Printer Friendly

Utilization of assay performance characteristics to estimate hemoglobin [A.sub.1c] result reliability.

Laboratory QC procedures are implemented to detect, reduce, and correct deficiencies in the testing process, with the goal of quickly identifying important errors before patient results are released (1). Historically, several options have been available for meeting CLIA QC requirements for nonwaived testing. The traditional approach requires that 2 levels of external QC be run each day of testing. As manufacturers, large reference laboratories, and hospital laboratories began collecting QC data, they noted that some test systems rarely failed QC and questioned the frequency at which external QC was required. In response, CLIA developed the Equivalent QC option, which reduced the number of external QC tests required for eligible methods. This Equivalent QC option will be discontinued in 2016. Recently, the Centers for Medicare and Medicaid Services announced a new type of QC plan, the Individualized QC Plan (IQCP), [7] beginning January 2014, that allows laboratories to utilize risk management strategies to design a QC program. Laboratories will be able to choose either the traditional approach of testing 2 levels of QC per assay, per day of patient testing, or they may elect to develop the newly introduced IQCP, which determines analytical QC frequency by utilizing risk management principles.

A risk assessment can be performed to determine if the current QC practice is adequate or requires revision (2). Currently there is minimal guidance available regarding how laboratories may quantitatively estimate risk to optimize analytical QC criteria appropriate for an IQCP (2). For the laboratory, risk is related to the chance of producing and reporting unreliable patient results, which are defined as results containing measurement errors that exceed an allowable total error ([TE.sub.a]) specification. Evaluation of analytical performance characteristics, assay requirements, a metrics, and statistical QC plans is one way to estimate risk during the analytical phase of testing.

The expected number of unreliable patient results reported when an assay is out of control is a useful metric for characterizing a laboratory's QC strategy in relation to its analytical performance capabilities. When an out-of-control condition occurs in the laboratory, the percentage of unreliable patient results produced while the out-of-control condition exists will differ from the in-control percentage of unreliable results. The number of unreliable patient results produced because of an out-of-control condition will depend on the change in percentage of unreliable results due to the out-of-control condition and the number of patient samples examined before the laboratory's QC procedures detect the out-of-control condition.

Hemoglobin (Hb) [A.sub.1c] is an ideal assay to pilot risk assessment of reporting unreliable patient results because (a) the majority of manufactured Hb [A.sub.1c] assays in the US are certified by the NGSP (formerly the National Glycohemoglobin Standardization Program) with stringent analytical performance requirements, (b) multiple testing methods and technologies are available and used in laboratory and/or point-of-care (POC) settings, and (c) there is considerable knowledge about the clinical impact of test results. Further, the prevalence of diabetes mellitus and prediabetes is increasing around the world and may climb to 50% of the population in the US by 2020 (3). Given that current guidelines recommend the use of Hb [A.sub.1c] for the diagnosis and monitoring of diabetes (4), laboratories may see substantial increases in the volume of Hb [A.sub.1c] orders.

Currently all Hb [A.sub.1c] assay manufacturers standardize to the NGSP Reference Method (or technically NGSP Designated Comparison Method) (5). In 2013, the College of American Pathologists (CAP) proficiency testing acceptance limit decreased to [+ or -]6%, leading laboratories to closely scrutinize their Hb [A.sub.1c] assay performance characteristics and QC practices. There is, however, limited information available regarding the risk of reporting erroneous Hb [A.sub.1c] results when using NGSP-certified methods. Although the CAP GH2 proficiency testing survey highlights the accuracy and variation within and between Hb [A.sub.1c] assays, less is known about how commonly recommended routine QC practices and assay performance affect the reliability of an Hb [A.sub.1c] result even when these assays pass proficiency testing.

The aim of this study was to evaluate the risk of reporting unreliable Hb [A.sub.1c] results when using currently available NGSP-certified Hb [A.sub.1c] methods. Six different Hb [A.sub.1c] assays across 4 academic medical centers were evaluated using assay performance characteristics according to CLSI protocols. In the new era of risk-based QC plans, this provides one example of quantitative risk estimates that can guide QC strategies appropriate for an IQCP.

