Privacy-preserving machine learning across hospitals to reduce bias

Federated learning algorithms to overcome statistical and algorithmic bias and privacy concerns in machine learning for health

NIH-funded research Columbia University Health Sciences · NIH-11194404

This project builds methods so hospitals can train shared computer programs on medical records without sharing raw data, aiming to make those programs fairer for different patient groups.

Quick facts

Grant typeR01 grant
Study typeNIH-funded research
Funding institutionColumbia University Health Sciences NIH-funded
Lab location1 site (New York, United States)
Project IDNIH-11194404 on NIH RePORTER

What this research studies

If you are a patient, this work aims to help hospitals build shared computer tools without sending your raw medical records elsewhere. The team will create new mathematical and algorithmic add-ons to federated learning that measure how different each site's data are and decide when adding a site helps or hurts the shared model. They will use techniques like optimal transport to compare data distributions and secure multiparty computation to keep information private during training. These methods will be tested using clinical datasets from multiple institutions to see if resulting models are more accurate and less biased.

Who could benefit from this research

Good fit: Patients whose medical records are stored at participating hospitals or health systems, especially those from diverse backgrounds, could be indirectly included in this work.

Not a fit: Patients whose health systems do not take part or those with extremely rare conditions not represented in partner datasets may not see benefit from these models.

Why it matters

Potential benefit: If successful, patients could get more accurate and fair clinical decision tools that protect their privacy when hospitals collaborate.

How similar studies have performed: Early federated learning projects in medicine show promise on imaging and EHR tasks, but combining bias-correction via optimal transport and strong privacy methods is relatively new and not yet widely proven.

Where this research is happening

New York, United States

Researchers

About this research

  1. This is an active NIH-funded research project — typically early-stage science, not a clinical trial accepting patient enrollment.
  2. Some NIH-funded labs run parallel clinical studies or seek volunteers for related work. To check, contact the principal investigator or institution listed above.
  3. For full project details, budget, and progress reports, visit the official NIH RePORTER page below.
Last reviewed 2026-06-13 by the Find a Trial editorial team. Information on this page is for educational purposes and is not medical advice. Always consult qualified healthcare professionals about clinical trial participation.