What is it about?
We have made stratification an accessible approach to adjust for confounding when the outcome is continuous. Our proposed stratification approach allows researchers to first focus on assessing the presence and direction of an association while avoiding model misspecification, leaving the quantification of the exposure effect on outcome to the residual diagnostic stage.
Featured Image
Why is it important?
To ascertain the relationship between the exposure and continuous outcome, adjusting for confounders is an important step in many epidemiological studies. The stratification approach for confounder adjustment is robust to model misspecification when compared with the commonly used approach that includes the confounders as predictors in the regression models. Our work extends the commonly used regression models in matched designs for binary and survival outcomes to continuous outcomes.
Perspectives
The work from this article is made possible by standing on the shoulders of giants from multiple disciplines such as economics, epidemiology and statistics. With the aim to address the issue of confounding in research projects dealing with continuous outcomes while minimizing model misspecification, it was amazing to witness the same underlying regression model that was used in, for example, matched case-control studies, being used to analyse continuous outcomes when the stratification approach is employed to account for confounding. I hope that you will find this intriguing and the proposed stratification approach useful in your future data analyses.
Chuen Seng Tan
National University of Singapore
Read the Original
This page is a summary of: A stratification approach using logit-based models for confounder adjustment in the study of continuous outcomes, Statistical Methods in Medical Research, December 2017, SAGE Publications,
DOI: 10.1177/0962280217747309.
You can read the full text:
Resources
Contributors
The following have contributed to this page
