Lecture 'Analytical tools for clinically useful tests'
- For whom
- Alumni , Business , Employees , Students
- When
- 13-09-2024 from 12:00 to 13:30
- Where
- Auditorium A2 - Building S9 - Campus De Sterre, Krijgslaan 281, 9000 Gent
- Language
- English
- Organizer
- Department of Applied Mathematics, Computer Science and Statistics
- Contact
- els.goetghebeur@ugent.be
Tips by Lisa McShane (NCI, Associate Director, Division of Cancer Treatment and Diagnosis) for navigating through a sea of high-dimensional data and a jungle of powerful analytic tools to arrive at a clinically useful test
Clinically useful tests guide medical decisions for better health outcomes. Increasingly, high-dimensional data underlie their development and use, with data from omics assays, imaging, and wearable devices. Powerful data analysis tools, including statistical methods and machine learning to develop complex predictors of phenotypes or outcomes, led to optimism that medical test development could be greatly accelerated.
There were success stories, but few clinically useful tests given the investments and number of publications claiming to have developed useful predictors. Common pitfalls in the design, analysis, and interpretation of studies set to develop complex predictors are highlighted through examples in oncology.
This talk promotes critical thinking when assessing evidence for medical tests and when planning new studies. Recommendations are given to improve the design, conduct, reporting, and value of research for medical tests from complex predictors from high-dimensional data.