A clinical research group analyzes an electronic health record database to evaluate a pote… — USMLE Step 3

A clinical research group analyzes an electronic health record database to evaluate a potential link between a newly approved GLP-1 receptor agonist and the incidence of acute pancreatitis in adults with type 2 diabetes mellitus. The initial unadjusted analysis yields a relative risk (RR) of $2.1$ ( $95\%$ Confidence Interval [CI]: $1.5 - 2.8$ ), suggesting a statistically significant increase in risk. Suspecting an underlying methodological phenomenon, the researchers re-analyze the data by stratifying the patient population according to a history of severe hypertriglyceridemia. In the stratum of patients with a history of severe hypertriglyceridemia, the RR is $1.1$ ( $95\%$ CI: $0.7 - 1.6$ ). In the stratum of patients with normal triglyceride levels, the RR is $1.0$ ( $95\%$ CI: $0.8 - 1.3$ ). Which of the following concepts best explains why the initial unadjusted relative risk was significantly elevated?

The unadjusted analysis failed to account for a variable associated with both the medication use and the outcome.Answer
B
The medication exerts a different magnitude of biological effect depending on the patient's baseline lipid status.
C
The evaluating clinicians had preconceived expectations about the medication, leading to diagnostic distortion.
D
The probability of a true positive diagnosis was artificially altered by the disease prevalence in the sample.
E
The researchers utilized an inappropriate inferential statistical test to compare the categorical incidence rates.

Answer

The unadjusted analysis failed to account for a variable associated with both the medication use and the outcome.

The correct answer accurately describes the mechanism of a confounding variable. Confounding occurs when a third variable is associated with both the exposure (GLP-1 agonist use) and the outcome (acute pancreatitis), but is not on the causal pathway. In this scenario, patients with severe hypertriglyceridemia may be more likely to be prescribed a GLP-1 agonist and are independently at a higher risk for acute pancreatitis. In the crude analysis, this makes it appear as though the medication causes pancreatitis. However, when the researchers control for this third variable via stratification, the relative risk drops to approximately 1.0 in both strata. The disappearance of the association upon stratification is the classic hallmark of confounding.

Step-by-Step Solution

Analyze the unadjusted study results.

The initial crude relative risk is 2.1, indicating a statistically significant association between the GLP-1 receptor agonist and acute pancreatitis.

Establishing the baseline finding is necessary to understand how subsequent analysis changes the interpretation.

Evaluate the results of the stratified analysis.

When stratified by a history of severe hypertriglyceridemia, the relative risk in both the high-triglyceride group (

RR = 1.1

) and normal-triglyceride group (

RR = 1.0

) becomes non-significant and close to the null value of 1.0.

Stratification is the primary method used to distinguish between confounding and effect modification.

Determine the epidemiological concept driving the change.

Because the significant association completely disappears when the data is divided into strata based on a third variable, that third variable (hypertriglyceridemia) is a confounder.

A confounder creates a spurious association in crude data because it is independently linked to both the exposure and the outcome.

Key Concept

Confounding is present when a third variable is associated with both the exposure and the outcome, creating a false association in unadjusted data that disappears upon stratification.

Estimated Time:1m 15s

Question

Answer

Step-by-Step Solution

Key Concept

Common Mistakes

Topics

More questions from this topic