EUVIMEDOpdag EUVIMED
Validierungsstudie

Explainable AI to Guide the Design of Interpretable Prediction Models From High-Dimensional Data, Applied to Falls and Related Injuries in Adults Aged 50 Years or Older: Algorithm Development and Validation

Erklärbare KI zur Anleitung des Designs interpretierbarer Vorhersagemodelle aus hochdimensionalen Daten, angewendet auf Stürze und verwandte Verletzungen bei Erwachsenen ab 50 Jahren: Algorithmusentwicklung und Validierung

Junyu Yan iD, Damian Machlanski iD, Kurt Butler iD, Panagiotis Dimitrakopoulos iD, Ewen M Harrison iD, Bruce Guthrie iD, Sotirios A Tsaftaris iD

Journal of Medical Internet Research · 2026-11-12 · 28 · e105364

DOI: 10.2196/105364 ↗

Auf einen Blick

Ein erklärbares KI-gestütztes Empfehlungssystem verbesserte die interne Validierung eines Cox-Modells zur 1-Jahres-Risikoabschätzung für Stürze oder verwandte Verletzungen leicht. Die Transparenz des Modells blieb erhalten.

Fragestellung

Kann ein erklärbares KI-basiertes Recommender-System helfen, interpretierbare Vorhersagemodelle für das 1-Jahres-Risiko von Stürzen oder verwandten Verletzungen aus hochdimensionalen Daten zu verbessern?

Methode

Entwicklung und interne Validierung mit Daten aus DataLoch; Erwachsenen ab 50 Jahren. Random survival forest mit SHAP, Empfehlungen für einen Cox-Ansatz; 80/20-Teilung; Vergleich mit mehreren Cox-Modellen; zusätzlich 3 öffentliche Datensätze.

Ergebnisse

220.341 Patient:innen; 8.489 (3,85 %) mit FRI innerhalb 1 Jahres. C-index 0,719 vs. 0,715; Differenz 0,004 (95 %-KI 0,001–0,007). Bei 409 Features C-index 0,722 vs. 0,713 und bessere Kalibrierung.

Schlussfolgerung der Autor:innen

Die vorgeschlagene explorative KI-Empfehlung zeigt das Potenzial von erklärbarer KI, den Entwurf transparenter Vorhersagemodelle zu unterstützen.

Kernaussagen

  • 1-Jahres-Risiko für FRI bei Erwachsenen ab 50 Jahren modelliert
  • SHAP-basierte Empfehlungen für Cox-Modelle abgeleitet
  • Leichte Verbesserung von C-index und Kalibrierung

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 Journal of Medical Internet Research, published online on 2026-11-12. Written by Junyu Yan and 6 co-authors. Institution: School of Engineering, University of Edinburgh, Edinburgh, United Kingdom. Major MEDLINE topics: Accidental Falls, Artificial Intelligence, Algorithms, Wounds and Injuries. 64 references.

Auf einen BlickAbstractSubjectsBibliografiske oplysningerCiteRelated studies

Abstract (original language)

Background

Transparent prediction models, such as Cox regression, remain the standard for clinical risk prediction. However, they are challenging to design optimally by hand when tens or even hundreds of features require selection, transformation, or interaction modeling. While complex machine learning models offer high performance, their “black-box” nature limits the clinical trust, transparency, and interpretability required for decision-making.

Objective

This study aimed to develop an explainable AI-based recommender that provides data-driven feature-related recommendations that, when incorporated, improve the predictive performance of existing interpretable statistical models on high-dimensional data, and to demonstrate it by developing and internally validating a Cox model for the 1-year risk of falls or related injuries (FRIs).

Methods

We used linked primary and secondary care records from DataLoch (Lothian, Scotland) for adults aged 50 years or older registered with a contributing general practice between January 2015 and February 2020. The outcome was the first FRI within 1 year of the index date. A random survival forest was explained with Shapley Additive Explanations (SHAP), and its feature attributions were translated into recommendations for a Cox model: feature exclusions, nonlinear terms, and interactions. Data were split 80/20 into training and test sets. We compared the recommendations-augmented Cox model with the baseline Cox model, a Cox model with restricted cubic splines, and a penalized Cox model with all pairwise interactions. The metrics were Harrell concordance index (C-index), Uno C, calibration slope, and integrated Brier score. We repeated the analysis with all 409 available features and in 3 public datasets (German Breast Cancer Study Group 2 [GBSG2], AIDS Clinical Trial [ACT], and PeakVO2).

