Clinical Frailty Scale at presentation to the emergency department: interrater reliability and use of algorithm-assisted assessment

Key summary points Aim In this study, the interrater reliability of the Clinical Frailty Scale (CFS) ratings comparing assessments by both experienced and unexperienced staff (ED clinicians and a study team (ST) using a smartphone application to support CFS scoring) was evaluated. The feasibility of the CFS assignment at ED triage, defined as a majority of patients aged 65 or older assigned a CFS level at triage, was also investigated. Findings In this cross-sectional study of 1349 consecutive ED patients aged 65 years and older, the interrater reliability for CFS ratings was good for three different dyads assessed, whether used as an ordinal scale or as frailty categories (CFS 1–4 = non-frail to vulnerable; CFS 5–6 = mild to moderate frailty; CFS 7–9 = severe frailty to terminally ill). More than two-thirds (70.2%) of patients had a CFS rating assigned at triage. Message The CFS is a reliable scale for use in the ED and the implementation of frailty assessment could be facilitated with an algorithm-assisted assessment. Purpose The Clinical Frailty Scale (CFS) allows health care providers to quickly stratify older patients, to support clinical decision-making. However, few studies have evaluated the CFS interrater reliability (IRR) in Emergency Departments (EDs), and the freely available smartphone application for CFS assessment was never tested for reliability. This study aimed to evaluate the interrater reliability of the Clinical Frailty Scale (CFS) ratings between experienced and unexperienced staff (ED clinicians and a study team (ST) of medical students supported by a smartphone application to assess the CFS), and to determine the feasibility of CFS assignment in patients aged 65 or older at triage. Methods Cross-sectional study using consecutive sampling of ED patients aged 65 or older. We compared assessments by ED clinicians (Triage Clinicians (TC) and geriatric ED trained nurses (geriED-TN)) and a study team (ST) of medical students using a smartphone application for CFS scoring. The study is registered on Clinicaltrials.gov (NCT05400707). Results We included 1349 patients aged 65 and older. Quadratic-weighted kappa values for ordinal CFS levels showed a good IRR between TC and ST (ϰ = 0.73, 95% CI 0.69–0.76), similarly to that between TC and geriED-TN (ϰ = 0.75, 95% CI 0.66–0.82) and between the ST and geriED-TN (ϰ = 0.74, 95% CI 0.63–0.81). A CFS rating was assigned to 972 (70.2%) patients at triage. Conclusion We found good IRR in the assessme