A nurse's one-minute judgment separates a 3.6% from a 27% chance of dying this year

🎧 Listen: Weekly · Nov 9, from 09:59.

If the first story says that frailty should change how you prepare an older patient, the obvious next question is how you know who is frail. Not in a sentence in the notes. On a scale you can write down, repeat in six months, and hand to a colleague.

This French cohort tested the cheapest instrument available: the nurse already sitting with the patient.

What they did

Six dialysis centers in Alsace enrolled 792 prevalent adults on maintenance dialysis: 51% in-center haemodialysis, 41% in self-care or nocturnal units, 8% on PD. Mean age 69, 58% men, 46% diabetic, 13% dialyzing through a catheter. Sixty dialysis nurses, about ten per center, rated each patient once on Rockwood's nine-point Clinical Frailty Scale between December 2016 and December 2018, judging the patient's usual state over the preceding fortnight. Training was a single videoconference with dialysis-specific vignettes. Follow-up for death and transplant ran to 36 months.

The comparator was the REIN predictive score, a French registry instrument built from age, sex, comorbidities, albumin and mobility to predict three-month mortality in incident patients aged 75 and over. Here it was applied to a prevalent cohort and grouped in tertiles.

Worth noting what the scale is not. It is not a questionnaire, a grip dynamometer or a gait-speed test. It is a judgment about what the patient can and cannot do, anchored on the point where a person needs help with personal care. Level 9 sits outside that dependence gradient altogether: it means terminally ill.

What it predicted

Just under a third of the cohort (30.6%) was frail, meaning CFS 6 to 9; 40.9% were vulnerable at 4 to 5, and 28.5% were well at 1 to 3. The groups looked as different as the labels suggest: frail patients had a mean age of 76 and 88% dialyzed in-center, while well patients averaged 58 and two thirds were in self-care units.

Over follow-up, 202 patients died and 85 were transplanted, treated as a competing event. Twelve-month mortality was 3.56% (95% CI 1.13 to 5.98) in the well group, 10.49% (7.15 to 13.84) in the vulnerable group, and 27.27% (21.65 to 32.90) in the frail group. At 24 months the figures were 9.78%, 21.30% and 45.87%.

In a multivariable Fine-Gray model adjusted for age, sex, diabetes, catheter use and the REIN score, each CFS point carried a subdistribution hazard ratio of 1.33 (1.19 to 1.49). The REIN score held its own at 1.05 per point (1.02 to 1.09). The two were independent of each other, which is the useful finding: a functional read adds information that a comorbidity count does not contain.

Discrimination was comparable rather than superior. The C-statistic at 12 months was 0.73 (0.68 to 0.79) for the CFS against 0.69 (0.63 to 0.75) for the REIN score, p = 0.20; at 24 months, 0.71 against 0.67, p = 0.13. So the honest claim is that a one-minute nurse rating matched a registry score built from labs and comorbidities, and added to it when both were in the model.

The catch

Reliability. In 509 patients rated independently by two nurses, agreement on the full nine-point scale was only fair, with a kappa of 0.35. Collapsed into the three categories it rose to 0.60, moderate to substantial. Nurse one rated the patient frailer in 144 cases, nurse two in 133.

That has a direct practical consequence. Treat the categories as the unit you act on and the per-point score as a trend to watch, not a boundary to make decisions at. The authors found a Youden-derived threshold of CFS 6 or more for 12-month mortality, and they are careful to call it a risk-enrichment cut-point rather than a decision rule. It needs external validation before anyone uses it to steer a patient toward or away from dialysis.

The other limits are structural. This is a prevalent cohort, so it carries survivorship bias and says nothing about the first months on dialysis, which is exactly when risk peaks. It is one French region, with no ethnicity data. The timing of the rating within the dialysis session was not standardised, and there was no test-retest.

Why it earns a place in the clinic anyway

Because the alternative is nothing. The scale costs a minute, needs no equipment, can be repeated at every multidisciplinary review, and the people best placed to rate it are already in the room three times a week. A patient drifting from 4 to 6 over a year is telling you something no laboratory value will, and that drift is a reasonable trigger for closer follow-up, rehabilitation, or a conversation about what dialysis is still achieving.

What is untested is whether using it systematically improves anything patients would notice, or shifts the dialysis-versus- conservative balance at all. The authors say so plainly. Use it as a prompt, not a verdict.

The source: Ingwiller M, Chantrel F, Krummel T, et al. Kidney Med. 2026;8(11):101510. Read the original

🎧 Listen: Weekly · Nov 9, from 09:59.

Physician-led, AI-assisted. Dr. Wael Hussein chooses every item NephSpace covers and reviews and edits every story before it is published. The first draft was written in our own words with AI (Claude, by Anthropic) from the full text of the paper and checked against our reading notes; Dr. Hussein then reviewed and edited it. For education and information only, not medical advice. Please talk to your own physician or care team about your health. Clinicians: read the original before changing practice. See our disclaimer and disclosures. Spotted an error? Tell us.