Calibration Calibration Verification Analytical Measurement Range and Clinically

Calibration, Calibration Verification, Analytical Measurement Range, and Clinically Reportable Range For the CAP Instrumentation and Chemistry Resource Committees 1

The Truth, Please • What’s your response when • • You receive the results of a CAP Linearity Survey? You come to Checklist Question CHM. 13400 (one of several relating to Calibration Verification)? 2

The Truth, Please • What’s your response when • • You receive the results of a CAP Linearity Survey? You come to Checklist Question CHM. 13400 (one of several relating to Calibration Verification)? • If you’re anything like me, it’s not “Wow, let me at it. I was really looking forward to taking care of this!” • Who I Am: Gary L. Horowitz, MD, FCAP Director, Clinical Chemistry, Beth Israel Deaconess Medical Center Associate Professor of Pathology, Harvard Medical School Chair, CAP Chemistry Resource Committee 3

Terms Calibration 4

Terms that Instill Fear Calibration Verification Analytical Measurement Range (AMR) Clinically Reportable Range (CRR) 5

Terms that Instill Fear Calibration Verification CAP Linearity Surveys Analytical Measurement Range (AMR) Clinically Reportable Range (CRR) 6

Why We’re Here Today • confusing, but important, concepts • audioconference format: allows anyone to visit the web as often as needed • objectives: • explain the concepts • reveal how they relate to the accreditation process • show you in detail how the CAP Linearity Surveys enable you to address these issues for your laboratory • ambitious goals, but we have two excellent speakers, experts • in Laboratory Medicine • in CAP Surveys 7

Introductions • Greg Miller, Ph. D è è Professor of Pathology, Virginia Commonwealth University Consultant Member, CAP Chemistry Resource Committee 8

Introductions • Greg Miller, Ph. D è è Professor of Pathology, Virginia Commonwealth University Consultant Member, CAP Chemistry Resource Committee • Anthony Killeen, MD, Ph. D, FCAP è è è Associate Professor of Pathology, University of Minnesota Member, CAP Instrumentation Resource Committee Former Member and Chair, CAP Chemistry Resource Committee • Program developed as joint effort of these two CAP Resource Committees 9

Objectives • Differentiate between the terms calibration, calibration verification, analytical measurement range (AMR) and clinically reportable range (CRR). • Clarify the accreditation requirements regarding these items. • Explain techniques to perform the appropriate verifications. • Clarify CAP Linearity (LN) Survey reports. 10

CLIA 493. 2 Reportable range: span of test result values over which the lab can establish or verify the accuracy of the measurement response. 11

CAP (two concepts) CLIA 493. 2 Analytical measurement range: range of analyte values that a method can directly measure on the specimen without any dilution, or other pretreatment not part of the usual assay process. Reportable range: span of test result values over which the lab can establish or verify the accuracy of the measurement response. 12

CAP (two concepts) CLIA 493. 2 Clinically reportable range: range of analyte values that are reported as a quantitative result, allowing for specimen dilution or other pretreatment used to extend the actual AMR. Reportable range: span of test result values over which the lab can establish or verify the accuracy of the measurement response. 13

CAP Calibration: set of operations that establish the quantitative relationship between the reagent system / instrument response and the concentration or activity values of an analyte. CLIA Calibration: process of testing and adjusting to establish a correlation between the measurement response and the concentration. 14

CAP (two concepts) CLIA 493. 2 Calibration verification: assaying materials of known concentration to substantiate the system’s calibration throughout the reportable range. 15

CAP (two concepts) CLIA 493. 2 Calibration verification: process of confirming that the current calibration settings remain valid for a method. Calibration verification: assaying materials of known concentration to substantiate the system’s calibration throughout the reportable range. 16

CAP (two concepts) CLIA 493. 2 Analytical measurement range verification: process of confirming that the assay system will correctly recover the concentration or activity of the analyte over the AMR. Calibration verification: assaying materials of known concentration to substantiate the system’s calibration throughout the reportable range. 17

CLIA 493. 1255: Calibration and calibration verification procedures (a) Perform and document calibration procedures. 1. Follow manufacturer’s instructions. 2. Use criteria established by lab during validation of method when put in service. 3. Whenever calibration verification fails to meet the lab’s specifications. 18

CLIA 493. 1255: Calibration and calibration verification procedures (b) Perform and document calibration verification procedures. 1. Follow manufacturer’s instructions 2. Use criteria established by lab during validation of the method when put in service è Include minimum, mid-point and maximum value to verify the laboratory’s reportable range (this is AMR) 3. At least every 6 months or when the following occur: • • Change of reagents unless it does not affect the range to report patient results (this is AMR), or QC values (this is CAL) Major preventive maintenance (this is CAL) QC trend, shift or out of limits (this is CAL) Schedule for verifying reportable range (this is AMR) requires more frequent calibration verification 19

Calibration (two point) SIGNAL Analytical Measurement Range extends beyond the high calibrator = Reportable Results CONCENTRATION CALIBRATOR TARGET VALUES • May be measurement system specific 20

Calibration (two point) Example: typical serum glucose method • Low calibrator = reagent blank (0 mg/d. L) • High calibrator = 356 mg/d. L • AMR = 10 -600 mg/d. L • The AMR is based on the expectation that the signal is linearly related to glucose concentration over the 10 -600 mg/d. L range. • Verification of the AMR expectation is required because the calibrators do not define the complete measurement range. 21

Calibration (multi-point) Analytical Measurement Range SIGNAL = Reportable Results Note that when calibrators span the AMR, the calibration is fully defined over the AMR CONCENTRATION CALIBRATOR TARGET VALUES • May be measurement system specific 22

Accuracy (Trueness) Traceability Manufacturer provides calibration traceability to the highest order reference method or reference material available • Manufacturer’s product calibrator has method (and sometimes reagent lot) specific target values • Cannot mix reagents and calibrators from different manufacturers 23

Terminology Trueness: the mean bias between a routine method and a reference method determined from replicate measurements of multiple patients’ samples • Imprecision is removed (or minimized) by replication Accuracy: the difference between a routine method and a reference method for a single measurement on a single patient’s sample • Difference includes influence of both bias and imprecision for a single measurement Refer to the CLSI harmonized terminology database for more detailed definitions at www. clsi. org/resources 24

Calibration Verification • Verify that measurement system calibration has not changed since the last calibration event And, where applicable (many routine methods), • Verify that method calibration is in conformance to the manufacturer’s specifications • (i. e. is consistent with the manufacturer’s calibration traceability) • Special case (uncommon) when a lab can verify calibration traceability to an accuracy based reference system 25

Commutable Reference/ PT/ QC Material • Has equivalent mathematical relationships among the results of different measurement procedures for a reference material and for native clinical samples (International Standards Organization, Clinical and Laboratory Standards Institute derived). • Produces a numeric result from a routine method that is the same as would be obtained for a native clinical sample that has the same quantity of an analyte (measurand) present. 26

Non-commutable materials may have a matrix bias • Bias in a result that is attributable to modification of the usual clinical sample matrix caused by preparation of a reference material, PT material or QC material • CAP has investigated matrix bias in PT materials using fresh frozen off-the-clot samples (that are expected to be commutable with clinical samples) è CAP Home (www. cap. org) > CAP Reference Resources and Publications > Chemistry • CAP has developed accuracy based PT Surveys from this research 27

CLSI EP 14 to validate commutability Test Method 10 95% prediction interval 8 6 4 Clinical Specimen 2 0 0 2 4 6 8 10 Comparison Method (1 RMP if available) 28

CLSI EP 14 to validate commutability Test Method 10 95% prediction interval 8 6 4 Clinical Specimen RM Commutable 2 0 0 2 4 6 8 10 Comparison Method (1 RMP if available) 29

CLSI EP 14 to validate commutability Test Method 10 95% prediction interval 8 6 4 Clinical Specimen RM Commutable RM Not-Commutable 2 0 0 2 4 6 8 10 Comparison Method (1 RMP if available) 30

Effect of non-Commutable PT materials du Pont Dimension Cholesterol (mg/d. L) 35 0 Patient bias = 0. 2% 30 0 PT bias = -9. 5% 25 0 20 0 Patient seru m CAP Survey 15 0 20 0 25 0 30 0 35 0 Ab ell-Ken d all Ch ole sterol (m g/d L) Reference Method A d apt ed fro m: Nait o, et. al. , Arc h Path ol L a b Me d. 1 993; 11 7: 3 45 -51 31

CAP Accuracy Based Surveys • Glycohemoglobin (GH 2) – uses fresh whole blood with National Glycohemoglobin Standardization Program target values • Creatinine Accuracy Calibration Verification and Linearity (LN 24) – uses off-the-clot serum with NIST target values • Accuracy Based Lipids (ABL) – uses off-the-clot serum with CDC reference method target values • Additional accuracy based Surveys are in development for analytes with clinical practice guidelines 32

Calibration Verification Run assay materials with measurement system and compare results to measurement system appropriate target values NO Re-calibrate Result = Target ? YES Calibration is OK 33

Calibration Verification Materials • Method product calibrator(s) • Method vendor provided materials for calibration verification • Previously tested clinical specimens • Reference materials with matrix and target values appropriate for the method • PT (or PT validated) materials with matrix and target values appropriate for the method, including materials designed for calibration verification 34

Verification of Calibration Trueness (or accuracy for individual samples) • Requires a reference material that is commutable native clinical samples (not commonly available) with • Most PT materials are not commutable due to the cost to prepare commutable materials and the practical need to have many analytes in the same material 35

LN Surveys for Calibration Verification Lab verifies conformance to the manufacturer’s calibration, using: è Commutable reference materials that have been validated to give results equivalent to those for native clinical samples (e. g. CAP LN 24 Creatinine accuracy based Survey with NIST target values) è Non-commutable materials with target values appropriate for a specific method (e. g. most CAP LN Survey materials with peer group evaluation) 36

Calibration Verification Concentration Levels • Approximately the same as the calibrators used for method calibration • i. e. to confirm correct calibration • Can also verify the AMR if concentrations, or activities, of the calibrators cover the full measurement range. • LN Survey provides a series of defined concentrations to verify calibration over the AMR 37

Frequency of Calibration Verification • Manufacturer’s instructions • Troubleshoot a QC problem • Calibration OR calibration verification at least every 6 months (CLIA), or when: • Change in critical reagents, component, or maintenance • If cal <6 months do not need calibration verification unless part of manufacturer’s calibration process 38

CONCENTRATION RECOVERED Analytical Measurement Range Verification Non-linear region Specimen concentration relationship is linear over the AMR CONCENTRATION EXPECTED Based on: (a) admixture ratio (b) assigned value 39

Method Calibration and AMR Verification can be a single process Analytical signal = reportable values ANALYTICAL MEASUREMENT RANGE Calibrator must: 1. Have at least 3 values that span the AMR, and 2. Have target values specific for the method (and reagent lot, if necessary). Method Calibrator No separate verification of AMR is necessary 40

AMR Verification 1. Specimens (can use patients) with low and high values 2. Mix 1: 1 to get midpoint value 3. Mix low: mid and mid: high to get additional values 4. Expected values are: Sample Proportion Low 0% 0. 75 L + 0. 25 H 25% 0. 50 L + 0. 50 H 50% 0. 25 L + 0. 75 H 75% High 100% Concentration expected 20 mg/d. L 160 mg/d. L 299 mg/d. L 438 mg/d. L 578 mg/d. L Criteria for acceptability = recovery of concentration in proportion to the amount present (linear relationship) 41

AMR Verification • AMR verification can be based on demonstrating linearity either by measuring recoveries or by demonstrating expected proportions 42

