CORRELATION​

Method Comparison​​
INTRODUCTION
The following slides are intended to give a better understanding of performing a Correlation Study using the ALCOR ESR
analyzer.

Correlations (or method comparisons) are performed as part of method validation to assess whether two analytical methods are
comparable and is one way for the laboratory to verify the accuracy claims of the manufacturer.

Confidential & Proprietary Information – Property of ALCOR Scientific Inc.

ACCURACY

Accuracy is the ability of the new test system to produce results that are correct and within the manufacturer’s claims. When
using correlation studies the laboratory is either:

• verifying that results obtained are statistically comparable to the existing method; or

• verifying that results obtained are statistically comparable to the reference method

Measurement of accuracy can also be accomplished by testing quality control samples and/or participation in an external
quality assessment (EQA).

Confidential & Proprietary Information – Property of ALCOR Scientific Inc.

PROTOCOL

The Correlation Study Protocol for ALCOR ESR analyzers is used to evaluate the For complete Correlation Protocol
instruction, please refer to:
following performance characteristics:
• iSED Correlation Protocol (112-07-002,
• accuracy
Latest Revision); and/or

• reportable range • miniiSED Correlation Protocol (1017-07-01,

Latest Revision)

The specification for acceptance is based on the Clinical and Laboratory Standards Institute-
Procedures for the Erythrocyte Sedimentation Rate Test; Approved Standard – Fifth Ed., 2011
(CLSI).

Confidential & Proprietary Information – Property of ALCOR Scientific Inc.

SAMPLE REQUIREMENTS
FOR ANALYSIS
• Sixty (60) samples with a value distribution between 15 to 105 mm/hr, and at least 40% of
the samples with a pathological state (result over 35 mm/hr) Occasional blood samples may fail

• Samples from both males and females are to be tested to give a clear plasma-erythrocyte
interface after sedimentation; if this
• Samples must be collected in EDTA anticoagulant 13 x 75 mm closed sample tube
(lavender tube) with a minimum of 3.5mL sample volume occurs in either the test system or
standardized method, the pair of
• Samples must be non-hemolyzed and without the presence of clots
values should be eliminated from
• Samples must be processed within four (4) hours from venipuncture the data set.

• Hct and MCV data must be collected for each sample to aide in data analysis

Confidential & Proprietary Information – Property of ALCOR Scientific Inc.

STATISTICAL ANALYSIS
Statistical analysis is used to determine agreement or linearity checks between the test and standardized methods.

Unfortunately, there is no linearity material available for any ESR analyzer on the market and therefore, standard R evaluation
with the EP evaluator cannot be used to determine linearity. Instead, the CLSI procedure recommends that linear regression should
be performed using a method appropriate for the data eg, Passing-Bablok, Bland Altman.

Clinical and Laboratory Standards Institute

Procedures for the Erythrocyte Sedimentation Rate
Test; Approved Standard – Fifth Ed.

H02-A5, Vol. 31 No. 11., 2011

Confidential & Proprietary Information – Property of ALCOR Scientific Inc.

PASSING & BABLOK
REGRESSION ANALYSIS
The preferred method used for data analysis of the ALCOR analyzers is the Passing-
Bablok regression analysis because the data set for ESR testing does not follow
typical Gaussian distribution (non-parametric distribution). Additionally, this mode of
regression analysis is not as sensitive to errors and data outliers, which is important
since ESR is a measurement of a physiological reaction instead of a volume,
concentration, or number.

Passing & Bablok regression analysis calculates the cumulative sum (Cusum) test
value which determines the linear relationship between the two methods. A Cusum
test value of P>0.05 indicates that the data is linear.

Confidential & Proprietary Information – Property of ALCOR Scientific Inc.

DATA ANALYSIS USING
MEDCALC SOFTWARE
Once the information has been submitted, the data will be organized for statistical evaluation to determine equivalence.

Data analysis is then performed using MedCalc® Statistical Software, Version 19 and the resulting report is then provided
to the laboratory for their review.

Confidential & Proprietary Information – Property of ALCOR Scientific Inc. MedCalc Software bvba, Ostend, Belgium; https://www.medcalc.org; 2019
METHOD EQUIVALENCE
The test method and the standardized method can be considered y = 0.9808 x=1.1538
Proportional differences 
equivalent if: Slope 0.9808

95% Confidence limits 0.9000 to 1.0500

1. The 95% confidence interval for b (slope) contains the value 1 (b is the Systematic differences 

measurement of the proportional differences and is calculated by using median slopes of Intercept -1.1538

95% Confidence limits -2.1500 to 0.0000

all pairs)
Linear model validity 
Cusum test for linearity No significant deviation from linearity (P=0.35)
2. The 95% confidence interval for a (intercept) contains the value 0 (a is the
measurement of the systematic differences and is calculated by using median of all pairs)   Variable X Variable Y
Lowest value 1.0000 1.0000

Highest value 65.0000 64.0000

Arithmetic mean 18.8305 17.0847
Median 12.0000 10.0000
If the data are analyzed by paired t-test statistics, the values for mean differences,
Standard deviation 17.8385 17.0641
standard deviation (SD) or differences, t, and significance (P) should be included Standard error of the mean 2.3224 2.2216

in the report

Confidential & Proprietary Information – Property of ALCOR Scientific Inc.

