The European Biological Variation Study (EuBIVAS): Biological Variation Data for Coagulation Markers Estimated by a Bayesian Model.

BACKGROUND: For biological variation (BV) data to be safely used, data must be reliable and relevant to the population in which they are applied. We used samples from the European Biological Variation Study (EuBIVAS) to determine BV of coagulation markers by a Bayesian model robust to extreme observations and used the derived within-participant BV estimates [CVP(i)] to assess the applicability of the BV estimates in clinical practice. METHOD: Plasma samples were drawn from 92 healthy individuals for 10 consecutive weeks at 6 European laboratories and analyzed in duplicate for activated partial thromboplastin time (APTT), prothrombin time (PT), fibrinogen, D-dimer, antithrombin (AT), protein C, protein S free, and factor VIII (FVIII). A Bayesian model with Student t likelihoods for samples and replicates was applied to derive CVP(i) and predicted BV estimates with 95% credibility intervals. RESULTS: For all markers except D-dimer, CVP(i) were homogeneously distributed in the overall study population or in subgroups. Mean within-subject estimates (CVI) were <5% for APTT, PT, AT, and protein S free, <10% for protein C and FVIII, and <12% for fibrinogen. For APTT, protein C, and protein S free, estimates were significantly lower in men than in women ≤50 years. CONCLUSION: For most coagulation markers, a common CVI estimate for men and women is applicable, whereas for APTT, protein C, and protein S free, sex-specific reference change values should be applied. The use of a Bayesian model to deliver individual CVP(i) allows for improved interpretation and application of the data.

A Bayesian Approach to Biological Variation Analysis.

BACKGROUND: Biological variation (BV) data have many applications for diagnosing and monitoring disease. The standard statistical approaches for estimating BV are sensitive to "noisy data" and assume homogeneity of within-participant CV. Prior knowledge about BV is mostly ignored. The aims of this study were to develop Bayesian models to calculate BV that (a) are robust to "noisy data," (b) allow heterogeneity in the within-participant CVs, and (c) take advantage of prior knowledge. METHOD: We explored Bayesian models with different degrees of robustness using adaptive Student t distributions instead of the normal distributions and when the possibility of heterogeneity of the within-participant CV was allowed. Results were compared to more standard approaches using chloride and triglyceride data from the European Biological Variation Study. RESULTS: Using the most robust Bayesian approach on a raw data set gave results comparable to a standard approach with outlier assessments and removal. The posterior distribution of the fitted model gives access to credible intervals for all parameters that can be used to assess reliability. Reliable and relevant priors proved valuable for prediction. CONCLUSIONS: The recommended Bayesian approach gives a clear picture of the degree of heterogeneity, and the ability to crudely estimate personal within-participant CVs can be used to explore relevant subgroups. Because BV experiments are expensive and time-consuming, prior knowledge and estimates should be considered of high value and applied accordingly. By including reliable prior knowledge, precise estimates are possible even with small data sets.

Biological variation estimates for prostate specific antigen from the European Biological Variation Study; consequences for diagnosis and monitoring of prostate cancer.

BACKGROUND: Prostate-specific antigen (PSA) is central in the diagnosis of prostate cancer. However, high-quality biological variation (BV) estimates for PSA are scarce. Here BV estimates from the European Biological Variation Study (EuBIVAS) for total (tPSA), free (fPSA), conjugated PSA (cPSA), and percent free PSA (%fPSA) are provided. METHOD: EuBIVAS samples were collected weekly from thirty-seven healthy males (22-59 years) for 10 weeks. All samples, stored at -80 °C, were measured in duplicate with a Roche Cobas e801. Outlier and homogeneity analysis were performed followed by CV-ANOVA to determine BV, analytical variation, analytical performance specifications (APS), reference change values (RCV) and the number of samples required to estimate the homeostatic set points. RESULTS: Within-subject BV estimates were for tPSA 6.8% (6.1-7.4); fPSA 7.1% (6.5-7.7) cPSA: 8.8% (8.0-9.7) and %fPSA 5.3% (4.8-5.8), delivering RCV for increase of 15-20% and indicating that one sample is sufficient to estimate the homeostatic set points within ±15%. BV estimates for tPSA were lower than previously published estimates. Estimates for fPSA, cPSA and %fPSA have not previously been reported in healthy subjects. CONCLUSIONS: Highly powered EuBIVAS BV estimates of tPSA, fPSA, cPSA and %fPSA provide updated APS and RCV for monitoring for prostate cancer.

