Statistical aspects of the estimation of reference limits.

Area of Science:

• Environmental Science
• Biostatistics
• Toxicology

Background:

• Biological monitoring generates complex data requiring robust statistical analysis.
• Accurate interpretation of biological monitoring data is crucial for environmental and health assessments.

Purpose of the Study:

• To outline fundamental statistical methods applicable to biological monitoring.
• To provide practical examples for data description, outlier detection, and interval estimation.
• To guide sample size and power calculations for diagnostic quantitative tests.

Main Methods:

• Descriptive statistics and Gaussianity tests for data characterization.
• Identification and handling of outliers in biological datasets.
• Calculation of parametric and non-parametric estimates for fractiles and their standard errors.
• Parametric estimation of one-tail upper side tolerance intervals.
• Sample size and power determination based on diagnostic test cut-off points.

Main Results:

• Demonstrated application of statistical methods using two real-world biological monitoring datasets.
• Provided practical guidance on selecting appropriate statistical techniques for biological data.
• Illustrated the relationship between sample size, statistical power, and diagnostic test performance.

Conclusions:

• Effective statistical analysis is essential for reliable biological monitoring.
• The methods presented enhance the interpretation of monitoring data and inform test design.
• This work supports evidence-based decision-making in environmental health and toxicology.