Investigators conduct an observational cohort study to evaluate the association between th… — USMLE Step 2 CK

Investigators conduct an observational cohort study to evaluate the association between the use of a novel oral contraceptive pill (OCP) and the incidence of deep vein thrombosis (DVT). The initial crude analysis demonstrates a relative risk (RR) of $3.2$ ( $95\%$ CI: $2.1$ - $4.8$ ) for DVT in OCP users compared to non-users. The researchers then stratify the data by smoking status. In the stratified analysis, the RR for DVT is $1.1$ ( $95\%$ CI: $0.8$ - $1.5$ ) among smokers and $1.2$ ( $95\%$ CI: $0.9$ - $1.6$ ) among non-smokers.

Which of the following statements are correct regarding the role of smoking status and the appropriate epidemiological methods to address it? (Select 2 answers)

Smoking acts as a confounding variable because the stratum-specific risk estimates are similar to each other but differ significantly from the crude estimate.Answer
Multivariable regression analysis could be utilized during the data evaluation phase to appropriately adjust for the effects of smoking.Answer
C
Smoking demonstrates effect modification because the unadjusted risk estimate is greater than both of the stratum-specific risk estimates.
D
Matching participants by smoking status is a statistical adjustment technique applied after data collection to remove this effect.
E
The crude relative risk of $3.2$ represents the true biological association, while the stratified results reflect selection bias.

Answer

The correct statements are that smoking acts as a confounding variable (because stratum-specific estimates are similar to each other but different from the crude) and that multivariable regression is an appropriate analysis-phase method to adjust for it.

The data demonstrates confounding because stratifying by smoking status eliminates the apparent strong association seen in the crude analysis, yielding stratum-specific relative risks (

1.1

and

1.2

) that are similar to each other but different from the crude relative risk (

3.2

). Furthermore, multivariable regression is a standard analytical method used after data collection to adjust for identified confounders like smoking.

Step-by-Step Solution

Compare the unadjusted (crude) relative risk to the stratum-specific relative risks.

The crude RR is

3.2

. The stratified RRs are

1.1

(smokers) and

1.2

(non-smokers).

Comparing crude and stratified estimates is the primary method to distinguish between confounding and effect modification.

Determine the epidemiological role of the stratifying variable (smoking).

Because the stratified RRs (

1.1

and

1.2

) are similar to each other but differ significantly from the crude RR (

3.2

), smoking is a confounding variable.

Effect modification would result in stratum-specific RRs that differ significantly from each other.

Identify the appropriate stage and method to control for this variable.

Because confounding was identified, the researchers can use analysis-phase techniques like stratification (already done) or multivariable regression to adjust for it.

Methods like matching or restriction must be implemented during the design phase, before data collection.

Key Concept

Distinguishing confounding from effect modification and identifying appropriate methods to control for confounding at different study phases.

Question

Answer

Step-by-Step Solution

Key Concept

Common Mistakes

Topics

More questions from this topic