Ensuring Fairness in Machine Learning to Advance Health Equity
2018/12/03 by Alvin Rajkomar, Michaela Hardt, Michael Howell +4 · 1,227 citations
Health Professions · Medicine · Nursing · Psychology · #Actuarial science #Artificial Intelligence in Healthcare and Education #Artificial intelligence #Computer science #Distributive justice #Economic Justice #Economics #Equity (law) #Ethics in Clinical Research #Harm #Health care #Health equity #Healthcare cost, quality, practices #Implementation #Knowledge management #Machine learning #Management science #Medicine #Microeconomics #Nursing #Political science #Psychology #Public health #Risk analysis (engineering) #Social psychology #Software deployment #Software engineering
paper · open access · doi:10.7326/m18-1990
published in Annals of Internal Medicine 169(12), 866-872 (American College of Physicians)
openalex publication_date 2018/12/03 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/05
Abstract
Machine learning is used increasingly in clinical care to improve diagnosis, treatment selection, and health system efficiency. Because machine-learning models learn from historically collected data, populations that have experienced human and structural biases in the past-called protected groups-are vulnerable to harm by incorrect predictions or withholding of resources. This article describes how model design, biases in data, and the interactions of model predictions with clinicians and patients may exacerbate health care disparities. Rather than simply guarding against these harms passively, machine-learning systems should be used proactively to advance health equity. For that goal to be achieved, principles of distributive justice must be incorporated into model design, deployment, and evaluation. The article describes several technical implementations of distributive justice-specifically those that ensure equality in patient outcomes, performance, and resource allocation-and guides clinicians as to when they should prioritize each principle. Machine learning is providing increasingly sophisticated decision support and population-level monitoring, and it should encode principles of justice to ensure that models benefit all patients.
Cited by
- Advancing healthcare equity through dissemination and implementation science
- Cardiovascular informatics: building a bridge to data harmony
- Evaluating prediction model performance
- Safety boundary maintenance in consumer AI systems responding to pediatric health queries: a cross-platform benchmark evaluation under naturalistic and adversarially pressured conditions
- Utilizing a domain-specific large language model for LI-RADS v2018 categorization of free-text MRI reports: a feasibility study
- Interpretable Machine Learning for Air Pollution and Respiratory Health Prediction: A Socioeconomic Subgroup Analysis
- Fairness in Missing Data Imputation
- Target specification bias, counterfactual prediction, and algorithmic fairness in healthcare
- A Unifying Human-Centered AI Fairness Framework
- Subgroup Validity in Machine Learning for Echocardiogram Data
- Towards Robust and Fair Next Visit Diagnosis Prediction under Noisy Clinical Notes with Large Language Models
- Social Epidemiology: Past, Present, and Future
- Machine Learning Models for Predicting Smoking-Related Health Decline and Disease Risk
- SurvBench: A Standardised Preprocessing Pipeline for Multi-Modal Electronic Health Record Survival Analysis
- Assessing Social Determinants-Related Performance Bias of Machine Learning Models: A case of Hyperchloremia Prediction in ICU Population
- The BigScience ROOTS Corpus: A 1.6TB Composite Multilingual Dataset
- Before the Clinic: Transparent and Operable Design Principles for Healthcare AI
- Causal thinking for decision making on Electronic Health Records: why and how
- Opportunities for incorporating intersectionality into biomedical informatics
- Vision Transformer–based Decision Support for Neurosurgical Intervention in Acute Traumatic Brain Injury: Automated Surgical Intervention Support Tool
- Equality of opportunity in travel behavior prediction with deep neural networks and discrete choice models
- Building AI Companions that Prioritise Learning over Performance
- Machine Learning and Public Health: Identifying and Mitigating Algorithmic Bias through a Systematic Review
- Robust or Suggestible? Exploring Non-Clinical Induction in LLM Drug-Safety Decisions
- The Fairness of Risk Scores Beyond Classification: Bipartite Ranking and the xAUC Metric
- A Sociotechnical View of Algorithmic Fairness
- The Average Patient Fallacy
- Algorithmic individual fairness and healthcare: a scoping review
