Biostatistics 2
(ADVANCED COURSE)
The Biostatistics 2 course is designed as a continuation of Biostatistics 1, aimed at researchers, healthcare professionals, and scientists who wish to deepen their knowledge of statistical modeling and analysis. This course focuses on advanced statistical methods, including generalized linear models, logistic regression, and survival analysis. In addition to covering these advanced topics, the course also introduces the basics of SAS® programming, providing participants with the necessary skills to implement these statistical techniques in practice.
Through a combination of theoretical concepts and practical examples, participants will learn how to select and apply appropriate statistical models, assess model performance, and interpret results effectively using SAS®. By the end of the course, participants will be equipped to handle complex data sets, evaluate key assumptions, and make informed decisions based on advanced statistical analyses.
Course Program
1. General Linear Models (GLM models)
a. Multiple Linear Regression
– Diagnostics
– Criteria for model selection (adjusted R-square, AIC, AICC, SBC)
– Mallows’ Cp and Hocking’s criterion
b. Analysis of Covariance
c. Analysis of Variance
– LSMEANS and MEANS – what’s the difference?
– What if we are interested in the difference between specific groups or the difference between two groups compared to three?
2. Logistic Regression
– OLS vs. Maximum Likelihood Estimates
– Binary logistic regression with multiple predictors
– Odds Ratio in a logistic model with multiple predictors (qualitative and quantitative)
– Concept of concordance and prediction accuracy
– Criteria for model selection
3.Model Evaluation and ROC Curve
– Classification table
– Sensitivity and Specificity
– Youden Index and selection of optimal cutoff
4. Violated Assumptions and Consequences for Model Execution
5. The Problem of Outliers
6. The Problem of Multicollinearity in Regression
7. Confounding and Interactions in Models
8. Stratified Contingency Table Analysis (Cochran-Mantel-Haenszel Statistic)
9. Survival Analysis
– Non-parametric Survival Analysis
– Cox Proportional Hazards Model
– Kaplan-Meier Curve