Exposure assessment

Exposure Assessment - The "Who Got What?" Game

• Objective: To accurately classify individuals as exposed or unexposed to a specific risk factor. This is the critical first step in a cohort study.
• Methods of Assessment:
  • Questionnaires & Interviews: Directly ask participants. Prone to recall bias.
  • Records: Use pre-existing data like medical or employment records. More objective.
  • Biomarkers: Measure substances in biological samples (e.g., blood, urine). Most objective and precise.
  • Environmental Monitoring: Assess exposure levels in a specific location (e.g., air quality).

⭐ Exam Favorite: Inaccurate exposure assessment leads to misclassification bias. If this occurs randomly (non-differential), it biases the results toward the null hypothesis, underestimating the true association.

Measurement Methods - Tools of the Trade

Error generating content for this concept group: Invalid JSON response

Misclassification Bias - When Measurements Go Wrong

• Occurs when either exposure or outcome is inaccurately categorized, leading to incorrect subject assignment.
• Two main types:
  • Non-differential Misclassification:
    • Measurement error is the same across all study groups (e.g., exposed/unexposed).
    • Effect: Biases results towards the null (underestimates the true association).
  • Differential Misclassification:
    • Measurement error rates differ between study groups.
    • Example: Recall bias, where cases may remember past exposures more accurately than controls.
    • Effect: Can bias results either towards or away from the null.

⭐ In cohort studies, non-differential misclassification of the outcome is more common than of the exposure. It makes the groups appear more similar than they truly are, weakening the observed association.

High‑Yield Points - ⚡ Biggest Takeaways

• Exposure assessment aims to correctly classify subjects as exposed or unexposed, a process critical for a study's internal validity.
• Non-differential (random) misclassification occurs at equal rates in all study groups, biasing the measure of association towards the null.
• Differential (non-random) misclassification occurs at different rates between groups, which can bias the association either towards or away from the null.
• Blinding of investigators to outcome status prevents differential misclassification.
• Objective measures (e.g., biomarkers) are preferred over subjective reports to reduce error.