EUVIMEDEUVIMED entdecken
Crossref

Evaluating risk prediction models: the predictiveness curve and its geometric summaries

Wei-Min Chiu, Dai-Rong Tsai iD, Chun-Ju Chiang iD, Wen-Chung Lee iD

American Journal of Epidemiology · 04.11.2026 · 195 (11) · 2899-2906

DOI: 10.1093/aje/kwag123 ↗

Kurz gefasst

Fachartikel in der Zeitschrift American Journal of Epidemiology, online erschienen am 03.06.2026. Verfasst von Wei-Min Chiu und 3 weiteren Autor:innen. Beteiligte Einrichtung: Institute of Epidemiology and Preventive Medicine, College of Public Health, National Taiwan University, Taipei 100, Taiwan. Hauptthemen laut MEDLINE: Models, Statistical. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 19 Literaturangaben. Gefördert von: the National Science and Technology Council in Taiwan, the Health Promotion Administration, the Ministry of Health and Welfare in Taiwan.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Risk prediction models are widely used, but their evaluation is dominated by discrimination-based measures, such as the receiver operating characteristic (ROC) curve and AUC, which provide limited insight into how risks are distributed across a population. We extend the predictiveness curve to risk prediction models, characterize its geometric properties, and derive 3 geometric summaries—the Pietra index, Gini index, and scaled Brier score. To obtain well-calibrated and stable risk estimates, we implement a 3-step prediction post-processing procedure, consisting of cross-validation, isotonic recalibration, and bootstrap averaging. We illustrate the approach using toy examples and a nationwide Taiwan cohort of 23 839 lung cancer patients, in whom a Cox model was used to estimate the 5-year risk of fatality. The 3 indices captured complementary aspects of performance and supported model choice for specific purposes: the Pietra index for minimizing the gray zone, the Gini index for maximizing separation, and the scaled Brier score for increasing certainty. In the lung cancer application, the final post-processed model showed clear risk stratification: 25.2% of patients were very low risk (<10% fatality), and 50.2% were high risk (>75% fatality), whereas only 4.8% were near the average risk. Overall, the predictiveness curve and its geometric summaries provide a population-oriented framework for transparent, purpose-specific evaluation of risk prediction models.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Models, StatisticalHumansLung NeoplasmsRisk AssessmentTaiwanPrediction AlgorithmsProportional Hazards ModelsROC CurveFemale

Stichwörter der Autor:innen

Gini indexPietra indexpredictiveness curverisk prediction modelrisk stratificationscaled Brier score

Autor:innen und Institutionen

  • Institute of Epidemiology and Preventive Medicine, College of Public Health, National Taiwan University, Taipei 100, Taiwan.
  • Taiwan Cancer Registry, Taipei 100, Taiwan.
  • Institute of Health Data Analytics and Statistics, College of Public Health, National Taiwan University, Taipei 100, Taiwan.

ORCID-Profile: Dai-Rong Tsai, Chun-Ju Chiang, Wen-Chung Lee

Förderung

  • the National Science and Technology Council in Taiwan
  • the Health Promotion Administration, the Ministry of Health and Welfare in Taiwan

Zitieren

APA

Chiu, W. M., Tsai, D. R., Chiang, C. J., & Lee, W. C. (2026). Evaluating risk prediction models: the predictiveness curve and its geometric summaries. American Journal of Epidemiology, 195(11), 2899-2906. https://doi.org/10.1093/aje/kwag123

Vancouver

Chiu WM, Tsai DR, Chiang CJ, Lee WC. Evaluating risk prediction models: the predictiveness curve and its geometric summaries. American Journal of Epidemiology. 2026;195(11):2899-2906. doi:10.1093/aje/kwag123

BibTeX
@article{Chiu2026,
  author = {Chiu, Wei-Min and Tsai, Dai-Rong and Chiang, Chun-Ju and Lee, Wen-Chung},
  title = {Evaluating risk prediction models: the predictiveness curve and its geometric summaries},
  journal = {American Journal of Epidemiology},
  year = {2026},
  volume = {195},
  number = {11},
  pages = {2899--2906},
  doi = {10.1093/aje/kwag123}
}
Kostenlos starten →