51视频

Universal Journal of Applied Mathematics Vol. 11(1), pp. 8 - 15
DOI: 10.13189/ujam.2023.110102
Reprint (PDF) (222Kb)

Confounding and Effect Modification in Biostatistics: An Overview

M谩rio Basto ^1,*, Teresa Abreu ¹, Ricardo Gon莽alves ¹, Jos茅 M. Pereira ^2
1 Higher School of Technology, Polytechnic Institute of C谩vado and Ave, Barcelos, Portugal
² CICF - Research Center on Accounting and Taxation, Polytechnic Institute of C谩vado and Ave, Barcelos, Portugal

ABSTRACT

Understanding concepts like confounding and effect modification is essential to biostatistics due to their potential influence on the interpretation of statistical results. The ability to appropriately identify and comprehend the links between research variables relies on having a firm grasp of these ideas. Statistical methods, such as stratification-related procedures and logistic regression, can be used to account for potential confounding factors. This helps to determine the real link between two variables of interest by controlling for the potentially skewing effects of the confounding variables. The ability to identify subsets of the population that may be more or less receptive to an intervention necessitates an understanding of effect modification. If, for instance, one learns that regular exercise is particularly useful in warding off heart disease in younger people, one could direct preventative efforts toward them. This paper aims to highlight the existence of associations between two binary variables that may be misleading or distorted due to the existence of confounding or effect modifier variables, that must be accounted for. The Mantel-Haenszel analysis and logistic regression are two techniques addressed in this study that help to statistically adjust the association between variables. Typically, only one of these methods is used in a study. This paper contrasts and illustrates the application of both techniques. To do this, four hypothetical situations are examined in order to provide the researcher with a comparative analysis of the two procedures and how to interpret the outcomes from each. Although the data may need to be adapted for each of the analyses, the outcomes are usually the same. The results indicate that both approaches are useful as long as they are used correctly. Nevertheless, depending on the situation, one or the other may be advisable.

KEYWORDS
Odds Ratio, Confounding, Interaction, Effect Modification, Mantel-Haenszel Analysis, Logistic Regression

Cite This Paper in IEEE or APA Citation Styles
(a). IEEE Format:
[1] M谩rio Basto , Teresa Abreu , Ricardo Gon莽alves , Jos茅 M. Pereira , "Confounding and Effect Modification in Biostatistics: An Overview," Universal Journal of Applied Mathematics, Vol. 11, No. 1, pp. 8 - 15, 2023. DOI: 10.13189/ujam.2023.110102.

(b). APA Format:
M谩rio Basto , Teresa Abreu , Ricardo Gon莽alves , Jos茅 M. Pereira (2023). Confounding and Effect Modification in Biostatistics: An Overview. Universal Journal of Applied Mathematics, 11(1), 8 - 15. DOI: 10.13189/ujam.2023.110102.