Prognostic Markers of Frailty and Fall Risks in Older Adults: a Longitudinal Cohort Study using Explainable AI

Read full text Summarise with VIRAT
Authors: Tiago Nobrega, Tiago Santos, Henrique Anjos, Bruno Gomes, Fabio Cunha, Pedro Oliveira, Rui Terenas Baptista, Andreia Isabel Nogueira Pizarro, Jorge Augusto Pinto Silva Mota, Daniel Moreira‐Gonçalves, Rui Henriques, Rafael S. Costa
Cited by: 0
References: 38

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Abstract Frailty is a geriatric syndrome that reflects a state of increased vulnerability to endogenous and exogenous stressors, exposing individuals to a higher risk of premature death and adverse health outcomes. This study aims at identifying novel determinants of frailty and refining screening tools to improve the prognostication of understudied endpoints. To this end, we established a new longitudinal cohort study comprising 2,862 participants and 6,855 observations across up to four assessment moments, incorporating measures of anthropometry, functional performance, cognition, quality of life, fall history, and sociodemography. We combine unsupervised clustering to explore profile heterogeneity with supervised prediction for falls, hospitalization, and handgrip strength, using explainability machine learning approaches to connect model outputs to clinically meaningful patterns. Outcome-agnostic clustering separates functional profiles that align with different fall burdens, while a complementary outcome-aware analysis provides an integrated stratification combining functional limitations and perceived risks. Predictive models achieve moderate and consistent discrimination for fall prediction (AUROC = 0.636–0.696), and explainability approaches consistently emphasize key drivers including handgrip strength, self-report assessments, and other results from functional tests. Handgrip regression attains MAE ≈ 3.6 kg ( R 2 ≈ 0.49), while a dedicated CatBoost sarcopenia classifier improves detection (AU-ROC = 0.798, recall = 0.792) at the cost of low precision, consistent with screening-oriented use. Overall, the results support the feasibility of explainable AI for actionable risk stratification in community assessments, while highlighting constraints related to missingness, class imbalance, and healthy volunteer bias.

Fields of Study & Subjects

Frailty in Older Adults Balance Gait and Falls Prevention Nutrition and Health in Aging

Publication details

Journal / source
medRxiv
Publication date
22 March 2026
Language
EN
Document type
Preprint