A public health team conducts a large prospective cohort study to investigate the associat… — USMLE Step 2 CK

A public health team conducts a large prospective cohort study to investigate the association between a novel airborne industrial chemical exposure and the development of adult-onset asthma. The unadjusted (crude) analysis reveals a relative risk (RR) of $3.2$ (95% CI: $2.4 - 4.1$ , $p < 0.001$ ) for asthma in exposed workers compared to unexposed workers.

Concerned about potential extraneous variables, the researchers stratify the data by a history of tobacco smoking. The stratified analysis yields the following results:

Smoking Status	Relative Risk (RR)	95% Confidence Interval	p-value
Smokers	1.3	0.9 - 1.8	0.18
Non-smokers	1.2	0.8 - 1.7	0.22

The lead investigator reviews the data, noting that the 95% confidence intervals in the stratified groups cross $1.0$ and overlap with each other. Based on these statistical findings, the investigator concludes that smoking is an effect modifier and that the airborne chemical exposure does not have a true independent association with asthma.

Which of the following is the most accurate statistical critique of the study's findings and the investigator's conclusion?

The conclusion is incorrect; the data demonstrates that smoking is a confounding variable, as the stratum-specific relative risks are similar to each other but differ significantly from the crude relative risk.Answer
B
The conclusion is correct; the loss of statistical significance in both strata confirms that smoking modifies the biological effect of the chemical exposure.
C
The conclusion is incorrect; the loss of statistical significance in the strata is primarily due to a Type II error resulting from diminished statistical power after stratification, rather than confounding.
D
The conclusion is incorrect; because subjects were tracked based on their exposure status, this represents a case-control design where calculating relative risk is statistically invalid.
E
The conclusion is correct; the stratification process appropriately adjusted for the changing prevalence of asthma, which normalizes the positive predictive value of the exposure.

Answer

The data demonstrates that smoking is a confounding variable, as the stratum-specific relative risks are similar to each other but differ significantly from the crude relative risk.

In epidemiological and statistical analysis, stratification is a key method for distinguishing between confounding and effect modification. The crude analysis showed a strong relative risk of 3.2. When the data was stratified by smoking status, the relative risk in both groups plummeted to approximately 1.2 and 1.3. Because the stratum-specific estimates are similar to each other but significantly different from the crude estimate, smoking is acting as a confounding variable. The investigator's conclusion that smoking is an effect modifier is statistically incorrect; effect modification requires the effect size to differ meaningfully between the strata.

Step-by-Step Solution

Analyze the crude statistical findings.

The unadjusted relative risk is 3.2 (95% CI: 2.4 - 4.1), indicating a strong, statistically significant association between chemical exposure and asthma.

Establishing the baseline association is necessary before evaluating the impact of stratification.

Evaluate the stratum-specific point estimates.

In smokers, the RR is 1.3. In non-smokers, the RR is 1.2.

Comparing the magnitude of the effect across different strata reveals how the third variable influences the relationship.

Differentiate between confounding and effect modification based on the point estimates.

The stratum-specific RRs (1.2 and 1.3) are similar to each other but notably different from the crude RR (3.2). This statistical pattern defines confounding.

If the strata were significantly different from each other (e.g., 1.2 vs 5.5), it would indicate effect modification. Since they are nearly identical but lower than the crude, smoking is a confounder artificially inflating the initial unadjusted result.

Key Concept

Confounding vs. Effect Modification via Stratification

Estimated Time:2m 0s

Question

Answer

Step-by-Step Solution

Key Concept

Common Mistakes

Topics

More questions from this topic