Statistics for clinicians

S4C is a course intended to give clinicians the tool to understand indications, limits and proper interpretation of the most commonly used statistical analysis. It will be of significant help for a proper understanding of the literature and in choosing the appropriate test for your own studies. Participants will have their hands on SPSS to execute every test and learn the proper interpretation of its outputs.

Basic Level

Testing the significance of association of 2 qualitative variables:

• Statistics for clinicians essential:

  • The null and the alternative hypothesis

  • The unilateral and the bilateral study design

  • Correcting P value for the post hoc analysis

  • How to read statistical tables

  • The degree of freedom

  • Choosing the appropriate statistical test

  • Normality and variance equality

  • The foundation of multivariate analysis

• Types of statistical tests: 

  • Bivariate versus multivariate 

  • Paired versus unpaired 

  • Parametric versus distribution-free tests. 

• Testing the association of 2 qualitative variables: 

  • Chi-square and corrected Chi-square 

  • Fisher’s exact tests. 

• The comparison of 2 means: Student’s test (unpaired).

One-way ANOVA, Post Hoc analysis, correlation coefficient, simple regression, paired tests:

• The comparison of several means: One-way ANOVA and post-Hoc tests. 

• Testing the association of 2 quantitative variables: 

  • Correlation coefficient 

  • Linear regression analysis. 

Practical part I: Hands-on SPSS 

Paired designs: 

• paired Student test 

• McNemar’s test. 

Non-parametric tests: 

• Introduction to Distribution-free tests: 

  • Mann-Whitney test 

  • Kruskall Wallis test 

  • Wilcoxon sign test 

Time to event studies:

• Kaplan Meier 

• Log rank test 

Common indices of trial outcomes

• Relative and absolute risk 

• Odds ratio 

• Number needed to treat 

• Sensitivity 

• Specificity 

• Positive and negative predictive values 

• ROC analysis.

The multivariate analysis

• Multiple linear regression

• Logistic regression 

• Cox regression analysis