Comparison of Frailty Indicators Based on Clinical Phenotype and the Multiple Deficit Approach in Predicting Mortality and Physical Limitation

<jats:sec> <jats:title>Objectives</jats:title> <jats:p>To compare three simple bedside tools based on frailty phenotypes with a Frailty Index using the multiple deficit approach in the prediction of mortality and physical limitation after 4 years.</jats:p> </jats:sec> <jats:sec> <jats:title>Design</jats:title> <jats:p>Cohort study.</jats:p> </jats:sec> <jats:sec> <jats:title>Setting</jats:title> <jats:p>Hong Kong, China.</jats:p> </jats:sec> <jats:sec> <jats:title>Pariticipants</jats:title> <jats:p>Four thousand men and women aged 65 and older living in the community who were ambulatory enough to attend the study center.</jats:p> </jats:sec> <jats:sec> <jats:title>Methods</jats:title> <jats:p> Interviewers obtained information regarding physical, psychological, and functional health; body mass index ( <jats:styled-content style="fixed-case">BMI</jats:styled-content> ), grip strength, blood pressure, and ankle brachial index were determined. Three clinical frailty scales based on the Fried phenotype (Cardiovascular Health Study ( <jats:styled-content style="fixed-case">CHS</jats:styled-content> ); Fatigue, Resistance, Ambulation, Illness, and Loss ( <jats:styled-content style="fixed-case">FRAIL</jats:styled-content> ); and Hubbard) and a frailty index ( <jats:styled-content style="fixed-case">FI</jats:styled-content> ) were constructed from these variables, and their ability to predict incident mortality and physical function limitations was compared using receiver operating characteristic ( <jats:styled-content style="fixed-case">ROC</jats:styled-content> ) curves. </jats:p> </jats:sec> <jats:sec> <jats:title>Results</jats:title> <jats:p> All tools predicted adverse outcomes. More participants were categorized into frail and prefrail categories using the <jats:styled-content style="fixed-case">CHS</jats:styled-content> than with the other two clinical scales. For all frailty measures, with increasing levels of frailty, the sensitivity fell and the specificity increased to greater than 90%; the area under the <jats:styled-content style="fixed-case">ROC</jats:styled-content> curve values were approximately 0.6. </jats:p> </jats:sec> <jats:sec> <jats:title>Conclusion</jats:title> <jats:p> Simple frailty scores are comparable with a multidimensional deficit accumulation <jats:styled-content style="fixed-case">FI</jats:styled-content> in predicting mortality and physical limitations. The newer <jats:styled-content style="fixed-case">FRAIL</jats:styled-content> , proposed for use in a clinical setting, is comparable with other existing short screening tools, as well as tools based on the multiple‐deficits model used for research settings. Addition of a physical performance measure to screening tools may increase predictive accuracy. </jats:p> </jats:sec>