calculate_lod: Limit of Detection (LoD) calculations

In njjms/nicks: Basic Functions for Biostatistics

Limit of Detection (LoD) calculations

Description

These calculations are based on CLSI EP17-A2 (section 5.3.3.2 covers both parametric and nonparametric calculations)

Usage

calculate_lod(x, LoB = NA, alpha = 0.05)

Arguments

x	data.frame containing two columns: concentration and measurement
alpha	defaults to .05

Details

According to EP17, the minimal experimental design is: - 2 reagent lots - 1 instrument system - 3 days - 4 low measurand content samples - 2 replicates per sample

Note that multiplying all of this together gives us 48 samples. According to EP17, the minimum number of replicates for low measurand samples is 60, so at least one of these factors will have to be increased.

For these LoD calculations, it is assumed that variability is relatively consistent across low measurand concentrations. If this is not the case, a trial-and-error experimental design or precision profile approach may be needed.

Value

a list containing parametric LoD, a table of counts, pooled SD estimate, and per concentration SD estimates.

Examples

calculate_lod(x = data, LoB = .15)

njjms/nicks documentation built on May 4, 2022, 8:10 a.m.