Scholarlink

<p>Algorithmic bias occurs when predictive model performance varies meaningfully across sociodemographic classes, exacerbating systemic healthcare disparities. NYC Health + Hospitals, an urban safety net system, assessed bias in two binary classification models in our electronic medical record: one predicting acute visits for asthma and one predicting unplanned readmissions. We evaluated differences in subgroup performance across race/ethnicity, sex, language, and insurance using equal opportunity difference (EOD), a metric comparing false negative rates. The most biased classes (race/ethnicity for asthma, insurance for readmission) were targeted for mitigation using threshold adjustment, which adjusts subgroup thresholds to minimize EOD, and reject option classification, which re-classifies scores near the threshold by subgroup. Successful mitigation was defined as 1) absolute subgroup EODs &lt;5 percentage points, 2) accuracy reduction &lt;10%, and 3) alert rate change &lt;20%. Threshold adjustment met these criteria; reject option classification did not. We introduce a <InternalRef RefID="MOESM1">Supplementary Playbook</InternalRef> outlining our approach for low-resource bias mitigation.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Identifying and mitigating algorithmic bias in the safety net

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年6月5日

查看全文
  • Shaina Mackin,
  • Vincent J. Major,
  • Rumi Chunara,
  • Remle Newton-Dame

摘要

Algorithmic bias occurs when predictive model performance varies meaningfully across sociodemographic classes, exacerbating systemic healthcare disparities. NYC Health + Hospitals, an urban safety net system, assessed bias in two binary classification models in our electronic medical record: one predicting acute visits for asthma and one predicting unplanned readmissions. We evaluated differences in subgroup performance across race/ethnicity, sex, language, and insurance using equal opportunity difference (EOD), a metric comparing false negative rates. The most biased classes (race/ethnicity for asthma, insurance for readmission) were targeted for mitigation using threshold adjustment, which adjusts subgroup thresholds to minimize EOD, and reject option classification, which re-classifies scores near the threshold by subgroup. Successful mitigation was defined as 1) absolute subgroup EODs <5 percentage points, 2) accuracy reduction <10%, and 3) alert rate change <20%. Threshold adjustment met these criteria; reject option classification did not. We introduce a Supplementary Playbook outlining our approach for low-resource bias mitigation.