A vaccination campaign designed to increase the uptake of HPV vaccine was instituted in chosen counties of a certain state in order to educate parents not only about the disease itself, but also about why children should be vaccinated against this viral sexually transmitted disease. At the end of the campaign, children living in counties in which it was conducted were 3 times more likely to receive the HPV vaccine compared with children living in counties where no campaign was instituted. As well, after evaluating only the counties that were part of the vaccination campaign, the researchers found that families with higher incomes were 2 times more likely to vaccinate their children against HPV compared with families with lower incomes. What conclusion can be drawn from these results?

Family income appears to be an effect modifier.

The vaccination campaign appears to have been ineffective.

The vaccination campaign is the study outcome.

The vaccine uptake is the study exposure.

Family income appears to be a confounder.

Study X examined the relationship between coffee consumption and lung cancer. The authors of Study X retrospectively reviewed patients' reported coffee consumption and found that drinking greater than 6 cups of coffee per day was associated with an increased risk of developing lung cancer. However, Study X was criticized by the authors of Study Y. Study Y showed that increased coffee consumption was associated with smoking. What type of bias affected Study X, and what study design is geared to reduce the chance of that bias?

Observer bias; double blind analysis

A group of environmental health scientists recently performed a nationwide cross-sectional study that investigated the risk of head and neck cancers in patients with a history of cigar and pipe smoking. In collaboration with three teams of epidemiologists that have each conducted similar cross-sectional studies in their respective countries, they have agreed to contribute their data to an international pooled analysis of the relationship between non-cigarette tobacco consumption and prevalence of head and neck cancers. Which of the following statements regarding the pooled analysis in comparison to the individual studies is true?

The likelihood of type II errors is decreased.

It overcomes limitations in the quality of individual studies.

It is able to provide evidence of causality.

The level of clinical evidence is lower.

A 73-year-old man is brought to the emergency department by ambulance after being found to be non-communicative by his family during dinner. On presentation he appears to be alert, though he is confused and cannot follow instructions. When he tries to speak, he vocalizes a string of fluent but unintelligible syllables. Given this presentation, his physician decides to administer tissue plasminogen activator to this patient. This intervention best represents which of the following principles?

This does not represent prevention

Confounding variables — USMLE Lesson

Confounding Variables - The Meddling Third Wheel

An extraneous variable independently associated with both exposure and outcome, distorting the true relationship. It can create a spurious association or mask a real one.
Key Characteristics:
- Associated with the exposure.
- A risk factor for the outcome.
- Not on the causal pathway between exposure and outcome.

Causal diagram: Confounder, mediator, and collider

Methods to Control Confounding:
- Design Stage: Randomization, restriction, matching.
- Analysis Stage: Stratification, multivariable analysis.

⭐ In clinical trials, randomization is the most effective method to control for both known and unknown confounders.

Controlling Confounding - Taming the Bias

Methods to prevent or adjust for confounding variables, ensuring the observed association is between the true exposure and outcome.

During Study Design (Proactive)
- Randomization: Best method. Evenly distributes all potential confounders (known & unknown) between groups. Cornerstone of RCTs.
- Restriction: Limits entry to subjects with specific characteristics, thus eliminating the confounder (e.g., only non-smokers). Reduces generalizability.
- Matching: For each case, select controls with identical values for specific confounders (e.g., age, sex). Common in case-control studies.
During Data Analysis (Reactive)
- Stratification: Examines the exposure-outcome association within different strata (levels) of the confounder.
- Multivariable Analysis: Uses statistical models (e.g., logistic regression) to mathematically adjust for the effect of multiple confounders.

⭐ Randomization is the most effective method because it controls for both known and unknown confounders. However, it is not always feasible or ethical.

Confounder vs. Effect Modifier - Telling Them Apart

Confounder: An external variable associated with both exposure and outcome, distorting the true relationship. It creates a spurious association.
Effect Modifier: A variable that genuinely alters the strength or direction of the association between exposure and outcome. The effect is real but varies by stratum.

Confounding vs. Effect Modification with Stratification

To distinguish, perform stratification analysis.

After Stratifying by the Variable...	Confounding	Effect Modification
Measure of Association	Similar across all strata	Different across strata
Crude vs. Stratum-Specific	Crude measure is misleading	Crude measure is a weighted average
How to Handle	Adjust for it (e.g., matching, regression)	Report stratum-specific results

High‑Yield Points - ⚡ Biggest Takeaways

A confounder is a third variable associated with both the exposure and the outcome, but not on the causal pathway.

It can create a spurious association or mask a true one, leading to biased results.

Control for confounders using randomization, matching, restriction, stratification, or multivariate analysis.

Randomization is the best method to control for both known and unknown confounders.

Distinguish from effect modification, where an external variable truly alters the exposure's effect across different strata.

Unlock the full lesson and continue reading

Signup to continue reading this lesson and unlimited access questions, flashcards, AI notes, and more

Scan to download app

UNLOCK FREE ACCESS