Data from a multi-center retrospective cohort study are analyzed to determine the risk of … — USMLE Step 3

Data from a multi-center retrospective cohort study are analyzed to determine the risk of developing chronic obstructive pulmonary disease (COPD) following long-term occupational exposure to a specific industrial solvent. The investigators calculate a crude (unadjusted) relative risk of $4.2$ ( $95\%$ CI: $3.1$ - $5.6$ ) for COPD in exposed individuals compared to unexposed individuals.

Recognizing that tobacco use is a well-established contributor to respiratory pathology, the research team performs a stratified analysis based on smoking history. The stratum-specific results are as follows:
- Among participants with a history of smoking, the relative risk is $1.1$ ( $95\%$ CI: $0.7$ - $1.6$ ).
- Among participants with no history of smoking, the relative risk is $1.2$ ( $95\%$ CI: $0.8$ - $1.8$ ).

Based on these results, which of the following is the most accurate conclusion regarding the role of smoking in this study?

Smoking status is a confounding variable that accounts for the strong initial association between solvent exposure and COPD.Answer
B
Smoking status is an effect modifier that significantly changes the magnitude of the disease risk.
C
The crude relative risk estimate is inaccurate because positive predictive values fluctuate with the baseline prevalence of COPD.
D
The stratification into multiple subgroups requires the use of an analysis of variance (ANOVA) rather than relative risk.
E
The discrepancy between the crude and stratified analyses indicates the presence of diagnostic anchoring bias.

Answer

Smoking status is a confounding variable that accounts for the strong initial association between solvent exposure and COPD.

The correct answer accurately identifies the role of smoking as a confounding variable. Confounding occurs when an observed association between an exposure and an outcome is actually driven by a third variable that is associated with both. In statistical analysis, confounding is identified when the crude (unadjusted) measure of association is significant, but after stratifying the data by the suspected confounder, the stratum-specific measures are similar to each other and substantially closer to the null value (

1.0

). In this scenario, stratifying by smoking status reduces the relative risks to

1.1

and

1.2

, indicating that the initial strong association (

4.2

) was merely an artifact of the unequal distribution of smokers between the exposed and unexposed groups.

Step-by-Step Solution

Evaluate the crude (unadjusted) measure of association.

The unadjusted relative risk is

4.2

, indicating a strong apparent association between the industrial solvent and COPD.

Establishing the baseline finding before adjustment is necessary to understand the potential impact of a third variable.

Analyze the stratum-specific measures of association.

When stratified by smoking status, the relative risks drop to

1.1

and

1.2

, both of which are close to the null value of

1.0

Comparing stratified results to the crude result reveals whether the third variable is altering the true relationship between exposure and outcome.

Compare the crude and stratified results to distinguish between confounding and effect modification.

Because the stratum-specific relative risks are similar to each other but significantly different from the crude relative risk, the third variable (smoking) is identified as a confounder.

In confounding, stratification exposes and eliminates the spurious association. In effect modification, the stratum-specific risks would differ significantly from one another.

Key Concept

Distinguishing confounding from effect modification using stratified data analysis.

Question

Answer

Step-by-Step Solution

Key Concept

Common Mistakes

Topics

More questions from this topic