Confounding: What it is and how to deal with it

K. J. Jager, C. Zoccali, A. MacLeod, F. W. Dekker

Research output: Contribution to journalArticle

124 Citations (Scopus)

Abstract

As confounding obscures the 'real' effect of an exposure on outcome, investigators performing etiological studies do their utmost best to prevent or control confounding. Unfortunately, in this process, errors are frequently made. This paper explains that to be a potential confounder, a variable needs to satisfy all three of the following criteria: ( 1) it must have an association with the disease, that is, it should be a risk factor for the disease; ( 2) it must be associated with the exposure, that is, it must be unequally distributed between exposure groups; and ( 3) it must not be an effect of the exposure; this also means that it may not be part of the causal pathway. In addition, a number of different techniques are described that may be applied to prevent or control for confounding: randomization, restriction, matching, and stratification. Finally, a number of examples outline commonly made errors, most of which result from 'overadjustment' for variables that do not satisfy the criteria for potential confounders. Such an example of an error frequently occurring in the literature is the incorrect adjustment for blood pressure while studying the relationship between body mass index and the development of end-stage renal disease. Such errors will introduce new bias instead of preventing it.

Original languageEnglish
Pages (from-to)256-260
Number of pages5
JournalKidney International
Volume73
Issue number3
Early online date31 Oct 2007
DOIs
Publication statusPublished - Jan 2008

Keywords

  • confounding
  • randomization
  • matching
  • stratification
  • epidemiology
  • peritoneal-dialysis
  • controlled-trial
  • disease
  • hemodialysis

Cite this

Jager, K. J., Zoccali, C., MacLeod, A., & Dekker, F. W. (2008). Confounding: What it is and how to deal with it. Kidney International, 73(3), 256-260. https://doi.org/10.1038/sj.ki.5002650

Confounding : What it is and how to deal with it. / Jager, K. J.; Zoccali, C.; MacLeod, A.; Dekker, F. W.

In: Kidney International, Vol. 73, No. 3, 01.2008, p. 256-260.

Research output: Contribution to journalArticle

Jager, KJ, Zoccali, C, MacLeod, A & Dekker, FW 2008, 'Confounding: What it is and how to deal with it', Kidney International, vol. 73, no. 3, pp. 256-260. https://doi.org/10.1038/sj.ki.5002650
Jager KJ, Zoccali C, MacLeod A, Dekker FW. Confounding: What it is and how to deal with it. Kidney International. 2008 Jan;73(3):256-260. https://doi.org/10.1038/sj.ki.5002650
Jager, K. J. ; Zoccali, C. ; MacLeod, A. ; Dekker, F. W. / Confounding : What it is and how to deal with it. In: Kidney International. 2008 ; Vol. 73, No. 3. pp. 256-260.
@article{37641993d589429f819f47d723488bab,
title = "Confounding: What it is and how to deal with it",
abstract = "As confounding obscures the 'real' effect of an exposure on outcome, investigators performing etiological studies do their utmost best to prevent or control confounding. Unfortunately, in this process, errors are frequently made. This paper explains that to be a potential confounder, a variable needs to satisfy all three of the following criteria: ( 1) it must have an association with the disease, that is, it should be a risk factor for the disease; ( 2) it must be associated with the exposure, that is, it must be unequally distributed between exposure groups; and ( 3) it must not be an effect of the exposure; this also means that it may not be part of the causal pathway. In addition, a number of different techniques are described that may be applied to prevent or control for confounding: randomization, restriction, matching, and stratification. Finally, a number of examples outline commonly made errors, most of which result from 'overadjustment' for variables that do not satisfy the criteria for potential confounders. Such an example of an error frequently occurring in the literature is the incorrect adjustment for blood pressure while studying the relationship between body mass index and the development of end-stage renal disease. Such errors will introduce new bias instead of preventing it.",
keywords = "confounding, randomization, matching, stratification, epidemiology, peritoneal-dialysis , controlled-trial, disease, hemodialysis",
author = "Jager, {K. J.} and C. Zoccali and A. MacLeod and Dekker, {F. W.}",
year = "2008",
month = "1",
doi = "10.1038/sj.ki.5002650",
language = "English",
volume = "73",
pages = "256--260",
journal = "Kidney International",
issn = "0085-2538",
publisher = "Nature Publishing Group",
number = "3",

}

TY - JOUR

T1 - Confounding

T2 - What it is and how to deal with it

AU - Jager, K. J.

AU - Zoccali, C.

AU - MacLeod, A.

AU - Dekker, F. W.

PY - 2008/1

Y1 - 2008/1

N2 - As confounding obscures the 'real' effect of an exposure on outcome, investigators performing etiological studies do their utmost best to prevent or control confounding. Unfortunately, in this process, errors are frequently made. This paper explains that to be a potential confounder, a variable needs to satisfy all three of the following criteria: ( 1) it must have an association with the disease, that is, it should be a risk factor for the disease; ( 2) it must be associated with the exposure, that is, it must be unequally distributed between exposure groups; and ( 3) it must not be an effect of the exposure; this also means that it may not be part of the causal pathway. In addition, a number of different techniques are described that may be applied to prevent or control for confounding: randomization, restriction, matching, and stratification. Finally, a number of examples outline commonly made errors, most of which result from 'overadjustment' for variables that do not satisfy the criteria for potential confounders. Such an example of an error frequently occurring in the literature is the incorrect adjustment for blood pressure while studying the relationship between body mass index and the development of end-stage renal disease. Such errors will introduce new bias instead of preventing it.

AB - As confounding obscures the 'real' effect of an exposure on outcome, investigators performing etiological studies do their utmost best to prevent or control confounding. Unfortunately, in this process, errors are frequently made. This paper explains that to be a potential confounder, a variable needs to satisfy all three of the following criteria: ( 1) it must have an association with the disease, that is, it should be a risk factor for the disease; ( 2) it must be associated with the exposure, that is, it must be unequally distributed between exposure groups; and ( 3) it must not be an effect of the exposure; this also means that it may not be part of the causal pathway. In addition, a number of different techniques are described that may be applied to prevent or control for confounding: randomization, restriction, matching, and stratification. Finally, a number of examples outline commonly made errors, most of which result from 'overadjustment' for variables that do not satisfy the criteria for potential confounders. Such an example of an error frequently occurring in the literature is the incorrect adjustment for blood pressure while studying the relationship between body mass index and the development of end-stage renal disease. Such errors will introduce new bias instead of preventing it.

KW - confounding

KW - randomization

KW - matching

KW - stratification

KW - epidemiology

KW - peritoneal-dialysis

KW - controlled-trial

KW - disease

KW - hemodialysis

U2 - 10.1038/sj.ki.5002650

DO - 10.1038/sj.ki.5002650

M3 - Article

VL - 73

SP - 256

EP - 260

JO - Kidney International

JF - Kidney International

SN - 0085-2538

IS - 3

ER -