Course Description

Applied Logistic Regression


ABSTRACT

The aim of this course is to provide theoretical and practical training for biostatisticians and professionals of related disciplines in statistical modeling using logistic regression. The increasingly popular logistic regression model has become the standard method for regression analysis of binary data in the health sciences. The course is based on the text Applied Logistics Regression, 3rd Ed, Hosmer D, Lemeshow S and Sturdivant R. A Wiley-Interscience Publication, John Wiley & Sons, New York, NY, 2013.


OUTLINE

A. Introduction

B. Fitting the Logistic Regression Model

C. The Multiple Logistic Regression Model

D. Interpretation of Coefficients

E. The Multivariate Case: Statistical Adjustment

F. Interaction and Confounding

G. Stratified Analysis via Logistic Regression

H. Numerical Problems

I. Summary Measures of Goodness-of-Fit

J. Area Under the ROC Curve

K. ICU Example


INSTRUCTOR BIOS

David W. Hosmer, Ph.D. is Professor (Emeritus) of Biostatistics, Department of Public Health, School of Public Health and Health Sciences, University of Massachusetts/Amherst and Adjunct Professor of Statistics, Department of Mathematics and Statistics, University of Vermont. He is a coauthor of Applied Logistic Regression and Applied Survival Analysis: Regression Modeling of Time to Event Data, both published by John Wiley & Sons Inc. Dr. Hosmer has taught short courses on these subjects in the US and abroad for over 25 years.

Stan Leneshow, Ph.D., is Founding Dean of OSU’s College of Public Health, serving in that capacity for 10 years. He has been with the University since 1999 as a biostatistics professor in the School of Public Health and the Department of Statistics, director of the biostatistics core of the Comprehensive cancer Center and Director of the University’s Center for Biostatistics. His biostatistics research includes statistical modeling of medical data, sampling, health disparities and cancer prevention. He has published extensively in the applied and methodological literature and has coauthored three textbooks in the John Wiley & Sons Wiley statistics series, a leading publisher for the scientific, technical and medical communities worldwide. The textbooks he authored are: Applied Logistic Regression (now in its 3rd edition), Applied Survival Analysis (now in its 2nd edition) and Sampling Populations; Methods and Applications (now in its 4th edition). In 2003, Dr. Lemeshow was awarded the Wiley Lifetime Award. Professor Lemeshow maintains an ongoing relationship with Aarhus University, Denmark as an Honorary Professor in Biostatistics and is a member of the faculty of the Erasmus Summer Program, Rotterdam, Holland. He has taught more than 100 short courses on biostatistical methods in the country and abroad, including eight European countries, Australia, China, and India. See more at http://cph.osu.edu/people/slemeshow#sthash.bLMUoLJ6.dpuf

Rod Sturdivant, Ph.D., is an Associate Professor of Clinical Public Health in the Biostatistics Division of the College of Public Health at The Ohio State University. He assumed this position in 2013 after a distinguished 27 year career as an officer in the U.S. Army, retiring as a Colonel. Dr. Sturdivant holds a master’s degree in statistics and in operations research from Stanford and a Ph.D. in biostatistics from the University of Massachusetts – Amherst. He was selected as an Academy Professor in the Department of Mathematical Sciences at the United States Military Academy, West Point, where he completed his military service. During his time at West Point, Dr. Sturdivant was recognized as a Professor of Applied Statistics for his excellence across domains as a senior leader at the academy. He founded and directed the West Point Center for Data Analysis and Statistics (CDAS) and led projects to infuse technology and active, real world, problem driven curriculum into statistics and mathematics education. He has numerous publications in educational journals and has given presentations and workshops at national conferences involving math and statistics education. He is the co-principle investigator on one National Science Foundation (NSF) grant to develop resources for undergraduate statistics courses. He has published and presented extensively in the field of applied statistics.