Examkin

Question

Difficulty: HardBias, Confounding, and Validity

Investigators at a regional academic medical center are exploring the association between a newly marketed combined oral contraceptive and the incidence of venous thromboembolism. Initial analysis of the cohort yields a crude relative risk of 2.1 for venous thromboembolism among contraceptive users compared to non-users. The research team subsequently stratifies the participants based on the presence of the prothrombin G20210A mutation. In the stratum of participants without the mutation, the relative risk is 1.1. In the stratum of participants with the mutation, the relative risk is 6.4. Based on these statistical findings, which of the following conclusions and subsequent analytical steps are appropriate? (Select two that apply)

  1. The genetic mutation is demonstrating effect modification on the relationship between the contraceptive and thromboembolismAnswer
  2. B
    The genetic mutation is acting as a primary confounder in this cohort study
  3. The stratum-specific relative risks should be presented individually rather than calculating a single pooled estimateAnswer
  4. D
    Statistical adjustment techniques should be applied to generate a single pooled relative risk
  5. E
    Future studies should match participants on mutation status to eliminate this statistical distortion

Answer

The genetic mutation is demonstrating effect modification, and the stratum-specific relative risks should be presented individually.
The scenario describes a situation where the magnitude of the exposure's effect on the outcome differs significantly depending on the presence of a third variable (the prothrombin G20210A mutation). Because the relative risks in the two strata (1.1 and 6.4) are markedly different from each other, this third variable is acting as an effect modifier. This represents a true biological interaction rather than a statistical distortion. When effect modification is present, combining or pooling the data obscures the true relationship; therefore, the correct analytical approach is to report the stratum-specific risks separately to highlight the differing risk profiles.

Step-by-Step Solution

1
Analyze the crude relative risk provided in the scenario.
The initial unadjusted relative risk is 2.1.
This establishes the baseline association before accounting for any third variables.
2
Evaluate the stratum-specific relative risks.
The relative risk for participants without the mutation is 1.1, while the relative risk for those with the mutation is 6.4.
Comparing stratum-specific measures of association to each other and to the crude measure is the standard method for distinguishing between confounding and effect modification.
3
Interpret the statistical phenomenon occurring.
Because the stratum-specific relative risks (1.1 and 6.4) are markedly different from one another, the third variable (prothrombin mutation) is changing the magnitude of the exposure's effect.
A significant difference between stratum-specific risks defines effect modification (interaction), whereas if they were similar to each other but different from the crude risk, it would indicate confounding.
4
Determine the appropriate analytical reporting step.
The researchers must report the stratum-specific risks separately.
Effect modification is a biological reality, not an error or bias. Pooling the results would hide the fact that the drug carries a much higher risk in a specific genetic subpopulation.

Key Concept

Effect modification occurs when the effect of an exposure on an outcome is altered by a third variable. It is identified when stratified analysis reveals significantly different measures of association across strata, and it should be reported using stratum-specific estimates.
Rate this question

Topics

Bias, Confounding, and Validity

More questions from this topic