- Design
- prospective observational diagnostic accuracy study, two blinded assessors
- Population
- 200 patients aged 70 or over undergoing coronary angiography, mean age 77.3 years
- Primary outcome
- accuracy of physician eyeballing against the Fried frailty test
- Effect
- sensitivity 93.9%, NPV 98.5%, specificity 80.8%, PPV 49.2%, AUC 0.87
Two cardiologists independently assessed 200 patients aged 70 or over referred for coronary angiography, blinded to each other: once with a formal Fried test, once by eyeballing scored on a Fried-like scale. Mean age was 77.3 years and frailty prevalence by Fried criteria was 16.5%.
Eyeballing had sensitivity 93.9% and negative predictive value 98.5%, with specificity 80.8%, positive predictive value 49.2% and area under the curve 0.87. Specificity was lower in women and in hospitalised patients. Frailty by Fried criteria was associated with higher six-month mortality and rehospitalisation, as expected.
Those numbers give a clean rule with two halves. A clinician who looks at an older patient and judges them not frail is almost always right - a negative predictive value of 98.5% is as good as most tests used to exclude anything. But when eyeballing says frail, it is wrong about half the time, and that is exactly the direction where the consequences fall: a patient wrongly labelled frail may be steered away from revascularisation, or towards a conservative pathway, on an impression. So use the impression to skip the formal test when it is reassuring, and do the formal test whenever it is not. In women and in hospitalised patients, be more sceptical of the impression still.
- Accept a confident impression of not frail without a formal test in this setting
- Always confirm suspected frailty with a structured instrument before it changes management
- Be more cautious with the impression in women and in inpatients, where specificity was lower
- Do not let a frailty label alone exclude a patient from revascularisation
- Record which assessment was used, so the next clinician knows what the label rests on
Why it matters
Frailty judged by impression is already deciding who gets offered invasive cardiology, and it is right about half the time when it says yes.
Don't overread it
A single-centre study of 200 patients referred for angiography; the accuracy figures do not transfer to settings where frailty is more common.
The statistics, in plain English
A negative predictive value of 98.5% depends heavily on frailty being uncommon here at 16.5% - in a population where half are frail, the same test would miss far more. The positive predictive value of 49.2% means a coin toss: half the patients eyeballed as frail were not frail on formal testing. Area under the curve of 0.87 is good overall discrimination, but the asymmetry between ruling in and ruling out is what matters at the bedside, and 200 patients at one centre is a small base for the subgroup claims about women and inpatients.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for geriatrics, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free