The EuBIVAS: Within- and Between-Subject Biological Variation Data for Electrolytes, Lipids, Urea, Uric Acid, Total Protein, Total Bilirubin, Direct Bilirubin, and Glucose.

BACKGROUND: The European Federation of Clinical Chemistry and Laboratory Medicine European Biological Variation Study (EuBIVAS) has been established to deliver rigorously determined data describing biological variation (BV) of clinically important measurands. Here, EuBIVAS-based BV estimates of serum electrolytes, lipids, urea, uric acid, total protein, total bilirubin, direct bilirubin, and glucose, as well as their associated analytical performance specifications (APSs), are presented. METHOD: Samples were drawn from 91 healthy individuals (38 male, 53 female; age range, 21-69 years) for 10 consecutive weeks at 6 European laboratories. Samples were stored at -80 °C before duplicate analysis of all samples on an ADVIA 2400 (Siemens Healthineers). Outlier and homogeneity analyses were performed, followed by CV-ANOVA on trend-corrected data, when relevant, to determine BV estimates with CIs. RESULTS: The within-subject BV (CVI) estimates of all measurands, except for urea and LDL cholesterol, were lower than estimates available in an online BV database, with differences being most pronounced for HDL cholesterol, glucose, and direct bilirubin. Significant differences in CVI for men and women/women <50 years of age were evident for uric acid, triglycerides, and urea. The CVA obtained for sodium and magnesium exceeded the EuBIVAS-based APS for imprecision. CONCLUSIONS: The EuBIVAS, which is fully compliant with the recently published Biological Variation Data Critical Appraisal Checklist, has produced well-characterized, high-quality BV estimates utilizing a stringent experimental protocol. These new reference data deliver revised and more exacting APS and reference change values for commonly used clinically important measurands, thus having direct relevance to diagnostics manufacturers, service providers, clinical users, and ultimately patients.

Within-subject biological variation of activated partial thromboplastin time, prothrombin time, fibrinogen, factor VIII and von Willebrand factor in pregnant women.

BACKGROUND: During pregnancy, interpretation of results from coagulation parameters can be difficult as the physiological changes that occur may affect the biochemical parameters. The aim of this study was to describe the normal course of five coagulation parameters in healthy pregnancies, and to estimate the within-subject biological variation (CVI). METHODS: Blood samples were obtained every 4th week during pregnancy and three samples after delivery in 20 healthy women and every 4th week during a 40-week period in 19 healthy non-pregnant women. Activated partial thromboplastin time (APTT), prothrombin time (PT), PT International Normalized Ratio (INR), fibrinogen, factor VIII clot (FVIII:C) and von Willebrand factor antigen (vWF:Ag) were analyzed. The physiological changes during pregnancy were compensated by transformation into multiples of the median (MoM) and it is natural logarithm (lnMoM) in order to establish a kind of steady state, and CVI was calculated from the standard deviation. RESULTS: During pregnancy, APTT, PT and INR remained unchanged or decreased, depending upon the reagent used, while fibrinogen, FVIII:C and vWF:Ag increased gradually until delivery. The CVI in pregnancy were 2.2 and 3.0% for APTT, 2.3 and 2.6% for PT, 2.2 and 2.3% for INR, 7.2% for fibrinogen, 12.2% for FVIII:C and 11.3% for vWF:Ag, and corresponded with the CVI in non-pregnant women. CONCLUSIONS: Transformation of coagulation parameters in healthy pregnancies to MoM is a tool to establish a kind of steady state. Although there is a physiological change in these coagulation parameters during pregnancy, the CVI after lnMoM transformation was comparable with the CVI of non-pregnant women.

The Biological Variation Data Critical Appraisal Checklist: A Standard for Evaluating Studies on Biological Variation.

