Discussion: Logistic Regression

Articles

Articles
The Field Epidemiology Manual provides a set of core training materials reflecting the essential competencies required by intervention epidemiologists.

Discussion: Logistic Regression

  • Hello everybody,

    The FEM includes a chapter about Logistic regression and I accepted to be the editor for this chapter. It is the only chapter approaching multivariate analysis using regression models. It is OK for me and I think it will be very useful for many people. However, For future editions of the FEM would be better to have a different approach. Logistic regression should be part of a broader chapter or "hyperchapter" dealing with multivariate analysis in general and including at least the main regression models used in epidemiology depending on the question to be answerd, the characteristics of our data and the study design (logistic, conditional logistic, cox and poisson regression models).

    May be the ECDC could bring this issue up in a near future.

  • Frenando,

    that is for sure very relevant comment. And thank you for that. We will consider it for (near future) development of the FEM wiki and it should be the task of the advisory board to draw directions of the development.

  • As I commented elsewhere I think it would be great to link to other resources within the chapters. An example for a great resource on categorical data analysis including logistic regression is the following lecture from UC Berkeley:

    webcast.berkeley.edu/course_details_new.php

  • Looking at this chapter of Logistic Regression and the discussion we had in 2010/2011, I propose we should look into this topic again and compare it with the explanatory approach we currently take in the MultiVariable Analysis module in EPIET.


    What I consider useful in that curriculum, there is first a review of stratification and how this influences your bi-variable analysis (effect modification, confounding, both or none). Then to describe lineair regression analysis, and again describe how stratification influences your analysis. 


    This will then provide a good basis for understanding:



    • when you need to go further to multivariable analysis (the chapter then needs to include explanation of decision tree after stratification: when to progress to logistic regresstion and when not)

    • the basic principle of simple lineair regression and stratified lineair regression


    From there on the jump to logistic models is made.

    What do you think? Ingredients to reshape this set of articles?