Frailty index deficit interchangeability: an empirical test using random survival forests in adults with cardiovascular disease
Frailty-Index-Defizit-Austauschbarkeit: ein empirischer Test mit Random Survival Forests bei Erwachsenen mit kardiovaskulärer Erkrankung
Age and Ageing · 2026-09-04 · 55 (9)
DOI: 10.1093/ageing/afag262 ↗Auf einen Blick
Bei 3.669 Erwachsenen mit kardiovaskulärer Erkrankung sagte ein 46-Item-Frailty-Index die Mortalität besser voraus als zusammengesetzte FI-Scores. Die Itemauswahl machte bei ausreichender Itemzahl kaum einen Unterschied.
Fragestellung
Spielt die Auswahl der FI-Items bei der Vorhersage der Mortalität bei Erwachsenen mit kardiovaskulärer Erkrankung eine Rolle?
Methode
Analyse von 3.669 Erwachsenen mit kardiovaskulärer Erkrankung aus NHANES 1999–2018; 46-Item-FI. Cox- und Random-Survival-Forest-Modelle für Gesamt- und CVD-Mortalität.
Ergebnisse
Modelle mit 46 einzelnen Items schnitten besser ab als zusammengesetzte FI-Ansätze für beide Endpunkte (C-index: 0,71 [0,69–0,74] vs. 0,64–0,67). Die Leistung der importance-ranked FIs erreichte bei etwa 10 Items ein Maximum.
Einschränkungen
Randomly built FIs converged with importance-ranked composites only by ~35 items (all-cause) and 20 items (CVD).
Schlussfolgerung der Autor:innen
Wenn genügend Items enthalten sind, hat die konkrete Zusammensetzung des FI nur einen geringen Einfluss auf die Vorhersage. Das Zusammenfassen von Items kostet prognostische Information, erhält aber die Praktikabilität.
Kernaussagen
- 46 einzelne Items sagten Mortalität besser voraus als zusammengesetzte FI-Scores.
- Der C-index lag bei 0,71 [0,69–0,74] vs. 0,64–0,67.
- Short importance-ranked FIs profitierten eher vom Vermeiden von „Dilution“ als von optimaler Itemwahl.
Deutsche Zusammenfassung: KI-gestützt aus dem Original-Abstract erstellt; alle Zahlen wurden automatisch mit dem Original abgeglichen. Maßgeblich ist der Original-Abstract.
At a glance
Journal article in Age and Ageing, published online on 2026-09-04. Written by Jack Quach and 4 co-authors. Institution: Faculty of Health, Dalhousie University, Halifax, Nova Scotia, Canada. Major MEDLINE topics: Cardiovascular Diseases, Frailty, Geriatric Assessment. 33 references. The full text is freely available (open access, CC BY-NC 4.0). Funded by: Scotia Scholars Doctoral Award, Killam Predoctoral Scholarship Level 2, Nova Scotia Graduate Scholarship.
Topics: Gerontologische Pflege
Abstract (original language)
Background
The frailty index (FI) assumes that deficits are interchangeable; the type of individual items matters less than the total count. Machine learning is increasingly used to create shorter FIs, often motivated by perceived barriers to collecting many items in clinical settings despite availability of electronic health records. We explored whether FI item selection matters when predicting mortality in adults living with cardiovascular disease (CVD).
Methods
We analysed 3669 adults living with CVD from the National Health and Nutrition Examination Survey (1999–2018) with a 46-item FI. Cox and random survival forest models predicted all-cause and CVD-specific mortality using individual items or composite FI scores. Items were ranked by permutation importance. At each item count k (1-46), we compared the top k items as individual predictors, the top k items combined into a single FI and 200 FIs each built from randomly selected k items.
Results
Models using 46 individual items outperformed composite FI approaches for both outcomes (C-index: 0.71 [0.69–0.74] vs 0.64–0.67). Importance-ranked FI performance peaked at ~10 items then declined. Randomly built FIs improved steadily, converging with the importance-ranked composites by ~35 items (all-cause) and 20 items (CVD). Analyses using imputation produced consistent results.
Conclusions
When enough items are included, which specific items make up the FI has negligible influence on prediction. Combining items into a single score discards some prognostic information, but this trade-off maintains the generalisability that makes the FI practical. Short importance-ranked FIs benefit from avoiding dilution rather than capturing an optimal item set, and any such selection is outcome-specific and sample-dependent.
Subjects
MeSH terms (MEDLINE)
Author keywords
Authors and institutions
- Faculty of Health, Dalhousie University, Halifax, Nova Scotia, Canada.
- School of Physiotherapy, Dalhousie University, Halifax, Nova Scotia, Canada.
- Division of Geriatric Medicine, Dalhousie University, Halifax, Nova Scotia, Canada.
- Department of Medicine, Dalhousie University, Halifax, Nova Scotia, Canada.
- Institute of Sport Exercise & Health, Division of Surgery & Interventional Science, University College London, London, UK.
ORCID profiles: Jack Quach, Olga Theou, Kenneth Rockwood, Joanna Blodgett
Funding
- Scotia Scholars Doctoral Award
- Killam Predoctoral Scholarship Level 2
- Nova Scotia Graduate Scholarship
- Michael Smith Foreign Study Supplements
- Canadian Institutes of Health Research Canada Graduate Scholarships
Cite
APA
Quach, J., Theou, O., Rockwood, K., Kehler, S., & Blodgett, J. (2026). Frailty index deficit interchangeability: an empirical test using random survival forests in adults with cardiovascular disease. Age and Ageing, 55(9). https://doi.org/10.1093/ageing/afag262
Vancouver
Quach J, Theou O, Rockwood K, Kehler S, Blodgett J. Frailty index deficit interchangeability: an empirical test using random survival forests in adults with cardiovascular disease. Age and Ageing. 2026;55(9). doi:10.1093/ageing/afag262
BibTeX
@article{Quach2026,
author = {Quach, Jack and Theou, Olga and Rockwood, Kenneth and Kehler, Scott and Blodgett, Joanna},
title = {Frailty index deficit interchangeability: an empirical test using random survival forests in adults with cardiovascular disease},
journal = {Age and Ageing},
year = {2026},
volume = {55},
number = {9},
doi = {10.1093/ageing/afag262}
}
EUVIMED