BACKGROUND: Concern has been raised about the quality of available biological variation (BV) estimates and the effect of their application in clinical practice. A European Federation of Clinical Chemistry and Laboratory Medicine Task and Finish Group has addressed this issue. The aim of this report is to (a) describe the Biological Variation Data Critical Appraisal Checklist (BIVAC), which verifies whether publications have included all essential elements that may impact the veracity of associated BV estimates, (b) use the BIVAC to critically appraise existing BV publications on enzymes, lipids, kidney, and diabetes-related measurands, and (c) apply metaanalysis to deliver a global within-subject BV (CVI) estimate for alanine aminotransferase (ALT). METHODS: In the BIVAC, publications were rated as A, B, C, or D, indicating descending compliance for 14 BIVAC quality items, focusing on study design, methodology, and statistical handling. A D grade indicated that associated BV estimates should not be applied in clinical practice. Systematic searches were applied to identify BV studies for 28 different measurands. RESULTS: In total, 128 publications were identified, providing 935 different BV estimates. Nine percent achieved D scores. Outlier analysis and variance homogeneity testing were scored as C in >60% of 847 cases. Metaanalysis delivered a CVI estimate for ALT of 15.4%. CONCLUSIONS: Application of BIVAC to BV publications identified deficiencies in required study detail and delivery, especially for statistical analysis. Those deficiencies impact the veracity of BV estimates. BV data from BIVAC-compliant studies can be combined to deliver robust global estimates for safe clinical application.

The EuBIVAS Project: Within- and Between-Subject Biological Variation Data for Serum Creatinine Using Enzymatic and Alkaline Picrate Methods and Implications for Monitoring.

BACKGROUND: The European Federation of Clinical Chemistry and Laboratory Medicine (EFLM) European Biological Variation Study (EuBIVAS) has been established to deliver rigorously determined biological variation (BV) indices. EuBIVAS determined BV for serum creatinine using the enzymatic and alkaline picrate measurement methods. METHOD: In total, 91 healthy individuals (38 males, 53 females; age range, 21-69 years) were bled for 10 consecutive weeks at 6 European laboratories. An equivalent protocol was followed at each center. Sera were stored at -80 °C before analysis. Analyses for each patient were performed in duplicate within a single run on an ADVIA 2400 system (San Raffaele Hospital, Milan). The data were subjected to outlier and homogeneity analysis before performing CV-ANOVA to determine BV and analytical variation (CVA) estimates with confidence intervals (CI). RESULTS: The within-subject BV estimates [CVI (95% CI)] were similar for enzymatic [4.4% (4.2-4.7)] and alkaline picrate [4.7% (4.4-4.9)] methods and lower than the estimate presently available online (CVI = 5.9%). No significant male/female BV differences were found. Significant differences were observed in mean creatinine values between men and women and between Turkish individuals and those of other nationalities. Between-subject BV (CVG) estimates, stratified accordingly, produced CVG values similar to historical BV data. CVA was 1.1% for the enzymatic and 4.4% for alkaline picrate methods, indicating that alkaline picrate methods fail to fulfill analytical performance specifications for imprecision (CVAPS). CONCLUSIONS: The serum creatinine CVI obtained by EuBIVAS specifies a more stringent CVAPS than previously identified. The alkaline picrate method failed to meet this CVAPS, raising questions regarding its future use.

Biological Variation Estimates Obtained from 91 Healthy Study Participants for 9 Enzymes in Serum.

BACKGROUND: We sought to develop estimates of biological variation (BV) for 9 enzymes in blood serum as part of the European Biological Variation Study. METHODS: Ninety-one healthy study participants (38 male and 53 female, 21-69 years old) were phlebotomized in each of 10 consecutive weeks at 6 European laboratories. The same preanalytical sample-handling protocol was followed at each center before transport to San Raffaele Hospital, Milan, Italy, for analysis. Sera were stored at -80 °C before analysis in duplicate within a single run on an ADVIA 2400 Clinical Chemistry System (Siemens Healthcare) following a protocol designed to minimize analytical imprecision. Assay traceability was established using frozen sera with target values assigned by reference methods. The results were subjected to outlier analysis before CV-ANOVA to deliver valid BV estimates. Results for 9 enzymes were subsequently partitioned for graphical display allowing visual assessment of the effects of country of origin, sex, and age on BV estimates. RESULTS: We found no effect of country upon the observed variation, but overall sex-related differences were evident for alanine amino transferase (ALT), γ-glutamyl transferase (GGT), and creatine kinase (CK). The following estimates for within-subject BV (CVI) and between-subject BV (CVG), respectively, were obtained: ALT: 9.3%, 28.2%; aspartate aminotransferase: 9.5%, 20.3%; GGT: 8.9%, 41.7%; alkaline phosphatase : 5.3%, 24.9%; lactate dehydrogenase: 5.2%, 12.6%; CK: 14.5%, 31.5%; amylase: 6.8%, 30.4%; pancreatic α-amylase: 6.3%, 24.9%; and lipase (LIP): 7.7%, 23.8%. CONCLUSIONS: All CVI and some CVG estimates were lower than those reported in the online BV 2014 updated database. Analytical performance specifications derived from BV can be applied internationally.

