A Philosophy for Dealing with Hypothesized Uncontrolled Confounding in Epidemiological Investigations

Citation

Blair, Aaron; Stewart, Walter F.; Stewart, Patricia A.; Sandler, Dale P.; Axelson, Olav; Vineis, Paolo; Checkoway, Harvey; Savitz, David A.; Pearce, Neal; & Rice, Carol (1995). A Philosophy for Dealing with Hypothesized Uncontrolled Confounding in Epidemiological Investigations. Medicina del Lavoro, 86(2), 106-110.

Abstract

Control for confounding and evaluation of the potential for uncontrolled confound­
ing is necessary to ensure validity in epidemiologic investigations. Well estab­lished risk factors that could confound can be handled in a relatively straightforward fashion in the design phase of the study and in statistical analyses. However, new hypotheses regarding potential cofnounders are often raised after data collec­tion is complete. In situations where con­founding was not controlled by design or statistical analysis, the issue becomes how to evaluate reported associations in epi­demiologic studies when uncontrolled confounding is raised as an alternative explanation.

Reference Type

Journal Article

Year Published

1995

Journal Title

Medicina del Lavoro

Author(s)

Blair, Aaron
Stewart, Walter F.
Stewart, Patricia A.
Sandler, Dale P.
Axelson, Olav
Vineis, Paolo
Checkoway, Harvey
Savitz, David A.
Pearce, Neal
Rice, Carol