ABC of Epidemiology Linear and logistic regression analysis

Raw Text

Kidney International

Volume 73, Issue 7

Author links open overlay panel

G. Tripepi 1

K.J. Jager 2

F.W. Dekker 2 3

C. Zoccali 1

Show more

Add to Mendeley

Share

Cite

https://doi.org/10.1038/sj.ki.5002787

Get rights and content

Under an Elsevier

user license

In previous articles of this series, we focused on relative risks and odds ratios as measures of effect to assess the relationship between exposure to risk factors and clinical outcomes and on control for confounding. In randomized clinical trials, the random allocation of patients is hoped to produce groups similar with respect to risk factors. In observational studies, exposed and unexposed individuals may differ not only for the presence of the risk factor being tested but also for a series of other factors that are potentially related to the study outcome, thus making ‘confounding’ very likely. One of the most important uses of multivariate modeling is precisely that ‘of controlling for confounding’ to let emerge the effect of the risk factor of interest on the study outcome. In this paper, we will discuss linear regression analysis for the examination of continuous outcome data and logistic regression analysis for the study of categorical outcome data. Furthermore, we focus on the most important application of multiple linear and logistic regression analyses.

Previous article in issue

Next article in issue

KEYWORDS

epidemiology

linear regression analysis

logistic regression analysis

confounding

Recommended articles

Cited by (0)

Copyright © 2008 International Society of Nephrology. Published by Elsevier Inc. All rights reserved.

Single Line Text

Kidney International. Volume 73, Issue 7. Author links open overlay panel. G. Tripepi 1. K.J. Jager 2. F.W. Dekker 2 3. C. Zoccali 1. Show more. Add to Mendeley. Share. Cite. https://doi.org/10.1038/sj.ki.5002787. Get rights and content. Under an Elsevier. user license. In previous articles of this series, we focused on relative risks and odds ratios as measures of effect to assess the relationship between exposure to risk factors and clinical outcomes and on control for confounding. In randomized clinical trials, the random allocation of patients is hoped to produce groups similar with respect to risk factors. In observational studies, exposed and unexposed individuals may differ not only for the presence of the risk factor being tested but also for a series of other factors that are potentially related to the study outcome, thus making ‘confounding’ very likely. One of the most important uses of multivariate modeling is precisely that ‘of controlling for confounding’ to let emerge the effect of the risk factor of interest on the study outcome. In this paper, we will discuss linear regression analysis for the examination of continuous outcome data and logistic regression analysis for the study of categorical outcome data. Furthermore, we focus on the most important application of multiple linear and logistic regression analyses. Previous article in issue. Next article in issue. KEYWORDS. epidemiology. linear regression analysis. logistic regression analysis. confounding. Recommended articles. Cited by (0) Copyright © 2008 International Society of Nephrology. Published by Elsevier Inc. All rights reserved.