Statistics - Measurement & Trueness

Determined by obtaining several calibrators or controls of known concentration that are near the LoD

It is when any signal that is 10s above the blank could be measured accurately and therefore that value defined a related term.

Limit of Quantitation

Used to determine the accuracy (bias) of the assay.

Certified reference materials with an assay value specifically assigned for your method. 

Proficiency test samples

QC sera that have a sufficiently large peer group (>10 labs).

10 measurements run 5 times with each run on a different day.

Measure 20 samples with values near the LoB

At least 85% of the samples should be above the LoB, if the LoD is accurateIf 40-60 samples are available, the proportion nabove the LoB should be at easet 88%.

If the total error estimate is less than the CLIA TEa then the LoQ is at or near the LoD.

By the SEa (for non immunoassays).

This is equal to the TEa x .75, since random error should be no greater than 25% of the TEa.

Prior to reference material, a recover study validated the trueness.

A sample of known concentration is spiked with a known quantity of the analyte to produce an expected value. Assayed in duplicate. Take the average measured value divided by the expected value, multiplied by 100 to give a % recovery.

Quantifies the proportional error present in a method

Spike several normal patient samples with a tenfold greater concentration of the substance that is normally found (e.g. 10x the upper therapeutic level for a drug)

Perform a t test to determine if the bias (difference in means between spiked and unspiked) is significant.

• The principal interferents are:
• Hemolysis (free hemoglobin)
• Jaundice (bilirubin)
• Lipids (lipemia)
• Metabolites
• Antibodies

The principal interferents are: 

• Drugs and drug metabolites
• &
• Chemical additives used for sample collection

• 1. Exposure of samples to light
• 2. Use of the wrong collection device
• 3. Wrong order of draw
• 4. Excessive sample storage
• 5. Improper temperature or treatment of the sample.

LoD & LoB are different because of the uncertainty in the confidence interval for the blank. This arises because the blank reading cannot be normally distributed around a signal for57/300 zero since most instruments do not measure below zero (negative numbers).

Limit of Detection

Lowest signal that can be differentiated from the blank (zero calibrator) & the LoQ. 

It is the lowest concentration that gives a result within the lab's performance goal for accuracy (i.e. CLIA '88 TEa).

Limit of Blank