Underdiagnosis bias of artificial intelligence algorithms applied to chest radiographs in under-served patient populations
By Laleh Seyyed-Kalantari, Haoran Zhang, Matthew B. A. McDermott, Irene Y. Chen, Marzyeh Ghassemi. Nature Medicine, 2021
830 citations recorded by the source corpus.
Artificial intelligence (AI) systems have increasingly achieved expert-level performance in medical imaging applications. However, there is growing concern that such AI systems may reflect and amplify human bias, and reduce the quality of their performance in historically under-served populations such as female patients, Black patients, or patients of low socioeconomic status. Such biases are especially troubling in the context of underdiagnosis, whereby the AI algorithm would inaccurately label an individual with a disease as healthy, potentially delaying access to care. Here, we examine algorithmic underdiagnosis in chest X-ray pathology classification across three large chest X-ray datasets, as well as one multi-source dataset. We find that classifiers produced using state-of-the-art computer vision techniques consistently and selectively underdiagnosed under-served patient populations and that the underdiagnosis rate was higher for intersectional under-served subpopulations, for example, Hispanic female patients. Deployment of AI systems using medical imaging for disease diagnosis with such biases risks exacerbation of existing care biases and can potentially lead to unequal access to medical treatment, thereby raising ethical concerns for the use of these models in the clinic. Artificial intelligence algorithms trained using chest X-rays consistently underdiagnose pulmonary abnormalities or diseases in historically under-served patient populations, raising ethical concerns about the clinical use of such algorithms.
Subfields: Biomedical Engineering, Health Informatics, Radiology, Nuclear Medicine and Imaging
View Paper: https://www.semanticscholar.org/paper/ab5923948a3d8c15f964c1577eec2b3c379ba09e
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Biomedical Engineering
In Radiology, Nuclear Medicine and Imaging
Citations in the corpus
- AI recognition of patient race in medical imaging: a modelling study (2021, neutral): “9 Importantly, if used, such models would lead to more patients who are Black and female being incorrectly identified as healthy compared with patients who are White and male.”
- Bias in AI-based models for medical applications: challenges and mitigation strategies (2023, neutral): “Bias is usually de fi ned as a difference in performance between sub-groups for a predictive task 6,7 .”
- Coarse race data conceals disparities in clinical risk score performance (2023, background): “…by using coarse categories can exceed the coarse racial variation that has been the focus of an enormous amount of work on algorithmic fairness in medicine (Chen et al., 2018; Obermeyer et al., 2019; Zink et al., 2023; Boulware et al., 2021; Seyyed-Kalantari et al., 2021; Adam et al., 2022).”
- Artificial intelligence for digital and computational pathology (2023, background): “P a t c h i m p o r t a n c e”
- A Survey of Large Language Models for Healthcare: from Data, Technology, and Applications to Accountability and Ethics (2023, neutral): “As an example, neural models trained on publicly accessible chest X-ray datasets tend to exhibit un-derdiagnosis tendencies in marginalized communities, including female patients, Black patients, Hispanic patients, and those covered by Medicaid insurance [188].”
- Algorithm fairness in artificial intelligence for medicine and healthcare (2023, background): “…Institutes of Health (NIH) ChestX-ray 15 ) underdiagnose underserved populations; that is, the likelihood is greater of the algorithms incorrectly predicting ‘no symptomatic conditions (findings)’ for female patients, Black patients, Hispanic patients and patients with Medicaid insurance 11 .”