文献速递医学人工智能与信息学流行病学与生物统计

临床机器学习模型中的社会人口学偏倚:算法偏倚实例与机制的范围综述

Sociodemographic bias in clinical machine learning models: a scoping review of algorithmic bias instances and mechanisms.
PubMed 文献 · J Clin Epidemiol · 2024-11-10

本范围综述检索MEDLINE、EMBASE、PsycINFO和Web of Science,纳入91项对临床机器学习模型进行社会人口学偏倚评估的研究。多数研究仅评估单一因素,最常见为种族、性别/社会性别和年龄;74.7%报告存在偏倚,其中87%的偏倚针对社会经济弱势群体。结论认为应对所有临床机器学习模型完成偏倚评估。

要点

  • 91项研究中74.7%报告临床机器学习模型存在社会人口学偏倚。
  • 偏倚评估最多集中于种族(64.8%)、性别/社会性别(45.1%)和年龄(26.4%)。
  • 仅1.1%的研究评估了交叉性因素,其他社会人口学因素评估较少。

核心结论

  • 多数接受偏倚评估的机器学习算法在社会人口学因素上表现出偏倚。
  • 偏倚评估主要集中于种族、性别/社会性别和年龄,其他社会人口学因素及其交叉评估较少。
  • 鉴于潜在的健康公平影响,应对所有临床机器学习模型完成偏倚评估。

研究设计与样本

  • 采用范围综述设计,依据Joanna Briggs Institute指南实施,并按PRISMA范围综述扩展版报告。
  • 检索MEDLINE、EMBASE、PsycINFO和Web of Science,纳入评估临床机器学习算法社会人口学偏倚的研究。
  • 共识别6448篇文章,其中760篇报告临床机器学习模型,91篇(12.0%)完成偏倚评估并符合全部纳入标准。

主要发现

  • 多数研究评估单一社会人口学因素(n=56,61.5%)。
  • 最常评估的因素为种族(n=59,64.8%),其次为性别/社会性别(n=41,45.1%)和年龄(n=24,26.4%),仅1项研究(1.1%)评估交叉性因素。
  • 74.7%(n=68)的研究报告存在偏倚,18.7%(n=17)报告不存在偏倚,6.6%(n=6)未说明是否存在偏倚。
  • 在报告存在偏倚的研究中,87%报告偏倚针对社会经济弱势群体。

临床怎么用

  • 作者认为,鉴于潜在的健康公平影响,应对所有临床机器学习模型完成偏倚评估。

局限性

  • 原文未报告局限性。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

BACKGROUND AND OBJECTIVES: Clinical machine learning (ML) technologies can sometimes be biased and their use could exacerbate health disparities. The extent to which bias is present, the groups who most frequently experience bias, and the mechanism through which bias is introduced in clinical ML applications is not well described. The objective of this study was to examine instances of bias in clinical ML models. We identified the sociodemographic subgroups PROGRESS that experienced bias and the reported mechanisms of bias introduction. METHODS: We searched MEDLINE, EMBASE, PsycINFO, and Web of Science for all studies that evaluated bias on sociodemographic factors within ML algorithms created for the purpose of facilitating clinical care. The scoping review was conducted according to the Joanna Briggs Institute guide and reported using the PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) extension for scoping reviews. RESULTS: We identified 6448 articles, of which 760 reported on a clinical ML model and 91 (12.0%) completed a bias evaluation and met all inclusion criteria. Most studies evaluated a single sociodemographic factor (n = 56, 61.5%). The most frequently evaluated sociodemographic factor was race (n = 59, 64.8%), followed by sex/gender (n = 41, 45.1%), and age (n = 24, 26.4%), with one study (1.1%) evaluating intersectional factors. Of all studies, 74.7% (n = 68) reported that bias was present, 18.7% (n = 17) reported bias was not present, and 6.6% (n = 6) did not state whether bias was present. When present, 87% of studies reported bias against groups with socioeconomic disadvantage. CONCLUSION: Most ML algorithms that were evaluated for bias demonstrated bias on sociodemographic factors. Furthermore, most bias evaluations concentrated on race, sex/gender, and age, while other sociodemographic factors and their intersection were infrequently assessed. Given potential health equity implications, bias assessments should be completed for all clinical ML models.

来源:J Clin Epidemiol,查看原文。版权归原作者所有。

引用

  1. Colacci M, Huang YQ, Postill G, et al. Sociodemographic bias in clinical machine learning models: a scoping review of algorithmic bias instances and mechanisms. J Clin Epidemiol. 2024. PMID: 39532254. doi:10.1016/j.jclinepi.2024.111606

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。