A new indirect estimation of reference intervals: truncated minimum chi-square (TMC) approach

Area of Science:

• Clinical Chemistry
• Biostatistics
• Laboratory Medicine

Background:

• Estimating reference intervals (RIs) is crucial for interpreting laboratory test results.
• Existing methods struggle with highly skewed data and values below the detection limit.

Purpose of the Study:

• To introduce a new indirect model for estimating reference intervals (RIs).
• To address limitations of current RI estimation methods with skewed data and low values.

Main Methods:

• Utilizes a quantile-quantile plot for initial power normal distribution parameter estimation (λ, μ, σ).
• Employs iterative truncated minimum chi-square (TMC) estimation for parameter refinement.
• Calculates 95% RIs, confidence intervals, and tolerance limits (bootstrapping).
• Incorporates age-dependency using spline functions when applicable.

Main Results:

• The proposed model handles extremely skewed data with high percentages of values at or below the detection limit.
• Demonstrates superior fit to simulated data compared to other indirect methods.
• Provides a robust method for calculating reference interval limits and their confidence/tolerance intervals.

Conclusions:

• The new indirect RI estimation model offers a significant advancement for skewed laboratory data.
• Applicable in R and Excel, facilitating wider adoption in clinical laboratories.
• Improves the reliability of reference interval determination in challenging data scenarios.