23 May 2017
To: Recipients of EP15-A3
From: Jennifer K. Adams, MT(ASCP), MSHA
Subject: Corrections
This notification is to inform you of technical and editorial corrections made to CLSI
document EP15-A3, User Verification of Precision and Estimation of Bias; Approved
Guideline—Third Edition. The corrections are shown as highlighted and/or stricken text in the
items listed below.
Technical correction:
A correction was made in Section 3.5, The Verification Interval.
In equation (11), a superscripted “2” was added after the closing parenthesis. The
equation was revised to read:
Editorial corrections:
Two consistency corrections were made throughout the document as necessary:
– All instances of “sR” and “sWL” were corrected to use a lowercase “s.”
– All instances of “σ” in “σR” and “σWL” were italicized.
Corrections were made in Section 2.3.5, Imprecision Estimates by One-Way Analysis of
Variance.
In the Abbreviations list in Table 4, two instances of “total” were italicized. The
Abbreviations list was revised to read:
Abbreviations: ANOVA, analysis of variance; DF, degrees of freedom; DFtotaltotal, total
degrees of freedom; MS, mean squares; SS, sum of squares; SStotaltotal, total sum of
squares.
In the “Important Note” callout based on the last sentence of the last paragraph, the
subscripted “B” was capitalized. The callout was revised to read, “Estimating and
expressing sR, sbB, and sWL helps the laboratory identify and quantify the sources of
uncertainty.”
A correction was made in the Upper Verification Limit section of Section 2.3.6.2,
Performing the Comparisons. The “R” in “%CVR” was italicized and subscripted. The text
was revised to read, “...or as a %CV (%CV = %CVRR or %CV = %CVWL).”
A correction was made in Scenario E in Section 3.3, Target Values and Their Standard
Errors. “seRM” was italicized. The text was revised to read, “When working with a
commercial QC material supplied with a TV for which the standard error cannot be
estimated, set seRMseRM = 0.”
Corrections were made in Section 3.5, The Verification Interval.
In equation (14), the “s” in the first set of parentheses was italicized, italics were
removed from “0.975,” and the multiplication sign was changed to a multiplication dot.
The equation was revised to read:
VI = TV ± (m • ssec) = TV ± (t0.9750.975,df ו sec)
c
In Table 15B’s title, “NRM” was italicized and “RM” was subscripted. The text was revised
to read, “…With Five Replicates per Run, and NRMNRM = 10, 20, 50, 100 and ≥ 200
Laboratories.”
In Table 15C’s title, “z” was changed to “NRM.” The text was revised to read “…With Five
Replicates per Run, and zNRM = 10, 20, 50, 100 and ≥ 200 Laboratories.”
A correction was made in the second paragraph of Section 3.6, Interpretation. Italics were
removed from “0.975” and the multiplication sign was changed to a multiplication dot.
The text was revised to read, “…compare the defined allowable bias to the expanded
combined uncertainty (half width of the CI for the bias, t0.9750.975,df ו sec).”
c
A correction was made in Step 2 in Section 3.7.2, Worked Example 1B: Ferritin Precision
Example in Section 2.3.10 Using Statistics From Manufacturer’s Claims. The letter “A” was
removed before the equation, and the unit of measure was changed from “g/L” to “μg/L."
The equation was revised to read:
A correction was made in Step 9 in Section 3.7.5, Worked Example 3A: Testing of Sample
Spiked With Digoxin Using Statistics From the Precision Experiment (Extremely High sWL /
sR). A negative sign was added to “0.03.” The text was revised to read, “The observed
bias of −0.03 μg/L is not statistically significant.”
A correction was made in the second to last sentence in Section 3.7.7, Worked Example 4:
Testing of Sample Spiked With Digoxin Using Statistics From the Manufacturer’s Claims.
The “WL” was subscripted. The text was revised to read, “The interpolated %CVWLWL
claim was (6.48% + 1.91%) / 2 = 4.20%.”
A correction was made in Section B5 of Appendix B, Determining the Upper Verification
Limit Factor, F. An end parenthesis was added after “34.17” and italics were removed
from “UVL.” The text was revised to read “(For example, in a study involving two
samples, if df = 20, then X2 = 34.17). Then, calculate the UVLUVL factor, F = X2 / df .”
If you require any additional clarification regarding these corrections, please contact CLSI
Customer Service ([email protected]).
We appreciate your commitment to CLSI, and regret any inconvenience.
