Frailty in older adults: moving from measurement to management

Incorporating routine assessment of frailty into health care would benefit both older people and the health system

The past two decades have seen a tremendous research effort dedicated to defining, measuring and validating the frailty construct. Large cohort studies of older adults living in the community have consistently found that frail people are at risk of a range of adverse outcomes, including death, disability, and institutionalisation.1 More recently, there has been a move from population‐based cohort studies of frailty to analyses of data collected during routine clinical encounters. In this issue of the MJA, Khadka and colleagues2 contribute to this body of translational research with a large retrospective cohort study of community‐dwelling Australians undergoing Aged Care Assessment Program (ACAP) eligibility assessment.

Khadka and his co‐authors aimed to validate a 44‐item frailty index that they derived, using standardised methodology, from data routinely collected during ACAP assessments.2 In studies of community‐dwelling adults, frailty indexes typically have skewed (gamma) distributions, indicating a heterogeneous but largely robust population.3 However, Khadka and colleagues found their frailty index had a normal distribution (mean value, 0.20), characteristic of a more homogeneous and frail population.3 This finding was not surprising, as their participants were older than in other studies (mean age, 82 years) and were being assessed for community and residential care services. About 30% of the study population were frail at assessment (frailty index values greater than 0.25). However, the frailty index limit (indicated by the 99th percentile) was only 0.36. A submaximal limit of about 0.70 is typical for large population‐based samples,4 although a British study based on routine primary care data and with a similarly sized (but younger and more robust) sample also derived a frailty index with a lower than expected limit (0.49).5 Khadka and his co‐authors2 acknowledge their unexpected finding and briefly discuss the composition of their frailty index. However, the reasons for the low frailty index limit are not determined. Nevertheless, the predictive validity of their frailty index was clear: frailer adults were more likely to die or be admitted to permanent residential aged care at one, three, and five years after assessment.2

By stratifying their data by two‐year cohorts, Khadka and colleagues2 found that the frailty (but not the age) of community‐dwelling Australians undergoing aged care eligibility assessment was increasing. They hypothesise that this reflects a trend toward older adults opting to remain in their own homes longer.2 This novel finding suggests that incorporating frailty assessment into clinical encounters can yield important insights into the evolving care needs of our ageing population.

Frailty indexes derived from the interRAI Home Care assessment (a standardised assessment of community and permanent residential care needs of community‐dwelling older adults) have been reported for Switzerland in 20176 and New Zealand in 2018.7 Both studies ascertained that deriving a frailty index from routinely collected data was feasible and that the frailty index predicted adverse outcomes, including death, hospitalisation, falls, and permanent residential aged care. The frailty indexes in these smaller studies had normal6 or nearly normal7 distributions, slightly higher mean values of 0.246 and 0.27,7 and submaximal limits greater than 0.60. An evaluation of differences between assessment instruments and frailty index derivation, as well as between‐country differences in eligibility criteria for aged care needs assessment, may shed light on the divergent characteristics of these frailty indexes and that of Khadka and his co‐authors.2

Khadka and colleagues2 have shown that frailty assessment can be incorporated into the Australian health care system. We can learn from the United Kingdom, where routine frailty assessment with a validated tool (typically the electronic frailty index, automatically generated for all adults aged 65 years or more attending general practices) is supported by National Institute for Health and Care Excellence guidelines8 and the National Health Service England.9 The best frailty tool and the best source of data will continue to be debated, but the more important question is whether routine frailty assessment leads to better outcomes for older people.

Routinely quantifying frailty could achieve two key outcomes. Firstly, it could reduce the risk of adverse outcomes for individual patients by triggering tailored assessments and influencing treatment decisions. For example, Comprehensive Geriatric Assessment and multi‐disciplinary management are known to be beneficial for frail older inpatients,10 whereas invasive interventions11 and polypharmacy12 are associated with harm. Secondly, routine assessment of frailty could provide context‐specific, up‐to‐date, cross‐sectional, and longitudinal data that could guide the allocation of health care resources and inform health care policies that affect older people. Big data analyses of frailty in relation to national health care decision‐making are huge undertakings, but smaller scale evaluations, at the community service or hospital levels, would be an important step forward.

Our colleagues overseas have led the way defining frailty and translating its measurement into clinical practice. The ultimate aim is that routine frailty assessment leads to interventions that achieve better outcomes for frail older people. Perhaps Australian researchers could take the lead in putting what we have learnt about frailty into practice.