SAS Essential Training: 2 Regression Analysis for Healthcare Research

Deepen your SAS knowledge by learning how to conduct a regression analysis of a health survey data center using this popular data analytics platform.

SAS is a venerable data analytics platform that boasts millions of users worldwide and a slew of useful features. In this course, instructor Monika Wahi helps you deepen your SAS knowledge by showing how to use the platform to conduct a regression analysis of a health survey data center. Throughout the course, Monika demonstrates how to conduct regression analyses and present your model results in tables. She shows how to develop and present a linear regression model using PROC GLM as part of a hypothesis-driven analysis; how to do a logistic regression model in both PROC GENMOD and PROC LOGISTIC; and how to present and interpret your linear and logistic regression models. To wrap up, she goes over issues in regression and provides a few helpful tips.

Introduction

• Introduction to the course
• What you should know

1. Preparing for Linear Regression

• Linear regression and hypothesis review
• Plots for testing assumptions
• Stepwise linear regression modeling
• Basic PROC GLM code
• Reading PROC GLM output

2. Linear Regression Modeling

• Linear regression model presentation
• Linear regression: Early models
• Linear regression: Round 1
• Linear regression: The final model
• Linear regression model metadata
• Linear regression model fit
• Interpreting linear regression model

3. Preparing for Logistic Regression

• Hypothesis and odds ratio review
• Outcome distribution
• Basic PROC LOGISTIC code
• Basic PROC LOGISTIC output
• Stepwise logistic regression modeling

4. Logistic Regression Modeling

• Logistic regression: Early models
• Logistic regression: Round 1
• Logistic regression: The final model
• Logistic regression model metadata
• AIC and AUC for model fit
• Interpreting the logistic regression model

5. Model Presentation

• Presenting linear regression models
• Excel for linear regression models
• Presenting logistic regression models
• Excel for logistic regression models

6. Issues in Regression

• Collinearity in stepwise regression
• Interaction review
• Interactions in linear regression
• Interactions in logistic regression
• Interactions: Stratum-specific estimates
• -2 log likelihood for model fit

7. Regression Tips

• Categorizing continuous outcomes
• Categorizing continuous covariates
• Flags for ordinal value levels
• Strategically collapsing categories
• Choosing reference groups
• Describe your regression analysis

Conclusion

• Review of the process
• Next steps