Examkin

Question

Difficulty: EasyBias, Confounding, and Validity

A retrospective cohort study investigates the association between a new oral hypoglycemic agent and the risk of severe hypoglycemia. The initial unadjusted analysis shows a relative risk (RR) of 2.8 (95% CI: 1.9 - 3.7). The researchers note that patients prescribed the new agent generally have a longer duration of diabetes. They decide to stratify the patients into two groups: those with diabetes for less than 10 years and those with diabetes for 10 years or more. Following stratification, the RR for severe hypoglycemia associated with the new agent is 1.0 (95% CI: 0.8 - 1.2) in both stratified groups. Which of the following best describes the role of diabetes duration in this study?

  1. Confounding variableAnswer
  2. B
    Effect modifier
  3. C
    Dependent variable
  4. D
    Cognitive bias
  5. E
    Measurement bias

Answer

Confounding variable
The correct answer is Confounding variable. Confounding occurs when a third variable is associated with both the exposure and the outcome, creating a false association. The classic method to identify confounding is stratification. If stratifying by the suspected third variable causes the original association to disappear (the relative risks in the strata become approximately 1.0), then that third variable is a confounder.

Step-by-Step Solution

1
Analyze the crude (unadjusted) relative risk.
The unadjusted RR is 2.8, suggesting a significant positive association between the new drug and severe hypoglycemia.
Establishing the baseline association is necessary before evaluating the impact of stratification.
2
Evaluate the results after stratifying by diabetes duration.
In both the <10 years and ≥10 years strata, the RR becomes 1.0, meaning there is no association within the specific subgroups.
Stratification helps distinguish whether a third variable is driving the observed crude association or modifying it.
3
Interpret the change in relative risk.
Because the association completely disappears upon stratification, the original crude association was entirely driven by the unequal distribution of the third variable (diabetes duration) between the exposure groups.
This pattern (crude association disappears upon stratification) is the defining epidemiological signature of a confounding variable.

Key Concept

Confounding versus Effect Modification
Estimated Time:1m 0s
Rate this question

Topics

Bias, Confounding, and Validity

More questions from this topic