Recognising frailty in a more routine situation
A range of tests for identifying frailty are available, but the accuracy of these is uncertain. A review was undertaken to investigate the diagnostic accuracy of some simple tests for identifying frailty. The full detail of the review is available 11 but in summary, it searched for all studies that compared simple tests for identifying frailty (e.g. walking speed, grip strength, simple questionnaires) against a phenotype model, cumulative deficit model or comprehensive geriatric assessment.
The review found three studies that investigated seven simple methods for identifying frailty; these were:
- PRISMA 7 Questionnaire - which is a seven item questionnaire to identify disability that has been used in earlier frailty studies and is also suitable for postal completion. A score of > 3 is considered to identify frailty.
- Walking speed (gait speed) - Gait speed is usually measured in m/s and has been recorded over distances ranging from 2.4m to 6m in research studies. In this study, gait speed was recorded over a 4m distance.
- Timed up and go test - The TUGT measures, in seconds, the time taken to stand up from a standard chair, walk a distance of 3 metres, turn, walk back to the chair and sit down.
- Self-Reported Health - which was assessed, in the study examined, with the question 'How would you rate your health on a scale of 0-10'. A cut-off of < 6 was used to identify frailty.
- GP assessment - whereby a GP assessed participants as frail or not frail on the basis of a clinical assessment.
- Multiple medications (polypharmacy) - where frailty is deemed present if the person takes five or more medications.
- The Groningen Frailty Indicator questionnaire - which is a 15 item frailty questionnaire that is suitable for postal completion. A score of > 4 indicates the possible presence of moderate-severe frailty.
Slow walking speed (less than 0.8m/s or taking more than 5 secs to walk 4m); the PRISMA 7 questionnaire and the timed-up-and-go test (with a cut off score of 10 secs) had very good sensitivity but only moderate specificity for identifying frailty. This means that there are many fitter older people who will have a positive test result (false positives). For example, only one in 3 older people (over 75 years) with slow walking speed has frailty.
However, the accuracy of a test is related to the prevalence of a condition in a population. For example, older people who attend outpatient clinics, receive social services assessments or require ambulance crew attendance are more likely to have frailty. This means that the tests are likely to be more accurate in these situations, which supports a case finding approach to identifying frailty. The BGS therefore recommends, as the most suitable tests, the use of gait speed (taking more than 5 seconds to walk 4 m using usual walking aids if appropriate) or the timed up and go test (with a cut off score of 10s to get up from a chair, walk 3m, turn round and sit down). The PRISMA 7 questionnaire (with a cut-off score of >3) could be considered as an alternative for self-completion, including as a postal questionnaire.
Use of a two-step approach to diagnosis (for example the Prisma questionnaire followed by a gait speed test) would potentially improve accuracy but requires further investigation. Where possible, the BGS also advocates a brief clinical assessment to confirm the presence of frailty. This would help exclude some false positives (e.g. otherwise fit older people with isolated knee arthritis causing slow gait speed).
It is inappropriate to use the (Rockwood) Clinical Frailty Scale (CFS) 7 as a method of identifying frailty without a formal clinical assessment. The CFS was designed to be used to measure severity of frailty after a comprehensive geriatric assessment. It is not validated for measuring improvement in individuals after an acute illness for example.
Health and social care staff will therefore need to be familiar with the tests which might be used for recognising frailty and should be trained to use them.
If an older person is ill and there is reason to believe that their illness will affect their gait speed or ability to get up from a chair, the PRISMA 7 questionnaire based on their usual health status will identify those who are likely to have frailty (scoring yes to 3 or more questions). It can also be used for self-completion, including as a postal questionnaire.
There is some evidence that grip strength (using a hand held dynamometer) may be useful in situations where it is not feasible for the patient to get up and walk. However this measurement has not yet been tested for diagnostic accuracy of frailty 12.
In outpatient surgical settings, there is a lack of consensus on which tool should be used to identify frailty. Gait speed may help predict adverse outcomes, however, evidence is emerging for the use of the Edmonton Frail Scale 13. The strengths of this tool include brevity, clinical feasibility and identification of aspects of frailty amenable to preoperative optimisation (e.g. cognition, nutrition).