Materials and Methods

Hb [A.sub.1c] ASSAYS

Hb [A.sub.1c] was measured on 6 different analyzers across 4 academic medical centers. These included the Variant II Turbo (Bio-Rad), Variant II (Bio-Rad), and Tosoh G8 (Tosoh Bioscience), which are based on ion-exchange HPLC; the Capillarys 2 Flex Piercing (Sebia), which is based on capillary electrophoresis; the COBAS Integra 800 (Roche Diagnostics), which is based on agglutination immunoassay; and DCA Vantage (Siemens), which is based on immunoassay. Two different DCA Vantage instruments using 2 different lots of calibrator were evaluated. The Dimension ExL (Siemens) was also evaluated in this study. However, while this manuscript was in the review process, the manufacturer withdrew from the market the reagent lot that was evaluated. Therefore, these data have been excluded from the study.

All assays tested were NGSP certified as of September 2012.

NGSP SAMPLES

Forty NGSP secondary reference laboratory (SRL) target value-assigned samples (Dr. Randie Little, University of Missouri, performed testing and provided samples for this study for a fee; NGSP SRL) were sent to each laboratory and stored at -80[degrees]C until analysis.

PRECISION AND BIAS STUDIES

Precision for each assay was determined using the CLSI EP5-A2 protocol. Respective laboratory Hb [A.sub.1c] QC materials (both low QC and high QC) were assayed in duplicate twice per day (morning and afternoon) for a total of 20 days. Linear regression and bias were determined according to the CLSI EP9-A2 protocol. Eight of 40 NGSP SRL samples were thawed each day and tested in duplicate over a period of 5 days.

STATISTICAL ANALYSES

A representative patient distribution of Hb [A.sub.1c] values was obtained from 1 facility over a 2-week period. Sigma values [([TE.sub.a] - %Bias)/CV] for each instrument were calculated at each Hb [A.sub.1c] concentration and averaged over the observed Hb [A.sub.1c] patient distribution to obtain patient-weighted rvalues. Sigma values directly relate to the predicted probability of producing an unreliable patient result. Given a [TE.sub.a] specification and a procedure's %Bias and CV, the percentage of patient results predicted to be unreliable during stable operation is computed as:

In-control % unreliable

= 100{1 - [F([TE.sub.a] - %Bias/CV) - F(-[TE.sub.a] - %Bias/CV)]},

where F denotes the standard normal cumulative distribution function.

The expected number of unreliable final patient results [E([N.sub.uf])] owing to an out-of-control condition is defined as the predicted number of unreliable results produced from the inception of an out-of-control condition up to the last acceptable QC evaluation before the out-of-control condition's detection. These results are considered final because they were produced and reported before an acceptable QC evaluation. E([N.sub.uf]) depends on [TE.sub.a], the procedure's %Bias and CV, the laboratory's QC rules and frequency of QC evaluations, and the magnitude of the out-of-control condition. The method for computing the expected number of unreliable patient results has been described previously (6).

E([N.sub.uf]) was evaluated over a range of possible out-of-control conditions. Systematic error out-of-control conditions that cause a persistent systematic shift in results proportional to concentration were assessed over a range of negative and positive shifts spanning 2 multiples of [TE.sub.a]. The maximum predicted value of E([N.sub.uf]) over the range of out-of-control conditions was used to assess and compare performance of the different procedures in response to an out-of-control condition.

For these analyses, [TE.sub.a] was set to 5%, 6%, and 7% (to encompass the current state of Hb [A.sub.1c] testing acceptability in terms of current and previous manufacturer NGSP certification and proficiency testing through the CAP), QC rules were set to the 1:2s rule (with control limits set at mean [+ or -] 2 SD) with 2 QC levels, and the mean number of Hb [A.sub.1c] examinations between QC events was set to 100. Computations were performed using the MATLAB programming language (The Mathworks, Inc.).

