Conditions for Assessing Precision and Setting QC Statuses

Precision assessment occurs whenever a sample scheme analyte receives a result, or a result is cleared, and sets the sample scheme analyte's Precision Status.

Possible Status Outcomes of a Precision Assessment

Precision assessment resulting in a sample scheme analyte's Precision Status being set to something other than Not Required occurs where:

In all other cases, precision assessments sets the sample scheme analyte's Precision Status to Not Required.

Note: Where a sample that has a Primary Analytical Type of Blank, Standard or Spike AND also a Secondary Analytical Type of Duplicate or Replicate, then the sample undergoes both accuracy assessment and precision assessment. If the assessment outcomes are to be written to QC history, two records are written—one for the accuracy assessment, and one for the precision assessment.

A sample scheme analyte's Precision Status is set to Not Tested upon creation or when the result is cleared.

Abbreviations: SVA stands for scheme version analyte, SpecSVA stands for specification scheme version analyte, SSA stands for sample scheme analyte

Primary Analytical Type
and
Secondary Analytical Type

Result present?

Duplicate Original Sample result present?

Data Type of result/s

SSA QC Active?

SVA Duplicate Stat Det Limit defined?

SVA Duplicate Limiting Repeatability defined?

SVA Replicate Stat Det Limit defined?

SVA Replicate Limiting Repeatability defined?

SVA Repeatability Range Table defined?

SVA Reproducibility Range Table defined?

SVA Precision Script defined?

Allowable %Difference and Actual %Difference Calculations

Precision Assessment

SSA Precision Status

Duplicate-N/A
Replicate-N/A
Blank-Duplicate
Blank-Replicate
Standard-Duplicate
Standard-Replicate
Spike-Duplicate
Spike-Replicate

No

 

 

 

 

 

 

 

 

 

 

 

 

Not Tested

Duplicate-N/A
Replicate-N/A
Blank-Duplicate
Blank-Replicate
Standard-Duplicate
Standard-Replicate
Spike-Duplicate
Spike-Replicate

Yes

No

 

 

 

 

 

 

 

 

 

 

 

Not Tested

Unknown-N/A
Unknown-Blank
Unknown-Standard
Blank-N/A
Blank-Duplicate
Blank-Replicate
Standard-N/A
Standard-Duplicate
Standard-Replicate
Spike-N/A
Spike-Duplicate
Spike-Replicate

 

 

 

 

 

 

 

 

 

 

 

 

 

Not Required

Duplicate-N/A
Replicate-N/A
Blank-Duplicate
Blank-Replicate
Standard-Duplicate
Standard-Replicate
Spike-Duplicate
Spike-Replicate

Yes

Yes

Boolean, Date, DateTime, Image, Document

 

 

 

 

 

 

 

 

 

 

Not Required

Duplicate-N/A
Replicate-N/A
Blank-Duplicate
Blank-Replicate
Standard-Duplicate
Standard-Replicate
Spike-Duplicate
Spike-Replicate

Yes

Yes

Both Numeric

No

 

 

 

 

 

 

 

 

 

Not Required

Duplicate-N/A
Blank-Duplicate
Standard-Duplicate
Spike-Duplicate

Yes

Yes

Both Numeric

Yes

No

No

 

 

 

No

No

 

 

Not Required

Duplicate-N/A
Blank-Duplicate
Standard-Duplicate
Spike-Duplicate

Yes

Yes

Both Numeric

Yes

No

Yes

 

 

 

 

 

Allowable %Difference = DupLR

Actual %Difference = 100 * Absolute(SSA Numeric Final Value ) / Average (Numeric Final Value, orig SSA's Numeric Final Value )

Actual %Difference > Allowable %Difference

Failure

Actual %Difference > 1.96/3.09 * Allowable %Difference and Actual %Difference <= Allowable %Difference

 Warning

Actual %Difference <= 1.96/3.09 * Allowable %Difference

 Pass

Duplicate-N/A
Blank-Duplicate
Standard-Duplicate
Spike-Duplicate

Yes

Yes

Both Numeric

Yes

Yes

No

 

 

 

 

 

Allowable %Difference = 100 * DupSDL / Average (Numeric Final Value, orig SSA's Numeric Final Value )

Actual %Difference = 100 * Absolute(SSA Numeric Final Value - orig SSA Numeric Final Value ) / Average (Numeric Final Value, orig SSA's Numeric Final Value )

Actual %Difference > Allowable %Difference

Failure

Actual %Difference > 1.96/3.09 * Allowable %Difference and Actual %Difference <= Allowable %Difference

Warning

Actual %Difference <= 1.96/3.09 * Allowable %Difference

 Pass

Duplicate-N/A
Blank-Duplicate
Standard-Duplicate
Spike-Duplicate

Yes

Yes

Both Numeric

Yes

Yes

Yes

 

 

 

 

 

Allowable %Difference = 100 * DupSDL / Average (Numeric Final Value, orig SSA's Numeric Final Value ) + DupLR

Actual %Difference = 100 * Absolute(SSA Numeric Final Value - orig SSA Numeric Final Value ) / Average (Numeric Final Value, orig SSA's Numeric Final Value )

Actual %Difference > Allowable %Difference

Failure

Actual %Difference > 1.96/3.09 * Allowable %Difference and Actual %Difference <= Allowable %Difference

 Warning

Actual %Difference <= 1.96/3.09 * Allowable %Difference

 Pass

Replicate-N/A
Blank-Replicate
Standard-Replicate
Spike-Replicate

Yes

Yes

Both Numeric

Yes

 

 

No

No

 

 

 

 

 

Not Required

Replicate-N/A
Blank-Replicate
Standard-Replicate
Spike-Replicate

Yes

Yes

Both Numeric

Yes

 

 

No

Yes

 

 

 

Allowable %Difference = RepLR

Actual %Difference = 100 * Absolute(SSA Numeric Final Value - orig SSA Numeric Final Value ) / Average (Numeric Final Value, orig SSA's Numeric Final Value )

Actual %Difference > Allowable %Difference

Failure

Actual %Difference > 1.96/3.09 * Allowable %Difference and Actual %Difference <= Allowable %Difference

Warning

Actual %Difference <= 1.96/3.09 * Allowable %Difference

 Pass

Replicate-N/A
Blank-Replicate
Standard-Replicate
Spike-Replicate

Yes

Yes

Both Numeric

Yes

 

 

Yes

No

 

 

 

Allowable %Difference = 100 *RepSDL / Average (Numeric Final Value, orig SSA's Numeric Final Value )

Actual %Difference = 100 * Absolute(SSA Numeric Final Value - orig SSA Numeric Final Value ) / Average (Numeric Final Value, orig SSA's Numeric Final Value )

Actual %Difference > Allowable %Difference

Failure

Actual %Difference > 1.96/3.09 * Allowable %Difference and Actual %Difference <= Allowable %Difference

 Warning

Actual %Difference <= 1.96/3.09 * Allowable %Difference

 Pass

Replicate-N/A
Blank-Replicate
Standard-Replicate
Spike-Replicate

Yes

Yes

Both Numeric

Yes

 

 

Yes

Yes

 

 

 

Allowable %Difference = 100 * RepSDL / Average (Numeric Final Value, orig SSA's Numeric Final Value ) + RepLR

Actual %Difference = 100 * Absolute(SSA Numeric Final Value - orig SSA Numeric Final Value ) / Average (Numeric Final Value, orig SSA's Numeric Final Value )

Actual %Difference > Allowable %Difference

Failure

Actual %Difference > 1.96/3.09 * Allowable %Difference and Actual %Difference <= Allowable %Difference

 Warning

Actual %Difference <= 1.96/3.09 * Allowable %Difference

 Pass
Calculation of Percentage Difference from the Pair's Result (PD) following a Precision Assessment

Abbreviations: SpecSVA stands for specification scheme version analyte, SSA stands for sample scheme analyte

If the ssa's Precision Status is something other than Not Tested or Not Required, then calculate SSA's PD.

If the sample's Primary Analytical Type or Secondary Analytical Type is Duplicate, then the SSA's PD is calculated as follows:

Xbar = (original SSA's Final Numeric Value + duplicate SSA's Final Numeric Value) / 2

If Xbar = 0, then:

SSA's PD = 0; exit

If the USE_REP_AVERAGE_FOR_PERCENT_DIFFERENCE application preference is:

  • Not defined, or is defined and selected, then the percentage difference is calculated using the between the duplicate's result and the average of the duplicate's result and the original's result:

SSA's PD = Abs (duplicate SSA's Final Numeric Value—Xbar) * 100 / Xbar

  • Defined and is cleared, then the percentage difference is calculated using the difference difference between the duplicate's result and the original's result:

SSA's PD = Abs (duplicate SSA's Final Numeric Value—original SSA's Final Numeric Value) * 100 / Xbar

If the sample's Primary Analytical Type or Secondary Analytical Type is Replicate, then the SSA's PD is calculated as follows:

Xbar = (original SSA's Final Numeric Value + duplicate SSA's Final Numeric Value) / 2

If Xbar = 0, then:

SSA's PD = 0; exit

If the USE_REP_AVERAGE_FOR_PERCENT_DIFFERENCE application preference is:

  • Not defined, or is defined and selected, then the percentage difference is calculated using the between the replicate's result and the average of the replicate's result and the original's result:

SSA's PD = Abs (replicate SSA's Final Numeric Value—Xbar) * 100 / Xbar

  • Defined and is cleared, when calculating the SSA's PD, then the percentage difference is calculated using the difference between the replicate's result and the original's result:

SSA's PD = Abs (replicate SSA's Final Numeric Value—original SSA's Final Numeric Value) * 100 / Xbar

Note: When the sample's Secondary Analytical Type is Duplicate or Replicate (as in the case of standard pairs, blank pairs or spike pairs), then the SSA's PD is calculated as the percentage difference from the standard's Expected Value when Assessing Results against Accuracy Limits.

Calculation of Relative Percentage Difference from the Pair's Result (RPD) following a Precision Assessment

Abbreviations: SpecSVA stands for specification scheme version analyte, SSA stands for sample scheme analyte

If the SSA's Precision Status is something other than Not Tested or Not Required, then calculate the SSA's RPD (Relative Percentage Difference).

If the sample's Primary Analytical Type or Secondary Analytical Type is Duplicate, then the SSA's RPD is calculated as follows:

Allowable % Difference = 100 * SVA's Dup Statistical Detection Limit / Max (Abs (Average (duplicate SSA's Final Numeric Value, original SSA's Final Numeric Value)), 0.000000001) + SVA's Dup Limiting Repeatability

If the RPD_CALC_DL_TYPE application preference is:

  • Defined and is set to INTERNAL DL, then:

V1 = Max (duplicate SSA's Final Numeric Value, SVA's Internal Lower DL) and V2 = Max (original SSA's Final Numeric Value, SVA's Internal Lower DL)

  • Defined and is set to REPORT DL, then:

V1 = Max (duplicate SSA's Final Numeric Value, SVA's Report Lower DL) and V2 = Max (original SSA's Final Numeric Value, SVA's Report Lower DL)

  • Defined and is set to STAT DL and the sample's Primary Analytical Type or Secondary Analytical Type is Duplicate, then:

V1 = Max (duplicate SSA's Final Numeric Value, SVA's Dup Statistical Detection Limit) and V2 = Max (original SSA's Final Numeric Value, SVA's Dup Statistical Detection Limit)

  • Not defined, then:

V1 = duplicate SSA's Final Numeric Value and V2 = original SSA's Final Numeric Value

If the sample's Primary Analytical Type or Secondary Analytical Type is Replicate, then the SSA's RPD is calculated as follows:

Allowable % Difference = 100 * SVA's Rep Statistical Detection Limit / Max (Abs (Average (replicate SSA's Final Numeric Value, original SSA's Final Numeric Value)), 0.000000001) + SVA's Rep Limiting Repeatability

If the RPD_CALC_DL_TYPE application preference is:

  • Defined and is set to INTERNAL DL, then:

V1 = Max (replicate SSA's Final Numeric Value, SVA's Internal Lower DL) and V2 = Max (original SSA's Final Numeric Value, SVA's Internal Lower DL)

  • Defined and is set to REPORT DL, then:

V1 = Max (replicate SSA's Final Numeric Value, SVA's Report Lower DL) and V2 = Max (original SSA's Final Numeric Value, SVA's Report Lower DL)

  • Defined and is set to STAT DL and the sample's Primary Analytical Type or Secondary Analytical Type is Replicate, then:

V1 = Max (replicate SSA's Final Numeric Value, SVA's Rep Statistical Detection Limit) and V2 = Max (original SSA's Final Numeric Value, SVA's Rep Statistical Detection Limit)

  • Not defined, then:

V1 = replicate SSA's Final Numeric Value and V2 = original SSA's Final Numeric Value

Then:

SSA's RPD = Abs (V1—V2) / Average (V1, V2) * 100

Note: If the average of V1 and V2 is 0 (zero), then the SSA's RPD is set to 0 (zero).

Rounding of PD and RPD

After the PD and RPD are calculated, then they are rounded to 5 decimal places. This also applies when forcing reassessment of results either manually or via a workbook onLoad script.

If PD and/or RPD are changed in a workbook session, then only the rounded values are compared. Values with zeroes after the decimal places are managed, for example, 0.1100 and 0.11 are treated as the same value, at 5 decimal places.

Since historical calculations of these properties on sample scheme analytes are unrounded to 5 decimal places, if the results are reassessed in workbook, then those cells containing a rounded PD and RPD value are flagged as dirty and saved upon submit and the related QC History observations are updated, and created where they do not yet exist.

Composite QC Status

The sample scheme analyte's Precision Status, along with the Limit Status, Standard Status and Specification Status, is rolled into sample scheme analyte's Composite QC Status. When the Composite QC Status is Warning or Failure, the operator is notified by the workbook cell being highlighted yellow or red, respectively. A sample scheme analyte that has a Composite QC Status of Failure cannot be validated to have the sample scheme analyte's Workflow Status set to Completed.

If a sample scheme analyte's Composite QC Status of Failure is subsequently ignored by the operator, the Composite QC Status is changed to Ignore Failure. Where the result of the precision assessment is saved to QC history, the QC Status of the QC history accuracy record is changed from Failure to Ignore Failure.