An Algorithmic Bias Playbook — designed to help health care organizations find and fix biased algorithms that can lead to faulty decision-making — has recently been released by Berkeley Public Health Professor, Ziad Obermeyer, and a team of researchers including BIDS Alum Stephanie Eaneff and colleagues at the University of Chicago Booth School of Business.
The playbook is based on insights gleaned from the authors’ experiences over years of applied work, and describes four steps organizations can take to identify and mitigate biased algorithms. “Algorithmic bias is everywhere,” the authors posit. “Our work with dozens of organizations—health care providers, insurers, technology companies, and regulators—has taught us that biased algorithms are deployed throughout the health care system, influencing clinical care, operational workflows, and policy. Beyond healthcare, we've seen algorithmic bias influence decisions in criminal justice, finance, and other fields. But our work has also shown us that there are solutions, and we are sharing resources to help leaders, practitioners, and policymakers address the problem and mitigate algorithmic bias wherever they find it.”
The book presents concrete examples — “cautionary tales” — that enable users to define, identify, and measure racial bias in live algorithms, as well as “success stories” that demonstrate how bias can be mitigated, and how flawed algorithms can be “transformed into tools to fight injustice.” By improving strategies that enable better technical choices, the playbook’s practical approach to bias can be used by a variety of users to implement oversight structures that prevent bias, craft prospective guidance for industry, or even guide civil investigations.
Ziad Obermeyer and colleagues at the Booth School of Business release health care Algorithmic Bias Playbook
June 23, 2021 | Berkeley Public Health News