Results

The comparisons of measured Hb [A.sub.1c] values to target NGSP SRL results across the 6 Hb [A.sub.1c] assays are shown in Fig. 1 and summarized in Table 1. Based on data evaluated at 2 QC levels, the Variant II Turbo and Capillarys 2 Flex Piercing showed the smallest overall bias, and the Tosoh G8 and Integra 800 had the largest bias (Table 1). The squared correlation for all assays ranged from 0.989 [DCA Vantage-lot 1] to 0.999 (Variant II Turbo, Tosoh G8, Capillarys2 Flex Piercing) (Table 1). Percentage bias was calculated from the linear regression relationships over the range of NGSP target value assigned Hb [A.sub.1c] levels and found to differ significantly across assay platforms (Fig. 2). The Integra 800 and Bio-Rad Variant II showed the highest variability in percentage bias across the Hb [A.sub.1c] values tested.

Within-laboratory imprecision (CV) ranged from 1.28% for the Tosoh G8 to 2.97% for the Variant II Turbo when using low Hb [A.sub.1c] QCs (Table 1). At the high QC level, imprecision ranged from 0.8% for the Tosoh G8 to 2.65% for the DCA Vantage-lot 1.

A representative distribution of approximately 1500 Hb [A.sub.1c] patient results for 2 weeks was combined with the analytical performance characteristics of each assay shown in Table 1. Together they were used to generate patient-weighted [sigma] metrics and predicted probabilities of producing unreliable patient results (measurement errors exceeding [TE.sub.a]) during stable in control operation for each Hb [A.sub.1c] assay at [TE.sub.a] specifications of 7%, 6% (the current CAP proficiency testing acceptance limit), and 5% (Table 2).

Assuming a 1:2s QC rule with 2 QCs and a mean of 100 Hb [A.sub.1c] examinations between QC events, the predicted number of unreliable final patient results expected due to the existence of an out-of-control condition was computed over a wide range of possible magnitudes of systematic error out-of-control conditions (Fig. 3). The scales of the y axes in Fig. 3 were set the same for the E([N.sub.uf]) graphs shown to provide a head-to-head evaluation of each assay's systematic out-of-control conditions. The lines of the Capillarys 2 Flex Piercing (Fig. 3D) and the DCA Vantage-lot 2 (Fig. 3G) graphs are nearly flat due to these assays' negligible predicted number of unreliable final patient results, E([N.sub.uf]), over the wide range of possible magnitudes of systematic error out-of-control conditions. The maximum expected number of unreliable final patient results [Max E([N.sub.uf])] out of 100 events over the range of out-of-control conditions was determined for each assay and each [TE.sub.a] (Table 2).

The 6 assays evaluated in this study reflect a wide range of performance characteristics. The Integra 800 showed the poorest predicted performance. Using a [TE.sub.a] specification of 6%, its bias and imprecision were associated with a patient-weighted [sigma] value of only 0.36. If 100 patient samples were tested between QC events, 39.35% of results would be unreliable while in-control, and in the worst case, as many as 71 unreliable patient results (out of 100 results total) could be expected when an out-of-control event occurs (Table 2). In contrast, the Capillarys 2 Flex Piercing assay had imprecision and bias profiles that gave the highest patient-weighted [sigma] at each [TE.sub.a] tested. Only 0.02% of results are predicted to be unreliable while in-control, and < 1 unreliable patient result (out of 100 results total) would be expected even for the worst case out-of-control condition (Table 2). As the [TE.sub.a] specification is decreased, the patient-weighted a metrics decrease, with a corresponding increase in the risk of reporting unreliable patient results.

Discussion

QC plans are commonly generated to monitor stability of laboratory instruments and methods. More recently, improvements in instrumentation and assay technology have led to a transition from using QC to monitor instrument failure to using QC to minimize risk and/or mitigate residual risk of reporting an inaccurate result. Risk management strategies, popularized years ago in industry (7), have recently been touted as an alternative to a "one-size-fits-all" QC plan that is common in many laboratories (8, 9).

