American Medical Informatics Association Academic Forum Best Paper Award
Development of an Online Training Module on Algorithmic Bias in Health Care for Clinicians
Given the digital paradigm shift of health care and the increased data required in medical decision making, health care and clinicians are increasingly relying on artificial intelligence (AI) systems and this trend is expected to continue. However, clinicians currently do not receive training on how to best use AI systems or interpret their output. One of the concerns is that biases associated with the use of AI, algorithmic bias, will result in harms to patients, especially if clinicians are not sufficiently aware of the problem or how to work with AI systems for clinical care. Therefore, an online module was developed to train clinicians on recognizing and avoiding algorithmic bias in the use of artificial intelligence systems in health care including a review of basic concepts of algorithmic bias, recommendations on avoiding or minimizing algorithmic bias, and a set of cases to work through structured around real scenarios.