Prompt recognition of frailty in the emergency department (ED) is important to identify patients at higher risk of adverse outcomes. Despite this, few studies examine the diagnostic accuracy of screening instruments for frailty, instead focusing on predictive validity. We compared three commonly used, short frailty screens to an independent comprehensive geriatric assessment (CGA) in an urban University Hospital ED. Consecutive attendees aged ≥70 years were screened by trained raters, blind to the CGA, with the Variable Indicative of Placement risk (VIP), 3 and 4-item versions, Clinical Frailty Scale (CFS) and PRISMA-7. Accuracy was measured from the area under the ROC curve (AUROC). In total, 197 patients were included, median age 79 (±10); 46% were female. Half (49%) were confirmed as frail after CGA. All instruments differentiated frail from non-frail states, although the CFS (AUROC: 0.91) and PRISMA-7 (AUROC: 0.90) had higher accuracy compared to the VIP-4 (AUROC: 0.84) and VIP-3 (AUROC: 0.84). The CFS was significantly more accurate than the VIP-3 (p = 0.026) or VIP-4 (p = 0.047). There was no significant difference between the CFS and PRISMA-7 (p = 0.90). The CFS and PRISMA-7 were more accurate and should be considered in preference to the VIP (3 or 4-item versions) to identify frailty in EDs.

Keywords: PRISMA-7; clinical frailty scale; diagnostic accuracy; emergency department; frailty; screening; variable indicative of placement risk.