The goal of this study was to investigate the impact of differences in imprecision and bias for Hb [A.sub.1c] assays on the ability to meet quality goals in terms of patient risk when using currently available NGSP-certified as says. The improvements in Hb [A.sub.1c] instrumentation performance and standardization to the NGSP prompted CAP to reduce the recommended [TE.sub.a] from 7% to 6% in 2013, with the suggestion that these limits maybe further reduced in the future. Thus, a secondary interest of this study was the impact of varying [TE.sub.a]. A fixed QC rule (1:2s rule with 2 levels of QC) and frequency of QC evaluations (every 100 Hb [A.sub.1c] examinations) was assumed for each Hb [A.sub.1c] assay tested to ensure that differences in risk were a function of only [TE.sub.a], %CV, and %Bias. This is not meant to imply an endorsement of a particular QC rule or frequency at which QC should be run.

The overall imprecision and bias are important for interpretation of Hb [A.sub.1c] results. Currently, an intralaboratory imprecision (% CV) of <2% is recommended (10). All assays except the Bio-Rad Variant II Turbo (low QC = 2.97% CV), Roche Integra 800 (low QC = 2.4% CV), and Siemens DCA Vantage-lot 1 (high QC = 2.65% CV) met this goal at the 2 clinically relevant Hb [A.sub.1c] levels (low and high) tested. A %Bias of > [+ or -] 3 accounts for one-half the allowable limit ([+ or -] 6%) afforded by the CAP Hb [A.sub.1c] proficiency testing program. Four assays out of 6 in this study displayed a bias of >3%, indicating a potential larger role for bias in the overall assessment of Hb [A.sub.1c] method performance. In this study, bias was sometimes greater at lower or higher Hb [A.sub.1c] concentrations (Fig. 2).

Assessment of bias or poor calibration in a timely fashion is sometimes difficult without instituting additional checks into routine practice. In addition, targets of internal QC may be unreliable, and comparisons that include large numbers of NGSP target value-assigned samples in routine laboratory operations are not easily accomplished. One suggestion is that calibration verification samples (if available through the manufacturer or NGSP) be run alongside QC material after calibration and/or at some predefined time interval of routine testing. Additional analysis of smaller sample sizes of NGSP target value-assigned specimens may also be performed and compared within-laboratory or across-laboratories. Although each of these suggestions appears valuable, their cost and acceptability would have to be evaluated by each laboratory.

We used both precision and bias to generate several outcome metrics, including patient-weighted a metrics, in-control percentage unreliable patient results, and maximum expected number of unreliable patient results due to an out-of-control condition [Max E([N.sub.uf])]. Sigma values are directly related to the predicted probability of producing unreliable patient results during stable operation, which may be expressed in terms of defects per million opportunities (DPMO) (7). A 6-[sigma] method is associated with 3.4 DPMO and is classified as "world class quality." Patient-weighted [sigma] values are a weighted average of [sigma] values across a spectrum of patient results. We used these because they are a more accurate reflection of [sigma] metrics for a derived patient population (11). Our results demonstrate that there is not only substantial variability in the metrics across platforms as a result of differing analytical performance, but also a sizeable impact from adjusting the [TE.sub.a] specification.

At the time of this study, all but 1 Hb [A.sub.1c] assay (Capillarys 2 Flex Piercing) had proficiency testing data available through the CAP, and those assays that were in use for clinical practice at the 4 academic medical centers at the time of this study all successfully passed their CAP GH2 surveys, indicating the observed bias did not affect their ability to pass proficiency testing. However, note the existence of negative E([N.sub.uf]) values for some of the out-of-control conditions shown in Fig. 3. These reflect situations in which the magnitude and direction of the out-of-control condition negates the inherent bias in an assay, thereby reducing the likelihood of measurement errors exceeding [TE.sub.a] compared to the in-control state.

