'Frail' means different things on different tools

Clinical review: pending. The numbers on this page were checked against their sources; a clinician has not yet reviewed the page.

Frailty tools set their "frail" lines in different places, so the same person can be frail on one tool and not on another.

Introduction

"Frail" is not one fixed line. Each tool sets its own cut-point, and those cut-points pick out different groups of people. In one study of adults aged 50 to 104 in 11 European countries (SHARE), the share of people labelled frail ranged from 6% to 44% depending on the tool. Only 2.4% were frail on every tool.

This guide covers three things. It shows where each tool's "frail" line sits on a common scale, how prevalence changes between tools in the same people, and what the results from our own tools mean.

Sources: Sison 2023 [1], Theou 2013 [2], Frailty Science (Johns Hopkins) [3].

Where each tool's "frail" line sits on a frailty index

US researchers rebuilt nine frailty tools from survey answers and tests for the same 7,070 people. They were Medicare beneficiaries aged 65 and over living in the community (National Health and Aging Trends Study, 2015). The researchers lined up each tool with a 40-item frailty index by matching scores that sat at the same rank in the group. The chart and table show the frailty index that matched each tool's own "frail" cut-point.

A Clinical Frailty Scale score of 5, the scale's usual "frail" cut-point, matched a frailty index of 0.21. Going the other way, a frailty index of 0.25 matched a CFS score of about 5.3. At these cut-points, the CFS and the frailty index labelled similar shares of people frail. VES-13 and the Groningen Frailty Indicator labelled more people frail, and the other five tools fewer. On our CFS tools, level 4 is named "Living With Very Mild Frailty". This study counted only 5 or more as frail.

Source: eFrailty crosswalk data, from Sison 2023 (NHATS, US). CFS 8 and 9 are not covered, because terminal illness was not recorded in the survey.

How to use these numbers

• These are matches between groups, not conversions for one patient. A conversion can misclassify an individual, so use it only as a rough guide alongside clinical judgement.
• The authors say the crosswalk applies only to older adults living in the community.
• At group level the matches held up. When people were grouped as low, moderate or high risk on every tool through the crosswalk, 4-year mortality was similar across tools within each group (men and women aged 75 to 79).
• The VES-13 cut-point of 3 was first set to identify "vulnerable" older people at higher risk of death or functional decline. Sison and colleagues used it as the "frail" line.
• To look up the published group-level crosswalk, see the eFrailty frailty tool crosswalk, which is built on the same study. It carries the same caution about individual patients.

Sources: Sison 2023 [1], eFrailty crosswalk [4], Saliba 2001 [5].

Frailty index cut-points vary too

There is no single agreed frailty index cut-point. In 2013, Statistics Canada researchers noted that earlier community studies had used 0.21, 0.25 and 0.35. Two Canadian population studies used "above 0.21" as frail and 0.45 or more as "most frail". Others, including the US study above, used 0.25 or more. Sison and colleagues placed 0.25 at the move from pre-frail to mildly frail.

Sources: Hoover 2013 [6], Rockwood 2011 [7], Song 2010 [8], Sison 2023 [1], Best 2025 [9].

Same people, different prevalence

When researchers apply several tools to the same people, the share labelled frail changes a lot.

The tools also pick out people with different problems. In the NHANES study, 97.8% of people who were frail on the phenotype had difficulty with basic activities of daily living (ADL disability). The figure was 56.6% for people frail on the frailty index. The SHARE authors concluded that the scales "capture related but distinct groups" (Theou 2013). Frailty researchers at Johns Hopkins advise against treating tools as interchangeable, especially tools built on different ideas of frailty.

Sources: Theou 2013 [2], Blodgett 2015 [10], Afilalo 2017 [11], Sze 2019 [12], Frailty Science (Johns Hopkins) [3].

The eFI2 is a different kind of index

The UK electronic frailty index 2 (eFI2) also runs from 0 to 1, but it is built differently from a standard frailty index. The original eFI (2016) counts 36 deficits, each weighted equally. The eFI2 is a prediction model. It weights 36 factors by how strongly they predicted four events within one year: a new home care package, hospital admission for a fall or fracture, a move into a care home, or death. It was developed in 78,760 people aged 65 and over in Bradford, England, and tested in 660,417 in Wales, using 2018 records. In the Welsh data it predicted these events better than the original eFI (C-index 0.723 vs 0.687). Home care packages are not recorded in the Welsh data.

The British Geriatrics Society advises using the eFI2 to flag people who may be mildly, moderately or severely frail. A brief clinical assessment, such as the Clinical Frailty Scale, should then confirm it.

Populations: original eFI, UK primary care patients aged 65 to 95. eFI2, UK adults aged 65 and over.

The eFI2 values in the table are the British Geriatrics Society's rounded figures. The paper's exact cut-points are 0.0857, 0.1624 and 0.2392. The original eFI bands are four equal-width steps from 0 to the 99th-centile score (0.49). Because eFI2 items are weighted, its numbers are not on the same footing as a standard frailty index. For example, in the CARE75+ cohort a research frailty index of 0.25 ("any degree of frailty" in that study) matched an eFI2 of about 0.086.

Sources: Best 2025 [9], Clegg 2016 [13], British Geriatrics Society [14].

What Fit-to-Frail's categories mean

Fit-to-Frail's five categories (Poor, Fair, Average, Good and Excellent) are not frailty grades. They show where a person's frailty index falls among people of the same age and sex in the Canadian Longitudinal Study on Aging (CLSA), ages 45 to 85. The percentile is the share of those people whose frailty index is higher, so a higher percentile is better. People younger than 45 are compared with age 45, and people older than 85 with age 85. The comparison uses the norms method of Pérez-Zepeda and colleagues (2021).

