Skip to Content
Interactive Textbook on Clinical Symptom Research Logo


Home Button

Statistical Models for Prognostication
Author Bio
Introduction
Predictions: Statistical Models
Insight: Statistical Models
Ingredients: Statistical Models
Theoretical Aspects
Central Concepts
Regression Models
Regression
Currently selected section: Problems: Practical Advice
Example 1
Example 2
Chapter 8: Statistical Models for Prognostication: Practical Advice for Regression Modeling
        

What Not To Do

  • Search among many non-linear/interaction terms and fit the model as if pre-specified. A strong shrinkage is usually required for the estimated regression coefficients.

  • Search for an optimal cut-off of a continuous variable, and report as if pre-specified. Statistical procedures are available to adjust the p-value (Mazumdar and Glassman, 2000), but the dichotomization of a continuous variable usually implies a substantial loss of information.

  • Use stepwise selection with the default significance level of 5% in a small data set (Harrell et al., 1984) (Steyerberg et al., 2001).

  • Present a regression formula when you expect clinicians to use the model. Other formats, such a score charts, nomograms, and tables are easier to grasp.

Previous Page