What is it about?
This paper reviews how bias in electronic health records (EHRs) can affect healthcare and AI systems. It identifies six main sources of bias, including problems with data quality, clinician judgment, patient referrals, and algorithm design. The review shows that these biases often go unaddressed and can lead to unfair treatment and worse outcomes for certain groups.
Featured Image
Photo by National Cancer Institute on Unsplash
Why is it important?
Inherent biases can really hinder the performance of models or decision support systems which are trained or built using data, specifically electronic health records. These biases are inherent, and often times are hidden in the information we use for the development of these systems, which perpetuate the bias into clinical care.
Perspectives
What stood out to me in this paper is how deeply bias is embedded not just in the data, but in the entire healthcare system — from the way patients are referred and diagnosed to how algorithms are designed. It made me realize that building fair AI in medicine isn't just a technical challenge, it's a structural and human one. The fact that many biases, especially those tied to gender, social status, or unconscious clinician behavior, are rarely studied shows a major gap in current research. This paper reinforced for me the need to approach medical AI with humility, critical thinking, and a constant awareness of the real-world systems it's built on.
oriel perets
Ben-Gurion University of the Negev
Read the Original
This page is a summary of: Inherent Bias in Electronic Health Records: A Scoping Review of Sources of Bias, ACM Transactions on Intelligent Systems and Technology, September 2026, ACM (Association for Computing Machinery),
DOI: 10.1145/3757924.
You can read the full text:
Contributors
The following have contributed to this page
