Mount Sinai scientists introduce a new AI-based tool, AEquity, which helps decrease bias in clinical data to enhance equity, accuracy, and patient outcomes.

Researchers at Mount Sinai School of Medicine, Icahn have revealed AEquity, a novel instrument that identifies and eliminates bias in clinical data deployed to educate artificial intelligence models. The innovation will improve the validity and fairness of AI-driven clinical and financial healthcare applications.

The study that was published in the Journal of Medical Internet Research demonstrates that AEquity can examine images, patient records, and public health surveys to detect demographic imbalances and unintended inaccuracies. In case these biases are not controlled, they may result in incorrect diagnoses and disparities in patient outcomes in healthcare systems.

One of the authors of the research, Dr. Faris Gulamali, indicated that the tool would assist developers and health systems to identify bias and take the right actions. He clarified that he did not want the tools to be effective only for the groups that were more represented in the data.

AEquity is funded by the National Institutes of Health and has a broad diversity of machine learning models and can be applied on both inputs and outputs, including risk scores and diagnoses. Mount Sinai leaders feel this measure is a move that builds trust in AI and is pivotal toward developing more equitable healthcare systems.