- Revolutionizing emergency care: an overview of the transformative role of artificial intelligence in diagnosis, triage, and patient management
- Health GeoAI beyond algorithms: Embedding equity, accountability, and environmental responsibility
- Fertility preservation in endometriosis: evaluating surgical risks and emerging preservation approaches
- Chasing Your Long Tails: Differentially Private Prediction in Health Care Settings
- Hybrid Adaptive Conformal Offline Reinforcement Learning for Fair Population Health Management
- Counterfactual Reasoning for Fair Clinical Risk Prediction
- Leveraging Imperfection with MEDLEY A Multi-Model Approach Harnessing Bias in Medical AI
- Cross-Language Aphasia Detection using Optimal Transport Domain Adaptation
- FairTabGen: High-Fidelity and Fair Synthetic Health Data Generation from Limited Samples
- Competing Risks: Impact on Risk Estimation and Algorithmic Fairness
- FairPOT: Balancing AUC Performance and Fairness with Proportional Optimal Transport
- The Emerging Hazard of AI‐Related Health Care Discrimination
- SynthTextEval: Synthetic Text Data Generation and Evaluation for High-Stakes Domains
- Path-specific effects for pulse-oximetry guided decisions in critical care
- A Narrative Review on Large AI Models in Lung Cancer Screening, Diagnosis, and Treatment Planning
- Multi-Task Learning for Post-transplant Cause of Death Analysis: A Case Study on Liver Transplant
- Equitable Discrimination in Survival Prediction: The Maximum Expected C-Index
- Fairness Perceptions in Regression-based Predictive Models
- Discretization-free Multicalibration through Loss Minimization over Tree Ensembles
- Inherent Trade-offs in the Fair Allocation of Treatments
- Reliable fairness auditing with semi-supervised inference
- Machine learning in the social and health sciences
- Hurtful Words: Quantifying Biases in Clinical Contextual Word Embeddings
- FORCE: Feature-Oriented Representation with Clustering and Explanation
- Toward clinical digital phenotyping: a timely opportunity to consider purpose, quality, and safety. [europepmc]
- Artificial intelligence and algorithmic bias: implications for health systems. [europepmc]
- Back to the Future: Achieving Health Equity Through Health Informatics and Digital Health. [europepmc]
- Development and Reporting of Prediction Models: Guidance for Authors From Editors of Respiratory, Sleep, and Critical Care Journals. [europepmc]
- A Review of Challenges and Opportunities in Machine Learning for Health. [europepmc]
- Sex and gender differences and biases in artificial intelligence for biomedicine and healthcare. [europepmc]
- Real-World Integration of a Sepsis Deep Learning Technology Into Routine Clinical Care: Implementation Study. [europepmc]
- Predictably unequal: understanding and addressing concerns that algorithmic clinical prediction may increase health disparities. [europepmc]
- Digital technology, tele-medicine and artificial intelligence in ophthalmology: A global perspective. [europepmc]
- Identifying Ethical Considerations for Machine Learning Healthcare Applications. [europepmc]
- Use of Latent Class Analysis and k-Means Clustering to Identify Complex Patient Profiles. [europepmc]
- Artificial Intelligence for Mental Health Care: Clinical Applications, Barriers, Facilitators, and Artificial Wisdom. [europepmc]
- Does "AI" stand for augmenting inequality in the era of covid-19 healthcare? [europepmc]
- Structural Racism: The Rules and Relations of Inequity. [europepmc]
- Ethical Machine Learning in Healthcare. [europepmc]
- Accessing Artificial Intelligence for Clinical Decision-Making. [europepmc]
- Artificial intelligence for diagnosis and Gleason grading of prostate cancer: the PANDA challenge. [europepmc]
- Cross-ethnicity/race generalization failure of behavioral prediction from resting-state functional connectivity. [europepmc]
- Sources of bias in artificial intelligence that perpetuate healthcare disparities-A global review. [europepmc]
- Fairness of artificial intelligence in healthcare: review and recommendations. [europepmc]
- Principles and methods for transferring polygenic risk scores across global populations. [europepmc]
- Considerations for addressing bias in artificial intelligence for health equity. [europepmc]
- Artificial intelligence in clinical medicine: catalyzing a sustainable global healthcare paradigm. [europepmc]
- The value of standards for health datasets in artificial intelligence-based applications. [europepmc]
- Guiding Principles to Address the Impact of Algorithm Bias on Racial and Ethnic Disparities in Health and Health Care. [europepmc]
Related