Interestingly, the 2 different calibrator lots (lot 1 and lot 2) tested for the DCA Vantage point-of-care assay performed better and demonstrated higher patient-weighted [sigma] values than some of the clinical laboratory Hb [A.sub.1c] assays tested. Although the analytical performance of this method has already been shown to be superior to other POC methods (12), this is the first report demonstrating that analytical performance of the DCA assay alone can lead to a reduction in the maximum expected number of unreliable patient results from an out-of-control condition. Although this method performance superiority was evident for both calibrator lots, one caveat to this interpretation is that our assessment did not account for any potential preanalytical collection variables at the point-of-care level that may have affected the potential quality of point-of-care Hb [A.sub.1c] results.

Westgard QC rules have been available for many years as a guide for monitoring QC. However, laboratories have for the most part failed to optimize their QC procedures (7), opting instead for a one-size-fits-all 2-SD rule. It is important to note that large differences in analytical performance characteristics were observed based on the total volume of patient samples analyzed between QC events, indicating that a one-size-fits-all QC plan is not appropriate. Except for the Capillarys 2 Flex Piercing, the patient-weighted a metric for all platforms investigated at a [TE.sub.a] of 6% was <3, indicating that maximum QC (3 levels, 3 times per day) should be performed to achieve the necessary error detection. For this study, the set amount of patient testing between QC events was 100. E([N.sub.uf]) is proportional to the number of patient samples tested between QC events. If the number of patients tested between QC events changes, the risk of reporting unreliable results may also change. For example, if the number of Hb [A.sub.1c] patient samples tested between QC events was set at 10 instead of 100, the max E([N.sub.uf]) when using the Roche Integra 800 would be <1 out of 100 (at a [TE.sub.a] of 7%). Conversely, if you double the number of patient samples between QC events, E([N.sub.uf]) will also double. Exhaustive QC events are often cost prohibitive and can frequently result in a more complex mechanism of patient testing. The investigation of assay performance and potential approaches for its improvement may yield a better overall solution. A laboratory can alternatively implement different QC designs to reduce cost (13).

Our results show how analytical characteristics can be used to assess the risk of reporting an unreliable result. The model incorporates the 3 types of characteristics that contribute to patient risk: (a) the performance characteristics of the testing method (imprecision and bias), (b) the QC strategy used by the laboratory [number of QC samples, QC rule(s), and QC frequency], and (c) the quality required of the analyte ([TE.sub.a]). Each of these characteristics must be assessed by the laboratory to claim results are fit for their intended use. Of these, bias and [TE.sub.a] are characteristics laboratories likely may have the most difficulty with. However, evaluation can be performed first assuming zero bias and again at an alternative bias derived from peer group comparisons to assess the difference in patient risk implications. Likewise, if it is unclear what [TE.sub.a] to use, different quality specifications can be tested before implementation to assess the impact on patient risk.

This study demonstrates the importance of aligning the risk of reporting unreliable Hb [A.sub.1c] results with the instrumentation, assay, and patient volumes of the individual laboratory. Although currently available NGSP-certified Hb [A.sub.1c] assays can yield satisfactory results with external quality assessment programs, such as proficiency testing, it is important that the limitations of these assays are well understood by laboratory medicine professionals.

Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, oranalysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article.

Authors' Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts of interest:

Employment or Leadership: J. Yundt-Pacheco, Bio-Rad Laboratories; L. Kuchipudi, Bio-Rad Laboratories; C.A. Parvin, Bio-Rad Laboratories.

Consultant or Advisory Role: None declared.

Stock Ownership: J. Yundt-Pacheco, Bio-Rad Laboratories; C.A. Parvin, Bio-Rad Laboratories.

Honoraria: A. Woodworth, AACC; R. Molinaro, Bio-Rad, Inc., and Sebia, Inc.

Research Funding: None declared.

