Frailty Index

The frailty index (FI), developed by our group, is one of the leading paradigms for understanding frailty at population, personal, organ, and tissue levels (with cellular deficit evaluation in development). The frailty index is used in routine care all over the world, including in an electronic form in the National Health Service in England. To quantify the level of frailty of an individual, the frailty index approach focuses first on the number of their health deficits, and then on the nature of those deficits.

One of the strengths of this approach is that a frailty index can be constructed using existing clinical and population-based data: not every frailty index needs to include the same items to achieve closely comparable estimates in prevalence and rate of change. The frailty index is replicable across different databases because, as a state variable, it takes advantage of the high redundancy of the human organism.

Other advantages of this approach are:

• it can be applied to almost any health dataset
• it is a comprehensive assessment of health
• it provides a continuous score from fitness to frailty
• it is more sensitive to health changes when compared to other tools
• it does not have a ceiling or floor effect

The main challenge is that it needs at least 30 health variables to be included; if data are not already collected, 30 or more variables can be onerous to collect.

Frailty indices have been constructed in various population-based studies worldwide such as the American National Health and Nutrition Examination Survey, the Survey of Health and Ageing and Retirement in Europe, the Australian Longitudinal Study of Aging, the Beijing Longitudinal Study of Aging, the Mexican Health and Aging Study, and the Study on Global Ageing and Adult Health (China, Ghana, India, Mexico, Russia and South Africa). Frailty in Canada has been extensively investigated using data from the Canadian Study of Health and Aging (CSHA) and the National Population Health Survey (NPHS).

Our group has published over 100 papers on frailty using the CSHA and NPHS data. These studies were vital in setting the foundation for understanding the nature and quantification of frailty.

Specifically, they have shown that:

• the frailty index strongly predicts adverse outcomes including mortality, disability, and cognitive decline and it outperforms chronological age as a predictor of these outcomes;
• the distribution of frailty indices in community-living samples are left skewed (the majority of people are healthy with low frailty index scores);
• individuals do not exceed a limit of 0.7 in the frailty index (this appears to reflect the degree to which health problems can be tolerated);
• individual frailty index do not just increase with age, they can also decrease, reflecting improved health; and
• males have lower mean frailty index values than females of the same age, whereas females show better mean survival than males with the same frailty index value.

Frailty Indices based on the CLSA

We developed five frailty indices (FI) that can be used by researchers who have access to data from the Canadian Longitudinal Study on Aging (CLSA). One FI uses data from the pooled cohort (the Comprehensive and Tracking cohorts combined; FI-52). One uses data from the Tracking cohort (FI-66). The remaining three use the Comprehensive cohort (FI-Self Report, FI-Blood, FI-Examination). 

We also developed a standard frailty index based on the CLSA measures (FI-62) that can be used by health care professionals to measure frailty in their patients, by researchers who will collect their own frailty data, or by members of the public who want to calculate their own FI score.

FIs for researchers with access to CLSA:

Pooled cohort

Using standard methodology, we created a 52-item frailty index (FI-52) using baseline data from CLSA collected in the pooled tracking and comprehensive cohorts between 2011 and 2015 (Pérez-Zepeda et al., 2021). 

The items that comprise the CLSA FI-52 reference self-rated health, chronic conditions, activities of daily living, instrumental activities of daily living, cognitive function, and mental health. The proportion of health problems present in a given individual was used to estimate their frailty index score. 

Researchers who have access to CLSA data and who would like to replicate the CLSA FI-52 can use our syntax [PDF - 24KB]. Note that this syntax is formatted for STATA 16.0 and will work using only the baseline pooled cohort data.

Using weighted frailty index scores and the participants’ age and sex, researchers can use our syntax to calculate normative data for frailty in Canadian males and females over the age of 45. Our work used the first sampling weights provided by CLSA (2017), but can be adapted to newer versions of sampling weights (e.g. 2020).

Tracking cohort

Using the same methodology as the pooled cohort, we created a 66-item frailty index (FI-66) for the CLSA tracking cohort. The CLSA FI-66 includes the same items as the FI-52 plus an additional 13 mobility items and 1 weight loss item that were only collected in the tracking cohort (i.e. these items are not collected in the comprehensive cohort). 

Researchers who have access to CLSA data may replicate this FI using our syntax. They can also use a separate syntax to create normative frailty index data for males and females over the age of 45. 

Comprehensive cohort

We have developed (Blodgett et al., in press) three additional FIs for the CLSA comprehensive cohort:

1) The CLSA FI-Self Report – This FI was adapted from the CLSA FI-52 constructed for use with the pooled cohort. It consists of 48 self-reported items. Four cognitive scores from the CLSA FI-52 (see pooled cohort above) were excluded.

2) The CLSA FI-Blood – This FI consists of 23 biomarkers from chemistry and hematological reports. 

3) The CLSA FI-Examination – This FI consists of 47 items across six domains, including: physical performance, cognition, cardiac, anthropometric, spirometry, and hearing and vision. Due to the lack of clinical reference ranges and the informative variability in performance across most assessments, normative coding of many items was used.

FI for researchers, health care professionals, or members of the public:

We have developed a questionnaire for measuring an individual’s frailty index score using 62 items (CLSA FI-62). This is based on the CLSA FI-66 described above (see tracking cohort above) but excludes four cognitive tests that could not be converted to questionnaire form. Please note that all of our CLSA FIs combine five arthritis-related questions into two FI items. For this reason, our questionnaire includes 62 FI items derived from 65 questions. 

The questionnaire can be administered by researchers or health care professionals – or self-administered – to identify an individual’s frailty level. A printable version [PDF - 202KB] of the questionnaire is available with an accompanying scoring sheet [coming soon]. This allows users to calculate the frailty index score by hand; the resulting score can be compared to the normative frailty data of Canadians of the same age and sex available using this table. We have also created an online tool that automatically generates a frailty index score upon completion of the questionnaire.

Disclaimer: This tool is meant for education and research purposes only and should not be used for diagnosis or care planning, except under the direction of a health care provider. 

References

Theou O, Haviva C, Wallace L, Searle SD, Rockwood K. How to construct a frailty index from an existing dataset in 10 steps. Age Ageing. 2023;52(12):afad221.

Blodgett JM, Pérez-Zepeda MU, Godin J, Kehler DS, Andrew MK, Kirkland S, Rockwood K, Theou O. Frailty indices based on self-report, blood-based biomarkers and examination-based data in the Canadian Longitudinal Study on Aging. Age Ageing. 2022; 51(5): afac075

Pérez-Zepeda MU, Godin J, Armstrong JJ, Andrew MK, Mitnitski A, Kirkland S, Rockwood K, Theou O. 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): 447-456.

Searle SD, Mitnitski A, Gahbauer EA, Gill TM, Rockwood K. A standard procedure for creating a frailty index. BMC Geriatr. 2008 Sep 30;8:24.

Jones DM, Song X, Rockwood K. Operationalizing a frailty index from a standardized comprehensive geriatric assessment. J Am Geriatr Soc. 2004 Nov;52(11):1929-33.