Chap T. Le, of the Division of Biostatistics at the University of Minnesota, will present:
Abstract: Many diagnostic biomarkers are on continuous scale. It is desirable and important to dichotomize such a diagnostic marker at an optimal cutoff point, to classify subjects as “diseased” or “healthy”, and to bring the result for use in medical practice. I have proposed such a method by maximizing the “Youden Index” (Le, SMMR, 2006), which integrates both sensitivity and specificity. However, that approaches only consider the relationship between the continuous marker and the disease status, leaving out the subjects’ characteristics – such as age, race, and other biomarkers, which could greatly improve diagnostic marker cutoff optimization. In this talk, I propose a logistic regression-based method that takes account of patients’ characteristics in the dichotomization of the continuous diagnostic biomarker under investigation. Such a comprehensive evaluation using patients’ characteristics in the dichotomizing efforts will be very informative and very efficient for clinical research in many fields which is important at the time we focus more and more in individualized or personalized medicine.
A social tea will be held at 3:00 p.m. in A434 Mayo. All are Welcome.