Expert Testimony: None declared.

Patents: None declared.

Other Remuneration: A. Woodworth, AACC.

Role of Sponsor: No sponsor was declared.

References

(1.) Midyett R. "Empty QC": when QC is out of control. MLO Med Lab Obs 2012;44:22, 24, 26.

(2.) Williams M. Risk assessment and quality control: be ready for the new guidance. MLO Med Lab Obs 2012;44:18,20-1.

(3.) United Health Center for Health Reform and Modernization. The United States of Diabetes: challenges and opportunities in the decade ahead. http://www.unitedhealthgroup.com/~7media/UHG/ PDF/2010/UNH-Working-Paper-5-Fact-Sheet.ashx (Accessed May 2014).

(4.) American Diabetes Association. Standards of medical care in diabetes-2013. Diabetes Care 2013;36 Suppl 1:S11-66.

(5.) Little RR, Rohlfing CL, Sacks DB; National Glycohemoglobin Standardization Program (NGSP) Steering Committee. Status of hemoglobin [A.sub.1c] measurement and goals for improvement: from chaos to order for improving diabetes care. Clin Chem 2011; 57:205-14.

(6.) Parvin CA. Assessing the impact of the frequency of quality control testing on the quality of reported patient results. Clin Chem 2008;54:2049-54.

(7.) Schoenmakers CH, Naus AJ, Vermeer HJ, van Loon D, Steen G. Practical application of Sigma Metrics QC procedures in clinical chemistry. Clin Chem Lab Med 2011;49:1837-43.

(8.) CLSI. Laboratory quality control based on risk management: approved guideline. Wayne (PA): CLSI; 2011. CLSI document EP23-A.

(9.) Westgard JO. Perspectives on quality control, risk management, and analytical quality management. Clin Lab Med 2013;33:1-14.

(10.) Sacks DB, Arnold M, Bakris GL, Bruns DE, Horvath AR, Kirkman MS, et al. Guidelines and recommendations for laboratory analysis in the diagnosis and management of diabetes mellitus. Clin Chem 2011;57:e1-47.

(11.) Kuchipudi L, Yundt-Pacheco J, Parvin CA. Computing a patient-based sigma metric. Clin Chem 2010;56:A-107.

(12.) Lenters-Westra E, Slingerland RJ. Six of eight hemoglobin [A.sub.1c] point-of-care instruments do not meet the general accepted analytical performance criteria. Clin Chem 2010;56:44-52.

(13.) Westgard JO. Best practices for "Westgard Rules." http://www.westgard.com/lesson74.htm (Accessed October 2013).

Alison Woodworth, [1] Nichole Korpi-Steiner, [2] James J. Miller, [3] Lokinendi V. Rao, [4] John Yundt-Pacheco, [5] Lakshmi Kuchipudi, [5] Curtis A. Parvin, [5] Jeanne M. Rhea, [6] and Ross Molinaro [6] *

[1] Vanderbilt University, Nashville, TN; [2] University of North Carolina, Chapel Hill, NC; [3] University of Louisville, Louisville, KY; [4] UMass Memorial Medical Center, Worcester, MA; [5] Bio-Rad Quality System Division, Plano, TX; [6] Emory University School of Medicine, Atlanta, GA.

* Address correspondence to this author at: Emory University School of Medicine, Emory University Hospital Midtown, 550 Peachtree Ave. NE, Davis Fischer Bldg., Rm. 1239, Atlanta, GA 30308. Fax 404-727-9656; e-mail rjmolin@ emory.edu.

Received January 2, 2014; accepted April 21, 2014.

Previously published online at DOI: 10.1373/Clinchem.2013.220772

[7] Nonstandard abbreviations: IQCP, Individualized QC Plan; [TE.sub.a], allowable total error; Hb, hemoglobin; POC, point-of-care; CAP, College of American Pathologists; SRL, secondary reference laboratory; E(NUf), expected number of unreliable final patient results; Max E([N.sub.uf]), maximum E([N.sub.uf]); DPMO, defects per million opportunities.

