Location

Rochester, Minnesota

Contact

malik.momin@mayo.edu

SUMMARY

While there is growing interest in healthcare artificial intelligence (AI), much of the evidence supporting its benefits comes from proof-of-concept work. Under current standards, it can be easy to demonstrate strong performance of AI models in development settings. This means that proof-of-concept work is not enough to know if an AI model will work in real-world clinical applications.

The research of Momin M. Malik, Ph.D., examines ways that AI can be less effective in practice than expected and how researchers can develop better ways to catch this in advance. Dr. Malik explores this concept broadly across healthcare and other disciplines. He has applied this approach specifically in pediatric and adolescent medicine to topics such as care for patients with asthma, celiac disease, congenital heart disease and diabetes.

Focus areas

  • Socioeconomic variations in AI. Dr. Malik measures and understands the sources of socioeconomic variation in AI models of children with asthma by looking at health drivers in data coverage and detail in electronic health records.
  • AI model performance and better health for all. Dr. Malik adapts existing statistical methods and creates new statistical methods to rigorously evaluate AI model performance and promote better health for everyone. His work includes conducting uncertainty quantification and sample size calculations.
  • Monitoring of AI models. Dr. Malik creates frameworks and methods for monitoring AI models after they are put into practice. He considers potential feedback loops and potentially unavailable confirmation for negatively labeled instances.

Significance to patient care

AI programs that do not work can harm patient health, worsen outcomes, and waste time and money for patients, care teams and health systems. Current ways of testing AI tools are limited, which can make it seem as though AI will work when it may not.

Using better methods to test if AI works and paying attention to those results can help reduce the risk of harm and waste as AI is developed and used in real life.

PROFESSIONAL DETAILS

Primary Appointment

  1. Associate Consultant I, Pediatric Research, Department of Pediatrics

Academic Rank

  1. Assistant Professor of Biomedical Informatics

EDUCATION

  1. MS - Machine Learning Machine Learning Department, School of Computer Science
  2. PhD - Societal Computing School of Computer Science, Carnegie Mellon University
  3. MS - Computation, Organizations and Society School of Computer Science, Carnegie Mellon University
  4. MSc - Social Science of the Internet Oxford Internet Institute, University of Oxford
  5. AB - History and Science Harvard University

Clinical Studies

Learn about clinical trials that address specific scientific questions about human health and disease.

See my studies.

Explore all research studies at Mayo Clinic.

Publications

See the peer-reviewed findings I have published as a result of my research.

Review publications.
.
BIO-20601005

Mayo Clinic Footer