All you wanted to know about laboratory tests

All you wanted to know about laboratory tests, but were afraid to ask The informed patient meeting the laboratory director Metrology Accuracy and more 2 By Dietmar Stöckl STT Consulting -5 All you wanted to know about laboratory tests -5 -

Accuracy and more Today What would Westgard say? Sounds good! All you wanted to know about laboratory tests -5 - 2

Repetition Trueness, NOTE 2 Measurement trueness is inversely related to systematic measurement error. Bias [of measurement] Estimate of a systematic measurement error. Random error, NOTE 2 Random measurement errors of a set of replicate measurements form a distribution that can be summarized by its expectation, which is generally assumed to be zero, and its variance. All you wanted to know about laboratory tests -5 - 3

What would Westgard say? JIM! How does that sound to you? Similar, but not the same! All you wanted to know about laboratory tests -5 - 4

How do they compare? VIM Trueness, NOTE 2 … trueness is inversely related to systematic … error. Bias [of measurement] Estimate of a systematic … error. Random error, NOTE 2 Random … errors of a set of replicate measurements form a distribution that can be summarized by its … variance. All you wanted to know about laboratory tests -5 - 5

How do they compare? Westgard Bias Random error "Distribution-based", with specified probability (95%: two-, or one-sided) z x s (z = 1. 96 or 1. 65) Total error "Distribution-based" Bias + z x s VIM Bias [of measurement] Random error, NOTE 2 Note 2: "Distribution-based", probability not specified s 2 (variance) AND "result-based" (definition) Error "Result-based" Measured value minus a reference value. All you wanted to know about laboratory tests -5 - 6

How do they compare? Westgard Accuracy Like it or not, manufacturers and laboratories are left with the “systematic error concept of accuracy” VIM Accuracy Closeness of agreement between a measured quantity value and a true quantity value of the measurand. (westgard. com/essay 118. htm) Still differences! All you wanted to know about laboratory tests -5 - 7

Another voice Typical total error concepts are too simplistic. At least, sample related effects should be included. Krouwer JS. Setting performance goals and evaluating total analytical error for diagnostic assays. Clin Chem 2002; 48: 919 -27. All you wanted to know about laboratory tests -5 - 8

Errors – Summary Random error -Analytical: SDa (CVa) -Sample-related: SDs (CVs) “Total”: SDt = [SDa 2 + SDs 2] (may be expanded) Systematic error/Bias Bt = Experimental mean (sufficient n) minus “reference”/”true” Overall plus actual lot (may be expanded) Total error TE = Bt + z SDt Interferences Lipaemia, heterophilic antibodies, etc. >Estimate frequency and magnitude of effect >Treat separately by risk management procedures (my opinion) All you wanted to know about laboratory tests -5 - 9

In the end I type: . . . altogether, there still is hope. . . In my lifetime? All you wanted to know about laboratory tests -5 - 10

The next visit Uncertainty I love that one! All you wanted to know about laboratory tests -5 - 11