EUVIMEDEUVIMED entdecken
Crossref

De-biasing the bias: methods for improving disparity assessments with noisy group measurements

Solvejg Wastvedt iD, Joshua Snoke iD, Denis Agniel iD, Julie Lai, Marc N Elliott, Steven C Martino

Biometrics · 03.10.2024 · 80 (4)

DOI: 10.1093/biomtc/ujae155 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Biometrics, online erschienen am 28.12.2024. Verfasst von Solvejg Wastvedt und 5 weiteren Autor:innen. Beteiligte Einrichtung: Department of Statistics & Data Science, NORC at the University of Chicago, Chicago, IL 60603, United States. Hauptthemen laut MEDLINE: Algorithms, Bias, Bayes Theorem, Healthcare Disparities. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 22 Literaturangaben. Gefördert von: Robert Wood Johnson Foundation.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Health care decisions are increasingly informed by clinical decision support algorithms, but these algorithms may perpetuate or increase racial and ethnic disparities in access to and quality of health care. Further complicating the problem, clinical data often have missing or poor quality racial and ethnic information, which can lead to misleading assessments of algorithmic bias. We present novel statistical methods that allow for the use of probabilities of racial/ethnic group membership in assessments of algorithm performance and quantify the statistical bias that results from error in these imputed group probabilities. We propose a sensitivity analysis approach to estimating the statistical bias that allows practitioners to assess disparities in algorithm performance under a range of assumed levels of group probability error. We also prove theoretical bounds on the statistical bias for a set of commonly used fairness metrics and describe real-world scenarios where our theoretical results are likely to apply. We present a case study using imputed race and ethnicity from the modified Bayesian Improved First and Surname Geocoding algorithm for estimation of disparities in a clinical decision support algorithm used to inform osteoporosis treatment. Our novel methods allow policymakers to understand the range of potential disparities under a given algorithm even when race and ethnicity information is missing and to make informed decisions regarding the implementation of machine learning for clinical decision support.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

AlgorithmsBiasBayes TheoremHealthcare DisparitiesHumansEthnicityOsteoporosisRacial GroupsDecision Support Systems, ClinicalBiometryModels, Statistical

Stichwörter der Autor:innen

Bayesian improved surname geocodingalgorithmic fairnessrace imputationsensitivity analysis

Autor:innen und Institutionen

  • Department of Statistics & Data Science, NORC at the University of Chicago, Chicago, IL 60603, United States.
  • RAND Corporation, Pittsburgh, PA 15213, United States.
  • RAND Corporation, Santa Monica, CA 90401-3208, United States.

ORCID-Profile: Solvejg Wastvedt, Joshua Snoke, Denis Agniel

Förderung

  • Robert Wood Johnson Foundation

Zitieren

APA

Wastvedt, S., Snoke, J., Agniel, D., Lai, J., Elliott, M. N., & Martino, S. C. (2024). De-biasing the bias: methods for improving disparity assessments with noisy group measurements. Biometrics, 80(4). https://doi.org/10.1093/biomtc/ujae155

Vancouver

Wastvedt S, Snoke J, Agniel D, Lai J, Elliott MN, Martino SC. De-biasing the bias: methods for improving disparity assessments with noisy group measurements. Biometrics. 2024;80(4). doi:10.1093/biomtc/ujae155

BibTeX
@article{Wastvedt2024,
  author = {Wastvedt, Solvejg and Snoke, Joshua and Agniel, Denis and Lai, Julie and Elliott, Marc N and Martino, Steven C},
  title = {De-biasing the bias: methods for improving disparity assessments with noisy group measurements},
  journal = {Biometrics},
  year = {2024},
  volume = {80},
  number = {4},
  doi = {10.1093/biomtc/ujae155}
}
Kostenlos starten →