LAUSR

vestigators has flourished, yet controversy exists regarding the ascertainment of frailty syndrome among older patients. The tools to assess frailty are numerous, and some of these tools may result in misclassification of frailty status in some patients.1 Kundi et al2 used administrative claims to identify frail individuals and evaluated whether this assessment of frailty status predicted outcomes among older patients with acute myocardial infarction, heart failure, or pneumonia. The investigators used the Hospital Frailty Risk Score (HFRS), a clustering of diagnoses that were identified a priori using International Statistical Classification of Diseases and Related Health Problems, Tenth Revision codes, as markers for frailty.3 While HFRS may be a useful tool when using administrative codes to study older populations, the agreement between HFRS with measured frailty is quite low. Gilbert et al3 reported that the κ statistic for agreement was only 0.22 (95% CI, 0.15-0.30) when comparing the HFRS with a binary definition of frailty using the Fried frailty phenotype criteria (3 or more items present). While we believe the results of Kundi et al2 are useful for stratification and risk prediction, assessment of frailty syndrome using administrative data should use a better-validated tool.1,4