RECORDING
SAMPLE DATA
• For each sample, record ESR result data obtained using the test method and standardized method.

TEST METHOD "Y" STANDARDIZED METHOD "X"

ALCOR ESR Analyzer Model No. Information Brand/Model No. Information
Serial Number:   Serial Number:  

• Remember to include HCT and MCV information for every sample tested.
Sample # Sample ID Result "Y" Result "X" HCT MCV
1          
2           When testing is complete:
3          
4          
Send completed spreadsheet to
5          
techservice@alcorscientific.com for
data analysis and report.

Confidential & Proprietary Information – Property of ALCOR Scientific Inc.

FACTORS WHICH
IMPACT CORRELATION
During the correlation analysis the laboratory may see variations in individual test results due to several factors. Individual pairs
of data should not be used to confirm or deny the acceptability of a correlation study.

Factors which could affect the correlation studies include:

• method variation (aggregation vs. sedimentation)

• Distribution of sample results

• Sample stability

Confidential & Proprietary Information – Property of ALCOR Scientific Inc.

METHOD VARIATION
ALCOR’s ESR systems directly measure the first phase of RBC sedimentation (aggregation) unlike traditional ESR
methodologies which measure the effect of gravity on RBC (sedimentation) .

Two factors that historically have caused inaccurate or falsified values when measuring ESR by sedimentation are hematocrit
(Hct) and mean corpuscular volume (MCV).

The ALCOR ESR test systems are less impacted by Hct and MCV, thus values obtained for ESR provide a more accurate
picture of inflammation.

Confidential & Proprietary Information – Property of ALCOR Scientific Inc.

THE EFFECT OF HCT &
MCV ON RBC

HCT is the ratio of RBC to volume of whole blood sample (%) Indices Effect on ESR (sedimentation)
Samples which have a low ratio of RBCs to plasma (low Hct), will sediment
Low HCT (<40) Falsely elevated ESR
more quickly.
High HCT (>40) Falsely decreased ESR

MCV is the size of the RBC

Low MCV (<80) Falsely decreased ESR
Small RBCs have less surface area and are out of balance with the
High MCV (>80) Falsely elevated ESR
surrounding plasma. The smaller RBCs cannot overcome the force
imposed by the surround plasma and will not sediment by gravity in the
traditional methods.

Confidential & Proprietary Information – Property of ALCOR Scientific Inc.

FABRY CORRECTION
Individual pairs of data or a complete data set can be evaluated for the
Thomas L. Fabry (1987) proposed that ESR
effects of hematocrit.
results obtained by sedimentation should be
normalized using an Hct value of 40%. By
Formula
applying a correction factor to the observed
Corrected ESR = (Ref. ESR Result x 15) / (55 – Ref. Hct)
sedimentation results, the effects of anemia (low
Hct) are removed and a more accurate clinical
Sedimentation HCT Corrected ESR ALCOR ESR evaluation can be performed.
55 20.6 24 30

90 25.4 46 40

25 24.1 12 15 Reference

Confidential & Proprietary Information – Property of ALCOR Scientific Inc.

DISTRIBUTION OF
SAMPLE RESULTS
CLSI recommends and the ALCOR protocol calls for an analysis of results with a value
We understand testing can only be
distribution between 15 and 105 mm/hr with at least 40% of the results being in the done on samples available to you.
pathological state (result >35 mm/hr). This is important as the statistical method being Therefore, it is important to
used is the Passing-Bablok regression. understand the need to exclude
certain results data points (sample
However, this does not mean that results outside of the 15 to 105 mm/hr range should be
results) for proper analysis.
omitted or not submitted. In fact, for a thorough correlation analysis, all results should
be submitted for a technical review.

During the review, an assessment will be made to determine if some or all the values
outside of the suggested range should be used in the analysis.

Confidential & Proprietary Information – Property of ALCOR Scientific Inc.

SAMPLE STABILITY

Sample stability when using the ALCOR ESR analyzers is good for up to 24 hours at room temperature or refrigerated (112-07-502).
Sample stability varies amongst the ESR systems; thus, it is recommended to follow the CLSI recommendation.

The ALCOR Correlation Protocols align with the CLSI recommendation of testing samples within 4 hours of venipuncture and at
room temperature.

Confidential & Proprietary Information – Property of ALCOR Scientific Inc.

THANK YOU!