Volume 47, Issue 6 p. 749-754

Assessment and Control for Confounding by Indication in Observational Studies

Bruce M. Psaty MD, PbD

Corresponding Author

Bruce M. Psaty MD, PbD

Cardiovascular Health Research Unit, the Department of Medicine, University of Washington, Seattle

Cardiovascular Health Research Unit, the Department of Epidemiology, University of Washington, Seattle

Cardiovascular Health Research Unit, the Department of Health Services, University of Washington, Seattle

Cardiovascular Health Research Unit, Suite

1360, 1730 Minor Ave., Seattle, WA 98101.

Search for more papers by this author
Thomas D. Koepsell MD, MPH

Thomas D. Koepsell MD, MPH

Cardiovascular Health Research Unit, the Department of Medicine, University of Washington, Seattle

Cardiovascular Health Research Unit, the Department of Epidemiology, University of Washington, Seattle

Cardiovascular Health Research Unit, the Department of Health Services, University of Washington, Seattle

Search for more papers by this author
Danyu Lin PhD

Danyu Lin PhD

Cardiovascular Health Research Unit, the Department of Biostatisrics, University of Washington, Seattle

Search for more papers by this author
Noel S. Weiss MD, DrPH

Noel S. Weiss MD, DrPH

Cardiovascular Health Research Unit, the Department of Epidemiology, University of Washington, Seattle

Search for more papers by this author
David S. Siscovick MD, MPH

David S. Siscovick MD, MPH

Cardiovascular Health Research Unit, the Department of Medicine, University of Washington, Seattle

Cardiovascular Health Research Unit, the Department of Epidemiology, University of Washington, Seattle

Search for more papers by this author
Frits R. Rosendaal MD, PhD

Frits R. Rosendaal MD, PhD

Department of Clinical Epidemiology, Leiden University Medical Center, Leiden, the Nethetlands

Search for more papers by this author
Marco Pahor MD

Marco Pahor MD

Department of Preventive Medicine, University of Tennessee, Memphis, Tennessee

Search for more papers by this author
Curt D. Furberg MD, PhD

Corresponding Author

Curt D. Furberg MD, PhD

Department of Public Health Sciences, Wake Forest University School of Medicine, Winston-Salem, North Carolina

Cardiovascular Health Research Unit, Suite

1360, 1730 Minor Ave., Seattle, WA 98101.

Search for more papers by this author
First published: 27 April 2015
Citations: 232

Presented at the National Meeting of the American Society of Hypertension, San Francisco, California, May 28–31, 1997.

The research reported in this article was supported, in part, by Grants HL40628 and HL43201 from the National Heart, Lung, Blood Institute, Grant AG09556 from the National Institute on Aging, Bethesda, Maryland, and grants from the NWO (Nederlandse Organisatie voor Wetenschappelijk Onderzoek) The Hague, The Netherlands. Dr. Psaty is a Merck/SER Clinical Epidemiology Fellow (cosponsored by the Merck Co. Foundation, Railway, NJ, and the Society for Epidemiologic Research, Baltimore, MD).

Abstract

In the evaluation of pharmacologic therapies, the controlled clinical trial is the preferred design. When clinical trial results are not available, the alternative designs are observational epidemiologic studies. A traditional concern about the validity of findings from epidemiologic studies is the possibility of bias from uncontrolled confounding. In studies of pharmacologic therapies, confounding by indication may arise when a drug treatment serves as a marker for a clinical characteristic or medical condition that triggers the use of the treatment and that, at the same time, increases the risk of the outcome under study. Confounding by indication is not conceptually different from confounding by other factors, and the approaches to detect and control for confounding — matching, stratification, restriction, and multivariate adjustment — are the same. Even after adjustment for known risk factors, residual confounding may occur because of measurement error or unmeasured or unknown risk factors. Although residual confounding is difficult to exclude in observational studies, there are limits to what this “unknown” confounding can explain. The degree of confounding depends on the prevalence of the putative confounding factor, the level of its association with the disease, and the level of its association with the exposure. For example, a confounding factor with a prevalence of 20% would have to increase the relative odds of both outcome and exposure by factors of 4 to 5 before the relative risk of 1.57 would be reduced to 1.00. Observational studies have provided important scientific evidence about the risks associated with several risk factors, including drug therapies, and they are often the only option for assessing safety. Understanding the methods to detect and control for confounding makes it possible to assess the plausibility of claims that confounding is an alternative explanation for the findings of particular studies. J Am Geriatr Soc 47:749–754, 1999.