Underestimation of risk due to exposure misclassification

Area of Science:

• Environmental Health
• Toxicology
• Epidemiology

Background:

• Exposure misclassification is a significant challenge in establishing dose-response relationships for risk assessment.
• Non-differential errors in exposure assessment can lead to underestimation of health risks.
• Sensitivity analysis can adjust for known degrees of misclassification.

Purpose of the Study:

• To quantify the full magnitude of measurement error in prenatal methylmercury exposure assessment.
• To compare laboratory imprecision with total measurement error in exposure biomarkers.
• To evaluate the impact of underestimating measurement error on risk assessment and precautionary decisions.

Main Methods:

• Utilized data from a prospective Faroese birth cohort study.
• Employed two biomarkers: mercury concentration in cord blood and maternal hair.
• Applied factor analysis and structural equation modeling to assess total imprecision.
• Included dietary questionnaire data on whale meat consumption as an exposure parameter.

Main Results:

• Total imprecision for cord blood mercury concentration (28-30% CV) and maternal hair mercury concentration (52-55% CV) greatly exceeded laboratory imprecision (<5% CV).
• Dietary questionnaire data exhibited even higher imprecision.
• Measurement error is substantially underestimated when relying solely on laboratory reproducibility or quality control data.

Conclusions:

• Laboratory data significantly underestimate the true measurement error in prenatal methylmercury exposure assessment.
• Accurate risk assessment and precautionary principle application necessitate realistic estimation of exposure measurement errors.
• Overlooking or underestimating exposure measurement errors can lead to inadequate risk management and unintended levels of precaution.