EUVIMEDAtrasti EUVIMED
Open access · CC BY 4.0

A Personalized Predictive Model That Jointly Optimizes Discrimination and Calibration

Tatiana Krikella iD, Joel A Dubin

Statistics in Medicine · 2025-05 · 44 (10-12)

DOI: 10.1002/sim.70077 ↗

At a glance

Journal article in Statistics in Medicine, published online on 2025-05-16. Written by Tatiana Krikella and 1 co-authors. Institution: Department of Statistics and Actuarial Science, University of Waterloo, Waterloo, Ontario, Canada. Major MEDLINE topics: Precision Medicine, Algorithms, Models, Statistical. Cited 3 times so far (Crossref); 30 references. The full text is freely available (open access, CC BY 4.0). Funded by: Natural Sciences and Engineering Research Council of Canada (NSERC).

SantraukaSubjectsBibliografiniai duomenysCiteRelated studies

Abstract (original language)

Precision medicine is accelerating rapidly in the field of health research. This includes fitting predictive models for individual patients based on patient similarity in an attempt to improve model performance. We propose an algorithm which fits a personalized predictive model (PPM) using an optimal size of a similar subpopulation that jointly optimizes model discrimination and calibration, as it is criticized that calibration is not assessed nearly as often as discrimination despite poorly calibrated models being potentially misleading. We define a mixture loss function that considers model discrimination and calibration, and allows for flexibility in emphasizing one performance measure over another. We empirically show that the relationship between the size of subpopulation and calibration is quadratic, which motivates the development of our jointly optimized model. We also investigate the effect of within‐population patient weighting on performance and conclude that the size of subpopulation has a larger effect on the predictive performance of the PPM compared to the choice of weight function.

Subjects

MeSH terms (MEDLINE)

Precision MedicineAlgorithmsModels, StatisticalHumansCalibrationComputer Simulation

Author keywords

Brier Scorecosine similaritymixture loss functionprecision medicineprediction modelsubpopulation

Authors and institutions

  • Department of Statistics and Actuarial Science, University of Waterloo, Waterloo, Ontario, Canada.

ORCID profiles: Tatiana Krikella

Funding

  • Natural Sciences and Engineering Research Council of Canada (NSERC)

Cite

APA

Krikella, T., & Dubin, J. A. (2025). A Personalized Predictive Model That Jointly Optimizes Discrimination and Calibration. Statistics in Medicine, 44(10-12). https://doi.org/10.1002/sim.70077

Vancouver

Krikella T, Dubin JA. A Personalized Predictive Model That Jointly Optimizes Discrimination and Calibration. Statistics in Medicine. 2025;44(10-12). doi:10.1002/sim.70077

BibTeX
@article{Krikella2025,
  author = {Krikella, Tatiana and Dubin, Joel A},
  title = {A Personalized Predictive Model That Jointly Optimizes Discrimination and Calibration},
  journal = {Statistics in Medicine},
  year = {2025},
  volume = {44},
  number = {10-12},
  doi = {10.1002/sim.70077}
}
Start for free →