Background: frailty is a state of vulnerability to adverse outcomes. Routine identification of frailty is recommended in international guidance. This systematic review investigates the diagnostic test accuracy (DTA) of simple instruments for identifying frailty in community-dwelling older people.
Methods: the review methodology followed Cochrane procedures. Databases were searched from January 1990 to October 2013. Prospective studies assessing the DTA of simple instruments for identifying frailty in community-dwelling older people (aged ≥65 years) as index tests against a reference standard phenotype model, cumulative deficit frailty index or comprehensive geriatric assessment were eligible for inclusion. Sensitivity, specificity, positive predictive value, negative predictive value and likelihood ratios were calculated for index tests. Risk of bias was assessed using the QUADAS-2 checklist.
Results: three studies involving 3,261 participants were included. Median frailty prevalence was 10.5%. Seven index tests were assessed: gait speed, timed-up-and-go test, PRISMA 7 questionnaire, self-reported health, general practitioner clinical assessment, polypharmacy and Groningen Frailty Index. For a gait speed of <0.8 m/s, the sensitivity = 0.99 and specificity = 0.64. For the PRISMA 7, the sensitivity = 0.83 and specificity = 0.83. For the timed get-up-and-go test of 10 s, the sensitivity = 0.93 and specificity = 0.62. DTA was notably lower for all other index tests. All three studies were judged at unclear risk of bias.
Discussion: slow gait speed, PRISMA 7 and the timed get-up-and-go test have high sensitivity for identifying frailty. However, limited specificity implies many false-positive results which means that these instruments cannot be used as accurate single tests to identify frailty.
Keywords: diagnostic test accuracy; frailty; instruments; older people; sensitivity; specificity.
© The Author 2014. Published by Oxford University Press on behalf of the British Geriatrics Society. All rights reserved. For Permissions, please email: email@example.com.