Integrating Machine Learning in Your Clinical Research

Machine Learning Tools for Clinical Researchers: A Pragmatic Approach Series

Machine learning analysis methods offer the opportunity to integrate and learn from large amounts of biological, clinical, and environmental data, and there is a growing interest in how these tools can be used to inform and individualize clinical decision making in a variety of disease areas. Machine learning can offer different, yet often complementary, insights compared to traditional statistical analyses to better understand heterogeneity in patient presentation, prognoses, and treatment response, generating critical data for precision medicine research. These methods can allow integration across diverse data types and large feature sets, overcoming some limitations of traditional tools to answer clinical questions. However, many clinical researchers have little exposure to machine learning methods, presenting a barrier to utilization of these tools themselves and/or to effective collaboration with methodologists in their own research.

The objectives of this series are to:

• Provide a background/foundation of knowledge regarding the use of machine learning tools in clinical questions
• Understand the strengths and limitations of these methods
• Recognize some real-world examples of applied machine learning methodology in clinical research
• Elucidate how machine learning can be used to advance precision medicine research

In this session, a panel discussion will focus on how researchers and clinicians at UNC can integrate machine learning techniques into their own clinical research. Are you a clinician with an idea for how patient care could be improved with computational decision support tools? Pitch your idea (5-10 minute overview) to assembled machine learning experts. Receive expert guidance and compete for funding from the UNC Program for Precision Medicine in Health Care for analytical support to develop your project.