Examkin

Question

Difficulty: MediumBias, Confounding, and Validity

A team of clinical epidemiologists is designing an observational cohort study to evaluate the association between a newly approved biologic therapy for severe asthma and the subsequent risk of developing serious respiratory infections. The investigators are concerned about a specific variable: baseline pulmonary function (measured by FEV1). Patients with poorer baseline FEV1 are more likely to be prescribed the new biologic therapy and are also independently at a higher risk of developing serious respiratory infections, regardless of the treatment they receive.

Which of the following methodological techniques, if appropriately implemented during the design or analysis phase of the study, would effectively control for the confounding effect of baseline pulmonary function? (Select all that apply)

  1. Matching each patient receiving the new biologic therapy with a control patient who has a similar baseline FEV1Answer
  2. Restricting the study enrollment exclusively to patients who have a baseline FEV1 between 60% and 80% of predictedAnswer
  3. Adjusting for baseline FEV1 as an independent covariate in a multivariable regression modelAnswer
  4. D
    Masking the outcome assessors to the treatment group allocation of each participant
  5. E
    Reporting stratum-specific relative risks to highlight differing magnitudes of the biologic's effect across baseline FEV1 categories

Answer

Matching each patient receiving the new biologic therapy with a control patient who has a similar baseline FEV1, Restricting the study enrollment exclusively to patients who have a baseline FEV1 between 60% and 80% of predicted, and Adjusting for baseline FEV1 as an independent covariate in a multivariable regression model.
Confounding occurs when an extraneous variable is associated with both the exposure and the outcome, skewing the true association. It can be controlled during the design phase of a study via restriction (limiting the study to one category of the confounder) or matching (pairing subjects so the confounder is equally distributed). In the analysis phase, confounding can be effectively controlled via multivariable regression modeling, which mathematically adjusts the data to isolate the independent effect of the exposure.

Step-by-Step Solution

1
Identify the research problem described in the scenario.
The study is observational, and baseline FEV1 is identified as a variable associated with both the exposure (biologic therapy) and the outcome (respiratory infections). Therefore, FEV1 meets the criteria for a confounder.
Recognizing the nature of the variable dictates which methodological techniques are appropriate.
2
Recall the established epidemiological methods used to control for confounding.
Confounding can be controlled in the design phase (via randomization, matching, or restriction) and in the analysis phase (via stratification with pooled estimates or multivariable regression).
This provides the framework for evaluating the answer choices.
3
Evaluate each option against the established methods for confounding control.
Matching and restriction are valid design phase techniques. Adjusting via multivariable regression is a valid analysis phase technique. Masking addresses bias, not confounding. Highlighting stratum-specific differences addresses effect modification, not confounding.
This directly filters the correct options from the distractors based on methodological principles.

Key Concept

Methods to control for confounding in observational studies (Matching, Restriction, Stratification, Multivariable Regression).
Rate this question

Topics

Bias, Confounding, and Validity

More questions from this topic