Because the comparison is by age, the same frailty index can fall in a different category at a different age. For men, for example, some frailty index values count as Poor at age 50 but Excellent at age 85. So do not read Poor as "frail" or Excellent as "not frail". When you need a measure of frailty, use the frailty index itself, which the result also shows. Frailty index cut-points from other studies came from other item lists and other populations.

The Pictorial Fit-Frail Scale (beta) gives a continuous score from 0 to 43. It has no official categories, so our tool shows none. It is not one of the tools in the crosswalk above.

Sources: Pérez-Zepeda 2021 [15], Sison 2023 [1], Hoover 2013 [6], Song 2010 [8]; the Fit-to-Frail Index Calculator and Pictorial Fit-Frail Scale pages on this site.

A note for handover

When you hand over, name the tool and give the score, not just the word "frail", for example "Clinical Frailty Scale 5" or "frailty index 0.28". Tools set their "frail" lines at different levels, so the word alone does not tell the next clinician how frail the person is. If you must compare across tools, treat any conversion as a rough guide for groups, not a measurement of one person.

Sources: Sison 2023 [1], Theou 2013 [2], eFrailty crosswalk [4], Frailty Science (Johns Hopkins) [3].

References

1. Sison SDM, Shi SM, Kim KM, Steinberg N, Jeong S, McCarthy EP, et al. A crosswalk of commonly used frailty scales. J Am Geriatr Soc. 2023;71(10):3189-98. doi:10.1111/jgs.18453. PMID: 37289180.
2. Theou O, Brothers TD, Mitnitski A, Rockwood K. Operationalization of frailty using eight commonly used scales and comparison of their ability to predict all-cause mortality. J Am Geriatr Soc. 2013;61(9):1537-51. doi:10.1111/jgs.12420. PMID: 24028357.
3. Frailty Science, Johns Hopkins Older Americans Independence Center. Guidance on selecting a frailty assessment instrument [Internet]. [cited 2026 Oct 12]. Available from: https://frailtyscience.org/frailty-assessment-instruments/
4. eFrailty. Frailty tool crosswalk [Internet]. [cited 2026 Oct 12]. Available from: https://efrailty.hsl.harvard.edu/FrailtyConversion.html
5. Saliba D, Elliott M, Rubenstein LZ, Solomon DH, Young RT, Kamberg CJ, et al. The Vulnerable Elders Survey: a tool for identifying vulnerable older people in the community. J Am Geriatr Soc. 2001;49(12):1691-9. doi:10.1046/j.1532-5415.2001.49281.x. PMID: 11844005.
6. Hoover M, Rotermann M, Sanmartin C, Bernier J. Validation of an index to estimate the prevalence of frailty among community-dwelling seniors. Health Rep. 2013;24(9):10-7. PMID: 24258362. Available from: https://www150.statcan.gc.ca/n1/pub/82-003-x/2013009/article/11864-eng.htm
7. Rockwood K, Song X, Mitnitski A. Changes in relative fitness and frailty across the adult lifespan: evidence from the Canadian National Population Health Survey. CMAJ. 2011;183(8):E487-94. doi:10.1503/cmaj.101271. PMID: 21540166.
8. Song X, Mitnitski A, Rockwood K. Prevalence and 10-year outcomes of frailty in older adults in relation to deficit accumulation. J Am Geriatr Soc. 2010;58(4):681-7. doi:10.1111/j.1532-5415.2010.02764.x. PMID: 20345864.
9. Best K, Shuweihdi F, Alvarez JCB, Relton S, Avgerinou C, Nimmons D, et al. Development and external validation of the electronic frailty index 2 using routine primary care electronic health record data. Age Ageing. 2025;54(4):afaf077. doi:10.1093/ageing/afaf077. PMID: 40163740.
10. Blodgett J, Theou O, Kirkland S, Andreou P, Rockwood K. Frailty in NHANES: comparing the frailty index and phenotype. Arch Gerontol Geriatr. 2015;60(3):464-70. doi:10.1016/j.archger.2015.01.016. PMID: 25697060.
11. Afilalo J, Lauck S, Kim DH, Lefèvre T, Piazza N, Lachapelle K, et al. Frailty in older adults undergoing aortic valve replacement: the FRAILTY-AVR study. J Am Coll Cardiol. 2017;70(6):689-700. doi:10.1016/j.jacc.2017.06.024. PMID: 28693934.
12. Sze S, Pellicori P, Zhang J, Weston J, Clark AL. Identification of frailty in chronic heart failure. JACC Heart Fail. 2019;7(4):291-302. doi:10.1016/j.jchf.2018.11.017. PMID: 30738977.
13. Clegg A, Bates C, Young J, Ryan R, Nichols L, Ann Teale E, et al. Development and validation of an electronic frailty index using routine primary care electronic health record data. Age Ageing. 2016;45(3):353-60. doi:10.1093/ageing/afw039. PMID: 26944937.
14. British Geriatrics Society. Electronic Frailty Index 2 (eFI2) [Internet]. Nimmons D, author. British Geriatrics Society; 2026 Sep 17 [cited 2026 Oct 12]. Available from: https://www.bgs.org.uk/electronic-frailty-index-2-efi2
15. Pérez-Zepeda MU, Godin J, Armstrong JJ, Andrew MK, Mitnitski A, Kirkland S, et al. Frailty among middle-aged and older Canadians: population norms for the frailty index using the Canadian Longitudinal Study on Aging. Age Ageing. 2021;50(2):447-56. doi:10.1093/ageing/afaa144. PMID: 32805022.