Calculation of the confidence bounds for the fraction nonconforming of normal populations of measurements in clinical laboratory medicine

TL;DR

This paper presents a method to calculate confidence bounds for the fraction nonconforming in normal populations of clinical laboratory measurements, accounting for unknown errors using the noncentral t-distribution.

Contribution

It introduces a detailed approach to compute confidence bounds for the fraction nonconforming considering both random and systematic errors in clinical lab data.

Findings

Confidence bounds can be calculated using the noncentral t-distribution.

Method accounts for unknown systematic and random errors.

Illustrated with practical examples.

Abstract

The fraction nonconforming is a key quality measure used in statistical quality control design in clinical laboratory medicine. The confidence bounds of normal populations of measurements for the fraction nonconforming each of the lower and upper quality specification limits when both the random and the systematic error are unknown can be calculated using the noncentral t-distribution, as it is described in detail and illustrated with examples.