Understanding Confounding in Research

Abstract

1. Kantahyanee W. Murray, PhD* 2. Anne Duggan, ScD† 1. *Ruth H. Young Center for Children and Families, University of Maryland School of Social Work, Baltimore, Md. 2. †Department of General Pediatrics & Adolescent Medicine, The Johns Hopkins University School of Medicine, Baltimore, Md. The association between indoor smoking and asthma exacerbation is well established. You learn that one of your asthmatic patients is exposed regularly to indoor smoking. Your patient's mother attributes her daughter's asthma exacerbations to the high level of air pollution in the community in which they live. You think that indoor smoking is a major contributor to her asthma exacerbations but are unable to convince the mother. You decide to search the literature for studies that have examined whether air pollution confounds the relationship between indoor smoking and asthma exacerbations. When establishing relationships of cause and effect, internal validity represents the truthfulness of conclusions about causal relationships. Internal validity means that a true cause-and-effect relationship exists between an exposure (the cause) and outcome (the effect) variable. Confounding is one of several threats to the internal validity of a research study. (1) Confounding is defined as a possible source of bias in studies in which an unmeasured third variable (the confounder) is related to the exposure of interest (although not causally) and causally related to the outcome of interest. (2) Understanding confounding is critical in determining what inferences can be drawn from study findings. Is indoor smoking a cause of asthma exacerbation or is a third factor (eg, air pollution) the true cause? Failure to detect confounding can lead …