Here we present a selection of relevant news articles and research regarding fair machine learning for health. The field is changing and growing, so if you have suggestions to include, please email email@example.com, and we will update the page.
Recent News, Op-Eds, and Whitepapers
"Fairness in Machine Learning in Health Workshop @ Data and Society, Meeting Summary", compiled by Kadija Ferryman and Marzyeh Ghassemi, Dec 2019.
"Bias In Medicine: Last Week Tonight with John Oliver" (US link), Aug 18, 2019.
"Fitbits and other wearables may not accurately track heart rates in people of color", Ruth Hailu. Stat News, July 24, 2019.
"We should treat algorithms like prescription drugs", Andy Coravos, Irene Chen, Ankit Gordhandas, and Ariel Dora Stern. Quartz, Feb 14, 2019.
"AI could worsen health disparities", Dhruv Khullar. New York Times, Jan 31, 2019.
Fair Regression for Health Care Spending. Zink and Rose.
Dissecting racial bias in an algorithm used to manage the health of populations. Obermeyer, Powers, Vogeli, Mullainathan. Science 2019.
Counterfactual Reasoning for Fair Clinical Risk Prediction. Pfohl, Duan, Ding, Shah. MLHC 2019.
Fairness without Harm: Decoupled Classifiers with Preference Guarantees. Ustun, Liu, and Parkes. ICML 2019.
Creating Fair Models of Atherosclerotic Cardiovascular Disease. Pfohl et al. AIES 2019.
The Disparate Impacts of Medical and Mental Health with AI. Chen, Szolovits, and Ghassemi, AMA Journal of Ethics, 2019.
Ensuring Fairness in Machine Learning to Advance Health Equity. Rajkomar et al, Annals of Internal Medicine 2018.
Racial Disparities and Mistrust in End-of-Life Care. Boag et al, MLHC 2018.
Machine Learning and Health Care Disparities in Dermatology. Adamson and Smith, JAMA Dermatology 2018.
Potential Biases in Machine Learning Algorithms Using Electronic Health Record Data. Gianfrancesco et al, JAMA Internal Medicine 2018.
Biases introduced by filtering electronic health records for patients with "complete data". Weber et al, J Am Med Inform Assoc 2017.
MIMIC-III: Critical care dataset of over 23,000 adult ICU patient stays and 20,000 pediatric ICU stays with treatments, labs, and clinical notes. Features include patient sex, race, and insurance type.