Table 1. Assay performance characteristics across platforms/
sites: imprecision and bias across 2 QC levels.

Assay platform                   Low QC

                    Mean   % Bias (a)   % CV (b)

Variant II          5.09     -4.99        1.43
Variant II Turbo    5.18     -0.08        2.97
Tosoh G8            5.75      3.99        1.28
Capillarys 2        5.24     -0.33        1.66
Integra 800         5.61      5.76        2.40
DCA Vantage-lot 1   5.31     -0.34        1.88
DCA Vantage-lot 2   5.23     -0.37        1.93

Assay platform               High QC            Linear       [r.sup.2]
                                            regression (c)
                    Mean    % Bias   % CV

Variant II          9.74     2.00    1.33   1.107x - 0.834     0.991
Variant II Turbo    10.07    0.10    1.81   0.999x + 0.012     0.999
Tosoh G8            9.60     4.98    0.80   1.064x - 0.130     0.999
Capillarys 2        7.93    -0.01    1.33   0.998x + 0.016     0.999
Integra 800         9.90     4.07    1.18   1.014x + 0.242     0.997
DCA Vantage-lot 1   10.31    2.72    2.65   1.038x - 0.161     0.989
DCA Vantage-lot 2   10.49    1.73    1.81   1.024x - 0.109     0.991

(a) % Bias = 100 x (observed mean-assigned value)/assigned value.

(b) Precision calculations follow CLSI EP-5A.

(c) Linear regression of assay results compared to NGSP results
measured on a Tosoh HPLC.

Table 2. Risk analysis for Hb [A.sub.1c] assays at 3 different
[TE.sub.a] limits.

                                 Patient-weighted sigma

Assay platform      7% [TE.sub.a]   6% [TE.sub.a]   5% [TE.sub.a]

Variant II              2.30            1.57            0.83
Variant II Turbo        2.67            2.29            1.90
Tosoh G8                2.27            1.43            0.59
Capillarys 2            4.56            3.90            3.25
Integra 800             0.85            0.36            -0.12
DCA Vantage-lot 1       2.84            2.36            1.88
DCA Vantage-lot 2       3.36            2.84            2.32

                                In-control % unreliable

Assay platform      7% [TE.sub.a]   6% [TE.sub.a]   5% [TE.sub.a]

Variant II              10.03           19.63           32.81
Variant II Turbo        1.23            2.96            6.62
Tosoh G8                1.19            7.80            28.53
Capillarys 2            0.00            0.02            0.19
Integra 800             25.51           39.35           55.64
DCA Vantage-lot 1       0.69            1.82            4.50
DCA Vantage-lot 2       0.05            0.29            1.32

                          Max E([N.sub.uf]) out of 100 events

Assay platform      7% [TE.sub.a]   6% [TE.sub.a]   5% [TE.sub.a]

Variant II              34.27           46.66           51.54
Variant II Turbo        5.94            11.00           18.27
Tosoh G8                13.67           49.92           83.60
Capillarys 2            0.12            0.60            2.58
Integra 800             60.11           71.48           74.00
DCA Vantage-lot 1       8.85            18.92           36.28
DCA Vantage-lot 2       2.06            6.30            16.55
COPYRIGHT 2014 American Association for Clinical Chemistry, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2014 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Endocrinology and Metabolism
Author:Woodworth, Alison; Korpi-Steiner, Nichole; Miller, James J.; Rao, Lokinendi V.; Yundt-Pacheco, John;
Publication:Clinical Chemistry
Article Type:Report
Date:Aug 1, 2014
Words:4573
Previous Article:Three of 7 hemoglobin [A.sub.1c] point-of-care instruments do not meet generally accepted analytical performance criteria.
Next Article:Quantification of serum IgG subclasses by use of subclass-specific tryptic peptides and liquid chromatography-tandem mass spectrometry.
Topics:

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters