# Evaluate underdiagnosis and overdiagnosis bias of deep learning model on primary open-angle glaucoma diagnosis in under-served patient populations

arXiv: https://arxiv.org/abs/2301.11315 · HTML: https://mlsanity.com/paper/2301.11315/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 9 of its 39 references that are in the map.
- First posted on arXiv 26 Jan 2023.

## Builds on (most cited first, 9 of 9)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [ImageNet Large Scale Visual Recognition Challenge](https://mlsanity.com/paper/1409.0575/) (2014)
- [Densely Connected Convolutional Networks](https://mlsanity.com/paper/1608.06993/) (2016)
- [A Survey on Bias and Fairness in Machine Learning](https://mlsanity.com/paper/1908.09635/) (2019)
- [On Fairness and Calibration](https://mlsanity.com/paper/1709.02012/) (2017)
- [CheXclusion: Fairness gaps in deep chest X-ray classifiers](https://mlsanity.com/paper/2003.00827/) (2020)
- [Disc-aware Ensemble Network for Glaucoma Screening from Fundus Image](https://mlsanity.com/paper/1805.07549/) (2018)
- [Clinical Evidence Engine: Proof-of-Concept For A Clinical-Domain-Agnostic Decision Support Infrastructure](https://mlsanity.com/paper/2111.00621/) (2021)
- [Exploring Gender Disparities in Time to Diagnosis](https://mlsanity.com/paper/2011.06100/) (2020)

## Built on by (most cited first, 5 of 5)

- [What is Fair? Defining Fairness in Machine Learning for Health](https://mlsanity.com/paper/2406.09307/) (2024)
- [Evaluating and Mitigating Bias in AI-Based Medical Text Generation](https://mlsanity.com/paper/2504.17279/) (2025)
- [Improving Fairness of Automated Chest X-ray Diagnosis by Contrastive Learning](https://mlsanity.com/paper/2401.15111/) (2024)
- [MeCaMIL: Causality-Aware Multiple Instance Learning for Fair and Interpretable Whole Slide Image Diagnosis](https://mlsanity.com/paper/2511.11004/) (2025)
- [Demographically-Conditioned Synthetic Medical Images for Bias Mitigation and Bias Detection in Disease Classifiers](https://mlsanity.com/paper/2607.14984/) (2026)

## Abstract (arXiv)

In the United States, primary open-angle glaucoma (POAG) is the leading cause of blindness, especially among African American and Hispanic individuals. Deep learning has been widely used to detect POAG using fundus images as its performance is comparable to or even surpasses diagnosis by clinicians. However, human bias in clinical diagnosis may be reflected and amplified in the widely-used deep learning models, thus impacting their performance. Biases may cause (1) underdiagnosis, increasing the risks of delayed or inadequate treatment, and (2) overdiagnosis, which may increase individuals' stress, fear, well-being, and unnecessary/costly treatment. In this study, we examined the underdiagnosis and overdiagnosis when applying deep learning in POAG detection based on the Ocular Hypertension Treatment Study (OHTS) from 22 centers across 16 states in the United States. Our results show that the widely-used deep learning model can underdiagnose or overdiagnose underserved populations. The most underdiagnosed group is female younger (< 60 yrs) group, and the most overdiagnosed group is Black older (>=60 yrs) group. Biased diagnosis through traditional deep learning methods may delay disease detection, treatment and create burdens among under-served populations, thereby, raising ethical concerns about using deep learning models in ophthalmology clinics.