News
December 11, 2019

Algorithmovigilance: Regenstrief experts lead discussion on monitoring AI in healthcare

image of computer algorithm closeup

Four Regenstrief research scientists and leaders provided insight into the need to monitor the use of artificial intelligence algorithms in healthcare at the American Medical Informatics Association’s Health Informatics Policy Forum.

The forum considered the future of clinical decision support in the era of big data and machine learning.  Regenstrief leaders President and Chief Executive Officer Peter Embí, M.D., M.S.; Vice President for Data and Analytics Shaun Grannis, M.D., M.S.; Vice President for Research Development Eneida Mendonca, M.D., PhD; and Chief Information Officer Umberto Tachinardi, M.D., MSc, engaged in the panel titled “Algorithmovigilance.”

As the panel moderator, Dr. Embí, who is also the outgoing chair of AMIA, defined algorithmovigilance as the systematic monitoring of computable algorithms to evaluate and respond to expected or unexpected health effects. He said monitoring is crucial to maintain quality, minimize harm and promote trust in healthcare AI. Dr. Embí stated that this issue is becoming increasingly important as more machine learning-derived algorithms are developed and used.

Dr. Mendonca spoke about some of the unintended consequences of machine learning tools. Because AI systems are constrained by learning from observational data, they will have an amplifying effect on existing behaviors, possibly exacerbating bias or other detrimental practices. She described the risk of using algorithms for a different purpose than the one they were created for and touched on other issues including the possibility of cyberattacks on these systems.

Dr. Grannis spoke about how information entropy can affect the functionality of algorithms, which is another reason that they need to be monitored and adjusted. Dr. Tachinardi discussed the difference between data needed for research and data needed by clinicians. Jordan Platt, PhD, MPH from the University of Michigan also participated in the panel.

The AMIA Health Informatics Policy Forum convened about 150 experts, stakeholders, and policymakers to discuss the opportunities and challenges related to the use of clinical decision support. The event emphasized how the informatics community recommends adaptive clinical decision support be developed, tested, deployed, monitored and evaluated.

About Regenstrief Institute

Founded in 1969 in Indianapolis, the Regenstrief Institute is a local, national and global leader dedicated to a world where better information empowers people to end disease and realize true health. The Regenstrief Institute and its researchers are responsible for a growing number of major healthcare innovations and studies. Examples range from the development of global health information technology standards that enable the use and interoperability of electronic health records to improving patient-physician communications, to creating models of care that inform practice and improve the lives of patients around the globe.

A key research partner to Indiana University, Regenstrief Institute is celebrating 50 years of healthcare innovation. Sam Regenstrief, a successful entrepreneur from Connersville, Indiana, founded the institute with the goal of making healthcare more efficient and accessible for everyone. His vision continues to guide the institute’s research mission.

Related News

Aaron Carroll, M.D.

Call for Papers on Artificial Intelligence Applied to Pediatric Care

Published in JAMA Pediatrics. Here is a link to the article. Regenstrief Institute author: Aaron E. Carroll, M.D., M.S.

Critical gap in colorectal cancer screening process: follow-up after positive stool test

INDIANAPOLIS – Screening for colorectal cancer presents a unique opportunity unavailable for most other types of cancers. Screening allows