Examkin

Question

Difficulty: MediumBias and Confounding in Research

A team of public health researchers conducts a retrospective cohort study using electronic health records to evaluate the effectiveness of a novel biologic therapy compared to standard inhaled corticosteroids (ICS) for preventing severe asthma exacerbations. The primary outcome is the incidence of asthma exacerbations requiring emergency department visits over a 12-month period.

Initial analysis of the data yields a crude relative risk (RR) of 1.65, suggesting that patients receiving the novel biologic have a higher risk of exacerbations compared to those receiving standard ICS.

Concerned about potential baseline differences between the treatment groups, the researchers perform a stratified analysis based on the patients' history of hospitalizations for asthma in the prior year (a marker of baseline disease severity). The results of the stratified analysis are as follows:

- Patients with ≥1\geq 1 prior hospitalization: RR=0.75RR = 0.75
- Patients with 0 prior hospitalizations: RR=0.72RR = 0.72

Which of the following epidemiologic phenomena best explains the difference between the crude and stratified relative risks?

  1. ConfoundingAnswer
  2. B
    Effect modification
  3. C
    Selection bias
  4. D
    Recall bias
  5. E
    Lead-time bias

Answer

Confounding
The correct answer is confounding. In this study, confounding by indication occurred because patients with more severe asthma (who inherently have a higher risk of exacerbations) were preferentially prescribed the novel, stronger biologic therapy. This made the biologic look harmful in the crude analysis (RR = 1.65). However, when researchers stratified the patients by their baseline severity (prior hospitalizations), they removed the confounding effect. The resulting stratum-specific RRs (0.75 and 0.72) demonstrated that the biologic is actually protective across all severity levels. Because the stratified RRs are similar to each other but completely different from the initial crude RR, this confirms confounding, not effect modification.

Step-by-Step Solution

1
Analyze the crude relative risk (RR).
The crude RR is 1.65.
This initial value suggests that the novel biologic is associated with an increased risk of exacerbations compared to standard ICS.
2
Analyze the stratum-specific relative risks.
The RRs are 0.75 and 0.72 for the two severity strata.
Stratification by baseline severity shows that within each specific severity group, the novel biologic actually reduces the risk of exacerbations.
3
Compare the crude RR to the stratum-specific RRs to determine the underlying phenomenon.
The stratum-specific RRs are similar to each other (0.75 and 0.72) but markedly different from the crude RR (1.65).
When stratified analysis yields relative risks that are similar to each other but notably different from the crude relative risk, it is the hallmark of confounding.

Key Concept

Confounding occurs when a third variable is associated with both the exposure and the outcome, distorting the crude measure of association. It is identified when stratified measures of association are similar to each other but differ significantly from the crude measure.
Rate this question

Topics

Bias and Confounding in Research

More questions from this topic