CheXclusion: Fairness gaps in deep chest X-ray classifiers
Open accessThe DOI matches its Crossref record; 55 of the 55 references it cites were found in a registry or an index.
Abstract
Machine learning systems have received much attention recently for their ability to achieve expert-level performance on clinical tasks, particularly in medical imaging. Here, we examine the extent to which state-of-the-art deep learning classifiers trained to yield diagnostic labels from X-ray images are biased with respect to protected attributes. We train convolution neural networks to predict 14 diagnostic labels in 3 prominent public chest X-ray datasets: MIMIC-CXR, Chest-Xray8, CheXpert, as well as a multi-site aggregation of all those datasets. We evaluate the TPR disparity - the difference in true positive rates (TPR) - among different protected attributes such as patient sex, age, race, and insurance type as a proxy for socioeconomic status. We demonstrate that TPR disparities exist in the state-of-the-art classifiers in all datasets, for all clinical tasks, and all subgroups. A multi-source dataset corresponds to the smallest disparities, suggesting one way to reduce bias. We find that TPR disparities are not significantly correlated with a subgroup's proportional disease burden. As clinical models move from papers to products, we encourage clinical decision makers to carefully audit for algorithmic disparities prior to deployment. Our supplementary materials can be found at, http://www.marzyehghassemi.com/chexclusion-supp-3/.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“For example, chest radiograph classifiers trained on publicly available datasets may have multiple racial, sex, and socioeconomic biases that reflect the patient distribution in their training data (16), resulting in poor generalizability.”
From Mitigating Bias in Radiology Machine Learning: 1. Data Handling (2022) · cites this paper as [16]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Seyyed-Kalantari et al [87] noted that convolutional neural networks will frequently underdiagnose Hispanic patients at a disproportionate rate due to the potential lack of access to health care and insurance type.”
From Barriers to and facilitators of artificial intelligence adoption in health care: scoping review (2024) · cites this paper as [87]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“The classifier may lead to disparity in the results by classifying the same information into different subgroups [125].”
From Artificial intelligence bias in medical system designs: a systematic review (2023) · cites this paper as [125]
In the citing paper
“Similarly, breast cancer detection systems may perform differently based on breast density variations across ethnic groups [117,118].”
From Artificial Intelligence in Clinical Oncology: From Productivity Enhancement to Creative Discovery (2025) · cites this paper as [118]
In the citing paper
“Then as well Chakravarty et al. [23] has an AUC score of 80% with FL environment, but with same dataset and ML algorithm article [67] and [68] have 86% and 87% AUC respectively.”
From A systematic review on federated learning in medical image analysis (2023) · cites this paper as [67]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Adam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946
ImageNet: A large-scale hierarchical image database
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., & Fei-Fei, L. (2009) · 2009 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2009.5206848 · Cited by 65,333 · cited 2×
“Data Processing: We resize all images to 224 x 224 and normalize via the mean and standard deviation of the ImageNet dataset [12].”
MethodDensely Connected Convolutional Networks
Huang, G., Liu, Z., Van Der Maaten, L., & Weinberger, K. Q. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.243 · Cited by 48,566 · cited 1×
“Model Architecture: We initialize a 121-layer DenseNet [19] with the ImageNet [12] pre-trained weights and train models with a binary cross entropy loss function.”
MethodDermatologist-level classification of skin cancer with deep neural networks
Esteva, A., Kuprel, B., Novoa, R. A., Ko, J., Swetter, S. M., Blau, H. M., & Thrun, S. (2017) · Nature · doi:10.1038/nature21056 · Cited by 14,362
High-performance medicine: the convergence of human and artificial intelligence
Topol, E. J. (2019) · Nature Medicine · doi:10.1038/s41591-018-0300-7 · Cited by 10,450 · cited 1×
“While this may seem to make a clear case for implementing AI-enabled diagnostic tools [20], moving such methods from paper to practice require careful thought [14, 46].”
BackgroundSimultaneous Statistical Inference
Miller, R. G. (1981) · Springer Series in Statistics · doi:10.1007/978-1-4613-8122-8 · Cited by 5,604
Wang, X., Peng, Y., Lu, L., Lu, Z., Bagheri, M., & Summers, R. M. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.369 · Cited by 4,616
(2018) · Cited by 3,510
Equality of Opportunity in Supervised Learning
Hardt, M., Price, E., & Srebro, N. (2016) · arXiv · doi:10.48550/arxiv.1610.02413 · Cited by 2,536 · cited 2×
“While there are multiple technical fairness definitions [8, 10, 17, 50], and different fairness notions are not always simultaneously achievable [8, 23], we target the equality of opportunity notion across the protected attributes of sex, age, race and insurance type, and focus on two aims: quantifying TPR disparity a…”
BackgroundFair prediction with disparate impact: a study of bias in recidivism prediction instruments
Chouldechova, A. (2017) · Big Data · doi:10.1089/big.2016.0047 · Cited by 2,159 · cited 3×
“While there are multiple technical fairness definitions [8, 10, 17, 50], and different fairness notions are not always simultaneously achievable [8, 23], we target the equality of opportunity notion across the protected attributes of sex, age, race and insurance type, and focus on two aims:…”
Background
Research works printed but not linked: 1
- [22]22. D. S. Char, N. H. Shah and D. Magnus, Implementing machine learning in health care addressing ethical challenges, New England Journal of Medicine 378 , 981 (2018).