Biological variation: Evaluation of methods for constructing confidence intervals for estimates of within-person biological variation for different distributions of the within-person effect.

BACKGROUND: Precise estimates of the within-person biological variation, CVI, can be essential both for monitoring patients and for setting analytical performance specifications. The confidence interval, CI, may be used to evaluate the reliability of an estimate, as it is a good measure of the uncertainty of the estimated CVI. The aim of the present study is to evaluate and establish methods for constructing a CI with the correct coverage probability and non-cover probability when estimating CVI. METHOD: Data based on 3 models for distributions for the within-person effect were simulated to assess the performance of 3 methods for constructing confidence intervals; the formula based method for the nested ANOVA, the percentile bootstrap and the bootstrap-t methods. RESULTS: The performance of the evaluated methods for constructing a CI varied, both dependent on the size of the CVI and the type of distributions. The bootstrap-t CI have good and stable performance for the models evaluated, while the formula based are more distribution dependent. The percentile bootstrap performs poorly. CONCLUSION: CI is an essential part of estimation of the within-person biological variation. Good coverage probability and non-cover probabilities for CI are achievable by using the bootstrap-t combined with CV-ANOVA. Supplemental R-code is provided online.

Estimates of Within-Subject Biological Variation of Protein C, Antithrombin, Protein S Free, Protein S Activity, and Activated Protein C Resistance in Pregnant Women.

BACKGROUND: In pregnancy, interpretation of results from coagulation parameters can be difficult because of the procoagulant physiological changes. The aim of this study was to describe the course of 5 coagulation parameters (thrombophilia markers) in healthy pregnancies, and to estimate and compare the within-subject biological variation (CVI) of these parameters in healthy pregnant and nonpregnant women. METHODS: Blood samples were obtained every 4th week during pregnancy and 3 samples after delivery in 20 healthy women and every 4th week during 40 weeks in 19 healthy nonpregnant women. Protein C (PC), antithrombin (AT), protein S free (PS free), protein S activity (PS activity), and activated protein C resistance (with factor V-depleted plasma) (APCR) were analyzed. Before the calculation of CVI, results were transformed into multiples of the median (MoM) and natural logarithm of MoM (lnMoM) to adjust for the physiological changes during pregnancy. RESULTS: During pregnancy, PC results showed large variability, AT decreased slightly, and PS free and PS activity decreased significantly. Both activated partial thromboplastin time tests used to calculate APCR decreased, and the APCR ratio was constant. The CVI (lnMoM) in pregnancy were for PC 8.4%, for AT 3.8%, for PS free 11.5%, for PS activity 9.3%, and for APCR 0.5%, and similar to corresponding results in nonpregnant women. CONCLUSIONS: Transformation of coagulation parameters in healthy pregnancies to lnMoM is a tool to establish a kind of steady state. Although there is a physiological change in PC, AT, and PS free and PS activity during pregnancy, the CVI was comparable with the CVI of nonpregnant women.

The variation in high sensitive cardiac troponin concentration during haemodialysis treatment is not similar to the biological variation observed in stable end stage renal disease patients.

BACKGROUND: Cardiovascular mortality is high in end stage renal disease (ESRD). This study aimed to: (1) calculate within-week within- and between-subject biological variation (CVI and CVG) for hs-cTn in ESRD; (2) determine the magnitude of hs-cTn concentration changes during haemodialysis (HD) treatment; and (3) compare the CVI and CVG to the within and between-subject variation of cTn concentration changes during HD treatments (CVDIFF-I and CVDIFF-G). METHODS: Serum samples were collected from 20 patients before and after 10 consecutive HD treatments. cTn were measured using the hs-cTnT (Roche Diagnostics) and the hs-cTnI (Abbott Diagnostics). The CVA, CVI, CVG, CVDIFF-I and CVDIFF-G, were estimated using nested ANOVA. RESULTS: The within-week data showed hs-cTnT CVA, CVI and CVG of 1.6, 7.3 and 94.4%. Reference change values (RCV) were estimated to -18.7-23.0%. The Index of individuality (II) was 0.08. Corresponding values for hs-cTnI were 5.3, 13.2 and 142.4%, whilst the RCV was -32.5-48.2% and the II was 0.10. The mean concentration of cTn decreased by -6.4% (hs-cTnT) and -7.6% (hs-cTnI) during HD treatment. The CVDIFF-I and CVDIFF-G was 4.2 and 6.3% for hs-cTnT, and 10.7 and 9.4% for hs-cTnI. The RCVDIFF was -18.2-5.4% (hs-cTnT) and -39.0-23.8% (hs-cTnI), respectively, and the IIDIFF-values were 0.7 and 1.3. CONCLUSIONS: The CVI and CVG are similar to earlier findings. Mean hs-cTn concentrations decreased during HD. The within-subject hs-cTn variation during HD is similar to the between-subject variation, i.e. determining a cut-off value for hs-cTn changes during HD may be useful.

