Examkin

Question

Difficulty: HardBias, Confounding, and Validity

Evaluate the following clinical research scenarios and match each to the specific epidemiological phenomenon or bias it most accurately represents.

  • In a 5-year longitudinal trial of a demanding intensive lifestyle intervention for type 2 diabetes, 45% of intervention subjects drop out compared to 10% of controls. Per-protocol analysis of the remaining subjects demonstrates a massive reduction in HbA1c.Attrition bias
  • A case-control study evaluates the link between chronic NSAID use and renal cell carcinoma. Researchers select both the cases and the controls exclusively from admitted patients on a general medical ward, spuriously altering the exposure odds.Berkson bias
  • A novel rapid biomarker for suspected pulmonary embolism is evaluated. Only patients testing positive for the biomarker proceed to receive the gold-standard CT pulmonary angiography, while negative patients are discharged without advanced imaging.Verification bias
  • A trial evaluating a biologic for asthma shows an overall relative risk of exacerbation of 0.85. Stratified analysis reveals a relative risk of 0.40 in patients with high baseline eosinophils and 1.05 in those with normal eosinophils.Effect modification

Answer

The correct matches are: 1) The diabetes dropout scenario represents Attrition bias. 2) The hospitalized control scenario represents Berkson bias. 3) The biomarker screening scenario represents Verification bias. 4) The asthma biologic stratum-specific scenario represents Effect modification.
Each scenario perfectly illustrates a classic epidemiological principle: differential loss to follow-up defines attrition bias; hospital-based control selection defines Berkson bias; non-random application of a diagnostic gold standard defines verification bias; and differential treatment effects across physiological subgroups defines effect modification.

Step-by-Step Solution

1
Analyze the diabetes lifestyle intervention trial.
Identified significant differential dropout rates prior to outcome measurement.
Unequal loss to follow-up compromises randomization and causes attrition bias, a subtype of selection bias.
2
Analyze the NSAID and renal cell carcinoma case-control study.
Identified that both cases and controls are selected from a hospitalized population.
Hospital-based selection alters baseline exposure frequencies compared to the general population, which specifically defines Berkson bias (admission rate bias).
3
Analyze the pulmonary embolism diagnostic study.
Identified that the gold standard test is not applied uniformly to all participants.
Applying the reference standard only to those who test positive on the index test leads to verification (workup) bias.
4
Analyze the asthma biologic trial.
Identified that stratification yields two distinct relative risks that differ significantly from each other and from the crude relative risk.
This biological interaction across subgroups confirms effect modification, distinguishing it clearly from confounding.

Key Concept

Differentiating specific subtypes of selection bias (Attrition, Berkson), information bias (Verification), and epidemiological phenomena (Effect modification) in clinical research.
Rate this question

Topics

Bias, Confounding, and Validity

More questions from this topic