Cited by
COVID-19 Image Data Collection: Prospective Predictions are the Future
Cohen, J. P., Morrison, P., Dao, L., Roth, K., Duong, T., & Ghassem, M. (2020) · Machine Learning for Biomedical Imaging · doi:10.59275/j.melba.2020-48g7 · Cited by 927
“This can lead to many negative outcomes if treated as a typical benchmark dataset [Maguolo and Nanni, 2020; Cohen et al., 2020b; Seyyed-Kalantari et al., 2020; Kelly et al., 2019].”
BackgroundSelf-citationSeyyed-Kalantari, L., Zhang, H., McDermott, M. B. A., Chen, I. Y., & Ghassemi, M. (2021) · Nature Medicine · doi:10.1038/s41591-021-01595-0 · Cited by 928
“However, they are evaluated in each of five models (the same model trained five times with five different random seeds 7 ), with the train– validation–test split kept fixed in the training of the five models.”
MethodSelf-citationAlgorithmic fairness in artificial intelligence for medicine and healthcare
Chen, R. J., Wang, J. J., Williamson, D. F. K., Chen, T. Y., Lipkova, J., Lu, M. Y., Sahai, S., & Mahmood, F. (2023) · Nature Biomedical Engineering · doi:10.1038/s41551-023-01056-8 · Cited by 750
AI recognition of patient race in medical imaging: a modelling study
Gichoya, J. W., Banerjee, I., Bhimireddy, A. R., Burns, J. L., Celi, L. A., Chen, L.-C., Correa, R., Dullerud, N., Ghassemi, M., Huang, S.-C., et al. (2022) · The Lancet Digital Health · doi:10.1016/s2589-7500(22)00063-2 · Cited by 611
“10 In related work, several groups reported that AI algorithms can identify various demographic patient factors.”
Self-citationEthical Machine Learning in Healthcare
Chen, I. Y., Pierson, E., Rose, S., Joshi, S., Ferryman, K., & Ghassemi, M. (2021) · Annual Review of Biomedical Data Science · doi:10.1146/annurev-biodatasci-092820-114757 · Cited by 645
“Choosing these patients can rely on parts of the EHR that may be skewed due to lack of access to care, or abnormalities in clinical care: e.g., economic incentives may alter diagnosis code logging (81), clinical protocol affects the frequency and observation of abnormal tests (62), historical racial mistrust may delay…”
BackgroundSelf-citation
Metadata
- Authors
- Seyyed-Kalantari, L., Liu, G., McDermott, M., Chen, I. Y., & Ghassemi, M.
- Affiliations
- University of Toronto (CA)
- Massachusetts Institute of Technology (US)
- Venue
- Biocomputing 2021
- Publisher
- WORLD SCIENTIFIC
- Published
- 2020-11
- Volume
- 26
- Pages
- 232-243
- OpenAlex
- W3007679811
- Cited by
- 275
- References
- 47
- Citations by year
- 2026: 21 · 2025: 58 · 2024: 45 · 2023: 66 · 2022: 45 · 2021: 29 · 2020: 11
- FWCI
- 19.39
- Type
- conference paper
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by-nc
- Locations
- Pacific Symposium on Biocomputing (PSB) (publishedVersion, cc-by-nc)
- PubMed (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Topics
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
- Physical Sciences › Computer Science › Artificial Intelligence › Explainable Artificial Intelligence (XAI)
- Keywords
- algorithmic fairness, algorithmic bias, chest X-ray classification, deep learning classifier, protected characteristics, socioeconomic status, insurance type, race, sex, algorithmic auditing, CheXpert, MIMIC-CXR, ChestX-ray14
- MeSH terms
- Machine Learning, Humans, X-Rays, Neural Networks, Computer, Computational Biology
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: CheXclusion: Fairness gaps in deep chest X-ray classifiers. https://app.cochat.ai/papers/10.1142/9789811232701_0022
- The paper
- Seyyed-Kalantari, L., Liu, G., McDermott, M., Chen, I. Y., & Ghassemi, M. (2020). CheXclusion: Fairness gaps in deep chest X-ray classifiers. Biocomputing 2021. https://doi.org/10.1142/9789811232701_0022
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.