Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Algorithms Trained on Normal Chest X-rays Can Predict Health Insurance Types

Chi-Yu Chen, Rawan Abulibdeh, Arash Asgari, Sebastián Andrés Cajas Ordóñez, Leo Anthony Celi, Deirdre Goode, Hassan Hamidi, Laleh Seyyed-Kalantari, Ned McCague, Thomas Sounack, Po-Chih Kuo

TL;DR

<3-5 sentence high-level summary>- The paper investigates whether normal chest X-ray images encode socioeconomic information by attempting to predict health insurance type, a proxy for SES. It compares three state-of-the-art vision models across two large datasets (MIMIC-CXR-JPG and CheXpert) and shows that the models achieve meaningful AUCs (around 0.70 and 0.68, respectively) even on radiographically normal images. Through patch-based localization and demographic-mediator analyses, the authors argue that the predictive signal is diffuse and not primarily mediated by basic demographics, suggesting environmental or acquisition-related sociodemographic fingerprints. The work calls for rethinking fairness in medical AI to disentangle and mitigate such embeddings rather than merely balancing datasets or thresholds.

Abstract

Artificial intelligence is revealing what medicine never intended to encode. Deep vision models, trained on chest X-rays, can now detect not only disease but also invisible traces of social inequality. In this study, we show that state-of-the-art architectures (DenseNet121, SwinV2-B, MedMamba) can predict a patient's health insurance type, a strong proxy for socioeconomic status, from normal chest X-rays with significant accuracy (AUC around 0.70 on MIMIC-CXR-JPG, 0.68 on CheXpert). The signal was unlikely contributed by demographic features by our machine learning study combining age, race, and sex labels to predict health insurance types; it also remains detectable when the model is trained exclusively on a single racial group. Patch-based occlusion reveals that the signal is diffuse rather than localized, embedded in the upper and mid-thoracic regions. This suggests that deep networks may be internalizing subtle traces of clinical environments, equipment differences, or care pathways; learning socioeconomic segregation itself. These findings challenge the assumption that medical images are neutral biological data. By uncovering how models perceive and exploit these hidden social signatures, this work reframes fairness in medical AI: the goal is no longer only to balance datasets or adjust thresholds, but to interrogate and disentangle the social fingerprints embedded in clinical data itself.

Algorithms Trained on Normal Chest X-rays Can Predict Health Insurance Types

TL;DR

<3-5 sentence high-level summary>- The paper investigates whether normal chest X-ray images encode socioeconomic information by attempting to predict health insurance type, a proxy for SES. It compares three state-of-the-art vision models across two large datasets (MIMIC-CXR-JPG and CheXpert) and shows that the models achieve meaningful AUCs (around 0.70 and 0.68, respectively) even on radiographically normal images. Through patch-based localization and demographic-mediator analyses, the authors argue that the predictive signal is diffuse and not primarily mediated by basic demographics, suggesting environmental or acquisition-related sociodemographic fingerprints. The work calls for rethinking fairness in medical AI to disentangle and mitigate such embeddings rather than merely balancing datasets or thresholds.

Abstract

Artificial intelligence is revealing what medicine never intended to encode. Deep vision models, trained on chest X-rays, can now detect not only disease but also invisible traces of social inequality. In this study, we show that state-of-the-art architectures (DenseNet121, SwinV2-B, MedMamba) can predict a patient's health insurance type, a strong proxy for socioeconomic status, from normal chest X-rays with significant accuracy (AUC around 0.70 on MIMIC-CXR-JPG, 0.68 on CheXpert). The signal was unlikely contributed by demographic features by our machine learning study combining age, race, and sex labels to predict health insurance types; it also remains detectable when the model is trained exclusively on a single racial group. Patch-based occlusion reveals that the signal is diffuse rather than localized, embedded in the upper and mid-thoracic regions. This suggests that deep networks may be internalizing subtle traces of clinical environments, equipment differences, or care pathways; learning socioeconomic segregation itself. These findings challenge the assumption that medical images are neutral biological data. By uncovering how models perceive and exploit these hidden social signatures, this work reframes fairness in medical AI: the goal is no longer only to balance datasets or adjust thresholds, but to interrogate and disentangle the social fingerprints embedded in clinical data itself.

Paper Structure

This paper contains 11 sections, 1 figure, 4 tables.

Table of Contents

  1. Introduction
  2. Methods
  3. Datasets
  4. Models and Metrics
  5. Experiments
  6. Health Insurance Prediction from CXRs
  7. Localization of insurance information on Xray - Patch-based training
  8. Experiments on Demographic Mediators
  9. Discussion
  10. Conclusion
  11. Dataset description

Figures (1)

  • Figure 1: (a) The left image provides an example of the Remove-One-Patch method. The right image shows insurance type prediction AUC when only the corresponding patch area in the 3x3 grid is removed. (b) The left image provides an example of the Keep-One-Patch method. The right image shows insurance type prediction AUC when only the corresponding patch area in the 3x3 grid is retained.