty. Adjusted HRs for both 1-year mortality and 1-year emergency hospitalisation in younger adults with mild, moderate, and severe frailty were greater than in older adults with equivalent frailty categorisation. Specifically, compared with fit older adults, age-sex adjusted 1-year mortality HRs were 1·94 (95% CI 1·80-2·09) in older adults with mild frailty, 2·99 (2·77-3·22) with moderate frailty, and 4·03 (3·72-4·36) with severe frailty. Compared with fit younger adults, age-sex adjusted 1-year mortality HRs were 3·15 (2·80-3·55) in younger adults with mild frailty, 5·88 (4·95-6·98) with moderate frailty, and 12·61 (9·76-16·30) with severe frailty (Z score p<0·001 for all comparisons). Compared with fit older adults, age-sex adjusted HRs for 1-year emergency hospitalisation were 2·30 (2·22-2·39) in older adults with mild frailty, 4·09 (3·94-4·25) with moderate frailty, and 6·76 (6·50-7·03) with severe frailty. Compared with fit younger adults, age-sex adjusted HRs for 1-year emergency hospitalisation were 3·16 (3·07-3·25) in younger adults with mild frailty, 6·64 (6·34-6·94) with moderate frailty, and 13·02 (12·04-14·09) with severe frailty (Z score p<0·001 for all comparisons). Similar associations were observed for 3-year mortality and emergency hospitalisation. INTERPRETATION: Similarly to older adults, the eFI identifies younger adults with frailty at high risk of mortality and emergency hospital admission. The eFI might be a tool to identify individuals for further assessment and intervention. FUNDING: Wellcome Trust and Chief Scientist Office.