Measured mg/d. L AMR Verification 600 400 200 0 0 50 100 Expected Proportion Linear relationship verifies recovery of correct concentration relationship over AMR. 43

Materials for AMR Verification • Linearity material of appropriate matrix • PT (or PT validated) materials • Previously tested clinical specimens, altered by admixture, dilution, spiking • Reference materials with matrix and target values appropriate for the method • Method product calibrator(s) 44

AMR: how close to low and high limits • Goal is to verify acceptable clinical performance • For low limit: use clinical judgment, e. g. • creatinine, 0. 2 or 0. 3 mg/d. L • For high limit: within 10 -15% of upper limit of AMR • Challenging to find residual samples for some analytes; need to use clinical judgment, e. g. • blood gas (use commercial materials) • cardiac markers, e. g. Troponin (use LN Survey) 45

CAP LN Surveys • Most evaluate performance vs. peer group due to commutability limitations (matrix effect bias) • LN 24 offers matrix-free creatinine challenges with reference method values assigned • Verify that calibration conforms to manufacturer’s specifications • Verify that calibration is correct over the analytical measurement range • When accuracy based (e. g. , LN 24), verify that calibration is traceable to a reference system 46

Example of LN 2 Report (Calcium) 47

Example of LN 2 Report (Calcium) 48

Example of LN 24 Report (Creatinine) 49

Example of LN 24 Report (Creatinine) 50

Example of a Non-Linear Result Creatinine 51

Calibration Verification and Linearity Users Guide 52

Clinically Reportable Range • Range of analyte values that will be reported as a quantitative result, allowing for specimen dilution or other pretreatment to extend the AMR • Policy based on the laboratory director’s medical judgment regarding the clinical use requirements for a laboratory test 53

Clinically Reportable Range • Truncation or extension of the AMR based on clinical requirements. è Established once when method is put in service è Does not need to be verified on a recurring schedule • Values outside the CRR are reported as < or > a numeric value. • Checklist question is: Are dilution protocols and diluents (or concentration protocols) specified for all methods for which the CRR exceeds the AMR? 54

Clinically Reportable Range Example: h. CG may have a CRR with both low and high limits. • AMR is 3 -1, 000 m. IU/m. L (method specification). • CRR is 5 -1, 000 m. IU/m. L (clinical usefulness). • Results of <3, 3, or 4 are reported as “<5 m. IU/m. L”. • Results >1, 000 are diluted and rerun to obtain quantitative values up to 1, 000 m. IU/m. L. • Results >1, 000 are reported as “>1, 000 m. IU/m. L”. • NOTE: CRR limits are the judgment of the lab director and may be different for different labs 55

Clinically Reportable Range Example: AST may have a CRR with only a low limit. • AMR is 4 -900 IU/L (method specification). • CRR is 4 IU/L to any value (clinical usefulness). • Results of <4 are reported as “<4 IU/L”. • Results >900 are diluted and re-assayed. • The upper CRR is not specified because specimens are diluted until a quantitative value is obtained. 56

Summary Calibration or Calibration Verification at least every 6 months. è Calibration verification is not necessary if calibration is performed at least every 6 months. 57

Summary Analytical Measurement Range verification at least every 6 months. è May be required when change a critical reagent lot or other analytical parameter that can influence the AMR. è Not required when a recalibration is performed. è Not required when method calibrators include at least 3 concentrations that span the full AMR. • REQUIRED for initial method validation. 58

Summary Clinically Reportable Range is established once when a method is introduced based on clinical requirements as defined by the lab director. è SOP for dilution or other specimen pre-treatment must be defined. è Not required to verify CRR on a recurring schedule 59

Summary For additional information on the content of the Audioconference or questions about the CAP Calibration Verification/Linearity Surveys, please contact the CAP: By telephone: Customer Contact Center at 800 -323 -4040, option 1# By email: contactcenter@cap. org By mail: Surveys Program 325 Waukegan Road Northfield IL 60093 -2750 60