Review Question - QID 104072

The described study was subject to confounding, because the increase in coffee consumption was associated with increased cancer, however there was an additional variable that more likely caused the increase in lung cancer. This other variable was smoking which was the confounding variable that threw off the interpretation of the results in this study.

This study showed a classic example of confounding. The observed variable, coffee consumption, turned out to be associated with smoking, which was not a condition considered in the chain of analysis. Accordingly, the study suggested an association between coffee and cancer which did not directly exist. Confounding occurs when a third factor is either positively or negatively associated with both the exposure and outcome of interest. Confounding can distort the true association between exposure and outcome. In non-observational studies, randomization is an important tool to help reduce or remove confounding bias.

Boyko discusses the problem of confounding in observational research. Since observational research does not have the benefit of randomization, there is an increased risk of confounding bias. Exposure to risk factors can be due to self-selection, occupation, or a number of other pre-existing conditions that cannot be readily controlled for via randomization. If the potential confounding variable can be identified, statistical techniques exist to correct for the bias. However, if the potential confounder cannot be identified, there is no widely accepted statistical technique for adjusting for potential confounding.

Hallas and Pottegård discuss some of the study design methods that may be utilized to control confounding bias in observational studies. For case-control or cohort design studies, it is difficult to control for confounding, and the potential confounders are specific to the particular association under investigation. However, using a crossover design where possible can help reduce confounding. In a crossover study every participant is exposed at some point to the test condition and the control condition allowing each subject to serve as their own control which could rule out any confounding variables.

Illustration A is a schematic diagram depicting the influence of confounding in observational studies.

Incorrect Answers:
Answer 1: Observer bias occurs when the observer is aware of which arm of the study the subject is in, and the observer interprets results in light of this awareness. Double blind analysis helps reduce observer bias.
Answer 2: Lead time bias occurs with earlier detection of a disorder which makes the prognosis seem better when in reality the disease was merely detected earlier. Placebo control is not directly related to lead time bias.
Answer 3: Selection bias occurs when the selected subjects are not representative of the population to be studied and can be controlled by randomization and random sampling.
Answer 4: Measurement bias is a systematic bias in regards to classifying subjects. Measurement bias can be reduced by using validated measurement techniques.