EUVIMEDOpdag EUVIMED
ValidierungsstudieKohortenstudieOpen access

Development and Internal Validation of an Electronic Medical Record–Based Machine Learning Model for Frailty Prediction in Hospitalized Patients: Retrospective Cohort Study

Entwicklung und interne Validierung eines auf elektronischen Krankenakten basierenden Machine-Learning-Modells zur Frailty-Vorhersage bei hospitalisierten Patient:innen: retrospektive Kohortenstudie

Georgina Christine Laurea Szabo, Madison Bills, Matthew Williams, Annie Conway, Leesa Van Bochove, Gerry O'Callaghan

JMIR Medical Informatics · 2026-10-28 · 14 · e95478-e95478

DOI: 10.2196/95478 ↗

Auf einen Blick

Ein EMR-basiertes Machine-Learning-Modell sagte Frailty bei hospitalisierten Patient:innen anhand von Daten aus den ersten 72 Stunden voraus. Das XGBoost-Modell zeigte im Testdatensatz eine gute Vorhersageleistung.

Fragestellung

Lassen sich routinemäßig erhobene EMR-Daten aus den ersten 72 Stunden nach Aufnahme nutzen, um Frailty bei hospitalisierten Patient:innen vorherzusagen?

Methode

Retrospektive Kohortenstudie mit Routinedaten des Royal Adelaide Hospital; overnight admissions mit stationärem Aufenthalt von mindestens 72 Stunden zwischen Juli 2023 und Dezember 2024. Training/Test-Split 80/20; Vergleich mehrerer Machine-Learning-Modelle.

Ergebnisse

26.824 Aufnahmen, davon 9.427 (35 %) frail. XGBoost im Testdatensatz: F2-Score 0,755 (95 %-KI 0,741–0,768), AUROC 0,777 (95 %-KI 0,764–0,790), Recall 0,873 (95 %-KI 0,857–0,887); expected-to-observed ratio 0,996 (95 %-KI 0,963–1,028).

Einschränkungen

Geringere Leistung bei 24 Stunden nach Aufnahme; reduzierte Leistung bei Indigenous patients.

Schlussfolgerung der Autor:innen

Ein EMR-basiertes Frailty-Vorhersagemodell mit Daten aus 72 Stunden nach Aufnahme kann Frailty früh im Krankenhausaufenthalt erkennen und gezielte Interventionen sowie die Einsatzplanung unterstützen.

Kernaussagen

  • Frailty wurde mit HFRS >4 definiert.
  • XGBoost erreichte im Testdatensatz F2 0,755 und AUROC 0,777.
  • Leistung war nach 24 Stunden geringer; bei Indigenous patients reduziert.

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

Validierungsstudie in JMIR Medical Informatics, published online on 2026-10-28. Written by Georgina Christine Laurea Szabo and 5 co-authors. Institution: SA Health, Commission on Excellence and Innovation in Health, Adelaide, South Australia, Australia. Major MEDLINE topics: Electronic Health Records, Machine Learning, Frailty, Hospitalization. 21 references. The full text is freely available in PubMed Central.

Auf einen BlickAbstractSubjectsBibliografiske oplysningerCiteRelated studies

Abstract (original language)

Background

Frailty is prevalent among hospitalized patients and is associated with increased risk of complications, prolonged hospitalization, higher mortality, and health care costs. Existing frailty screening tools, including the Clinical Frailty Scale (CFS) and Hospital Frailty Risk Score (HFRS), have limitations related to subjectivity, resource intensity, or reliance on retrospective data, limiting their use in routine acute care. Early identification of frailty is critical to support targeted interventions and resource allocation, yet scalable approaches for predicting frailty early during hospital admission remain limited.

Objective

This paper evaluates whether routinely collected electronic medical record (EMR) data available within the first 72 hours of admission can be used to predict frailty in hospitalized patients.

Methods

A retrospective cohort study was conducted using routinely collected EMR data from the Royal Adelaide Hospital (RAH), including overnight admissions with a length of stay of at least 72 hours between July 2023 and December 2024. Frailty was defined using the HFRS (score >4). Predictors were derived from EMR data available within 72 hours of admission, including demographics, medications, laboratory results, nursing assessments, and medical and allied health consult orders. Data were split into training (21,436/26,824, 80%) and test (5388/26,824, 20%) sets. Multiple machine learning models were trained, with the F 2 -score as the primary optimization metric. Model performance was evaluated using the F 2 -score, recall, area under the receiver operating characteristic curve (AUROC), calibration, and clinical utility analyses.

Results

This study included 26,824 admissions, of whom 9427 (35%) were classified as frail. The final XGBoost (XGB) model achieved an F 2 -score of 0.755 (95% CI 0.741-0.768), an AUROC of 0.777 (95% CI 0.764-0.790), and a recall of 0.873 (95% CI 0.857-0.887) on the test dataset, with good calibration and an expected-to-observed ratio close to 1 (0.996, 95% CI 0.963-1.028). Key predictors included polypharmacy, laboratory results (including creatinine, albumin, and sodium), allied health consult orders, and nursing assessment data. Model performance was lower at 24 hours postadmission and improved by discharge.

Conclusions

An EMR-based frailty prediction model using data available at 72 hours postadmission can identify patients with frailty early in their hospital stay, supporting targeted interventions and operational planning. Reduced performance in Indigenous patients highlights the need for culturally validated frailty indices and ongoing evaluation of equity in predictive modeling.

Subjects

MeSH terms (MEDLINE)

Electronic Health RecordsMachine LearningFrailtyHospitalizationRetrospective StudiesHumansFemalePredictive Learning ModelsMaleAgedAged, 80 and overROC Curve

Author keywords

allied health dataclinical decision supportclinical prediction modelselectronic medical recordsfrailtyinpatientsnursing documentationsupervised machine learning

Authors and institutions

  • SA Health, Commission on Excellence and Innovation in Health, Adelaide, South Australia, Australia.
  • Central Adelaide Local Health Network, Royal Adelaide Hospital, 1 Port Road, Adelaide, South Australia, 5000, Australia, 61 0432831520.
  • School of Public Health, Adelaide University, Adelaide, South Australia, Australia.
  • Adelaide EpiCentre, Central Adelaide Local Health Network, Adelaide, South Australia, Australia.

Cite

APA

Laurea Szabo, G. C., Bills, M., Williams, M., Conway, A., Van Bochove, L., & O'Callaghan, G. (2026). Development and Internal Validation of an Electronic Medical Record–Based Machine Learning Model for Frailty Prediction in Hospitalized Patients: Retrospective Cohort Study. JMIR Medical Informatics, 14, e95478-e95478. https://doi.org/10.2196/95478

Vancouver

Laurea Szabo GC, Bills M, Williams M, Conway A, Van Bochove L, O'Callaghan G. Development and Internal Validation of an Electronic Medical Record–Based Machine Learning Model for Frailty Prediction in Hospitalized Patients: Retrospective Cohort Study. JMIR Medical Informatics. 2026;14:e95478-e95478. doi:10.2196/95478

BibTeX
@article{LaureaSzabo2026,
  author = {Laurea Szabo, Georgina Christine and Bills, Madison and Williams, Matthew and Conway, Annie and Van Bochove, Leesa and O'Callaghan, Gerry},
  title = {Development and Internal Validation of an Electronic Medical Record–Based Machine Learning Model for Frailty Prediction in Hospitalized Patients: Retrospective Cohort Study},
  journal = {JMIR Medical Informatics},
  year = {2026},
  volume = {14},
  pages = {e95478--e95478},
  doi = {10.2196/95478}
}
Start for free →