Results

Of 220,341 patients (mean age 64, SD 12 years), 8489 (3.85%) had an FRI within 1 year. From 59 candidate features, the recommender excluded 15 (8 medication classes and 7 comorbidities). The final model retained 44 predictors (demographic and lifestyle factors, frailty, 21 comorbidities, and 15 medication classes), with nonlinear terms for age, frailty, BMI, and area deprivation, and interactions of age with stroke, osteoporosis, loop diuretics, and frailty. The augmented model had a C-index of 0.719 (95% CI 0.708-0.731) compared with 0.715 (95% CI 0.702-0.727) for the baseline model (difference 0.004, 95% CI 0.001-0.007), with 70 vs 73 parameters, and similar calibration (slope 0.962 vs 0.961). With all 409 features, the recommender recommended 304 exclusions, 4 nonlinear terms, and 2 interactions. The augmented model had a C-index of 0.722 compared with 0.713, with 129 vs 423 parameters, and better calibration (slope 0.952 vs 0.846). The method also proved effective on 3 additional public datasets, demonstrating wider applicability.

Conclusions

The proposed exploratory AI recommender demonstrates the potential of explainable AI to guide the design of transparent predictive models.

Subjects

MeSH terms (MEDLINE)

Accidental FallsArtificial IntelligenceAlgorithmsWounds and InjuriesHumansPrediction AlgorithmsMiddle AgedFemaleAgedMaleProportional Hazards ModelsPredictive Learning Models

Author keywords

Cox proportional hazardsShapley valuesexplainabilityexplainable artificial intelligencefeature attributionfeature interactionsfeature selectionhigh-dimensional datamachine learningprediction modelsrisk predictionstudy design

Authors and institutions

  • School of Engineering, University of Edinburgh, Edinburgh, United Kingdom.
  • Advanced Care Research Centre, Usher School of Population Health Sciences, University of Edinburgh, Edinburgh, United Kingdom.
  • Centre for Medical Informatics, Usher School of Population Health Sciences, University of Edinburgh, Edinburgh, United Kingdom.

ORCID profiles: Junyu Yan, Damian Machlanski, Kurt Butler, Panagiotis Dimitrakopoulos, Ewen M Harrison, Bruce Guthrie, Sotirios A Tsaftaris

Cite

APA

Yan, J., Machlanski, D., Butler, K., Dimitrakopoulos, P., Harrison, E. M., Guthrie, B., & Tsaftaris, S. A. (2026). Explainable AI to Guide the Design of Interpretable Prediction Models From High-Dimensional Data, Applied to Falls and Related Injuries in Adults Aged 50 Years or Older: Algorithm Development and Validation. Journal of Medical Internet Research, 28, e105364. https://doi.org/10.2196/105364

Vancouver

Yan J, Machlanski D, Butler K, Dimitrakopoulos P, Harrison EM, Guthrie B, et al. Explainable AI to Guide the Design of Interpretable Prediction Models From High-Dimensional Data, Applied to Falls and Related Injuries in Adults Aged 50 Years or Older: Algorithm Development and Validation. Journal of Medical Internet Research. 2026;28:e105364. doi:10.2196/105364

BibTeX
@article{Yan2026,
  author = {Yan, Junyu and Machlanski, Damian and Butler, Kurt and Dimitrakopoulos, Panagiotis and Harrison, Ewen M and Guthrie, Bruce and Tsaftaris, Sotirios A},
  title = {Explainable AI to Guide the Design of Interpretable Prediction Models From High-Dimensional Data, Applied to Falls and Related Injuries in Adults Aged 50 Years or Older: Algorithm Development and Validation},
  journal = {Journal of Medical Internet Research},
  year = {2026},
  volume = {28},
  pages = {e105364},
  doi = {10.2196/105364}
}
Start for free →