Reducing Racial and Ethnic Algorithmic Bias and Ensuring that VA Algorithms are Fair
Lauren Korshak, DHealth(c), MS, RCEP
Amol Navathe, MD, PhD
Ravi Parikh, MD, MPP
Suzanne Tamang, PhD, MS
Seminar date: 3/10/2021
Description: Predictive algorithms both inside and outside of VA do not always accurately represent risks of poor health outcomes in minority populations. This Cyberseminar will use existing two VA algorithms, the Office of Mental Health and Suicide Prevention’s Stratification Tool for Opioid Risk Mitigation and CAN Scores, to describe racial and ethnic algorithmic bias assessment methods and highlight strategies that can be used in development and evaluation of population-based risk stratification tools to help ameliorate unfairness.
The intended audience:
Researchers, clinicians, care coordinators, policy makers, educators, data analysts, Veteran stakeholders and representatives of vulnerable Veteran groups who are interested in understanding and/or promoting equitable health among all Veterans.