How to Calculate Limit of Detection
brightly coloured science test tubes image by Steve Johnson from <a href='http://www.fotolia.com'>Fotolia.com</a>
Limit of Detection (LoD) refers to the lowest analyte (chemical substance) concentration that can be distinguished from the Limit of Blank's (LoB) highest analyte concentration found when blank samples containing no analyte are tested. LoD is quantified by measuring the LoB and samples of known analyte concentration.
The calculation for determining the LoD is LoD equals LoB plus 1.645 times the standard deviation of the low concentration sample. Therefore, the first step in the calculation is determining the LoB.
Perform laboratory analyses to obtain measurements on blank samples. Your total sample number should be at least 20 replicates (identical samples). A sample size of 20 will provide enough data points to reduce the sample error in your results.
Calculate the average of the blank samples. This can be performed using the statistical software of your choice.
Calculate the sample standard deviation of the blank samples.
Determine the Limit of Blank. Limit of Blank equals the average concentration of the blanks plus 1.645 times the standard deviation of the blanks.
Perform laboratory measurements on at least 20 replicates of a blank sample.
Determine the mean value and standard deviation of the replicates.
Obtain the standard deviation of samples containing low concentration of the analyte being tested. What constitutes a "low concentration" is specified in the design of the lab test.
Calculate the Limit of Detection. Limit of Detection equals the Limit of Blank plus 1.645 times the standard deviation sample containing low concentration.
Confirm the Limit of Detection by observing the result values for samples containing the limit of detection concentration. Make sure that no more than 5 per cent of the values are less than the Limit of Blank. If this is the case, the Limit of Detection can be confirmed.
- brightly coloured science test tubes image by Steve Johnson from Fotolia.com