Sample collections from healthy volunteers for biological variation estimates' update: a new project undertaken by the Working Group on Biological Variation established by the European Federation of Clinical Chemistry and Laboratory Medicine.

BACKGROUND: Biological variation (BV) data have many fundamental applications in laboratory medicine. At the 1st Strategic Conference of the European Federation of Clinical Chemistry and Laboratory Medicine (EFLM) the reliability and limitations of current BV data were discussed. The EFLM Working Group on Biological Variation is working to increase the quality of BV data by developing a European project to establish a biobank of samples from healthy subjects to be used to produce high quality BV data. METHODS: The project involved six European laboratories (Milan, Italy; Bergen, Norway; Madrid, Spain; Padua, Italy; Istanbul, Turkey; Assen, The Netherlands). Blood samples were collected from 97 volunteers (44 men, aged 20-60 years; 43 women, aged 20-50 years; 10 women, aged 55-69 years). Initial subject inclusion required that participants completed an enrolment questionnaire to verify their health status. The volunteers provided blood specimens once per week for 10 weeks. A short questionnaire was completed and some laboratory tests were performed at each sampling consisting of blood collected under controlled conditions to provide serum, K2EDTA-plasma and citrated-plasma samples. RESULTS: Samples from six out of the 97 enroled subjects were discarded as a consequence of abnormal laboratory measurements. A biobank of 18,000 aliquots was established consisting of 120 aliquots of serum, 40 of EDTA-plasma, and 40 of citrated-plasma from each subject. The samples were stored at -80 °C. CONCLUSIONS: A biobank of well-characterised samples collected under controlled conditions has been established delivering a European resource to enable production of contemporary BV data.

Biological Variation: The Effect of Different Distributions on Estimated Within-Person Variation and Reference Change Values.

BACKGROUND: Good estimates of within-person biological variation, CVI, are essential for diagnosing and monitoring patients and for setting analytical performance specifications. The aim of the present study was to use computer simulations to evaluate the impact of various measurement distributions on different methods for estimating CVI and reference change value (RCV). METHOD: Data were simulated on the basis of 3 models for distributions of the within-person effect. We evaluated 3 different methods for estimating CVI: standard ANOVA, ln-ANOVA, and CV-ANOVA, and 3 different methods for calculating RCV: classic, ln-RCV, and a nonparametric method. We estimated CVI and RCV with the different methods and compared the results with the true values. RESULTS: The performance of the methods varied, depending on both the size of the CVI and the type of distributions. The CV-ANOVA model performed well for the estimation of CVI with all simulated data. The ln-RCV method performed best if data were ln-normal distributed or CVI was less than approximately 12%. The nonparametric RCV method performed well for all simulated data but was less precise. CONCLUSIONS: The CV-ANOVA model is recommended for both calculation of CVI and the step-by-step approach of checking for outliers and homogeneity in replicates and samples. The standard method for calculation of RCV should not be used when using CVs.

Feasibility of using self-reported patient data in a national diabetes register.

BACKGROUND: In order to improve recruitment of patients to the Norwegian diabetes register for adults, a questionnaire was designed to collect data directly from patients. The main aim of this study was to assess the agreement of questionnaire data with data reported to the Register from health care personnel during routine consultations. METHODS: Patient data were obtained by sending a questionnaire with 27 of the 41 Register variables to 3714 members of the Norwegian Diabetes Association. Questionnaire data were compared with data already in the Register. Paired t-tests, percentages of total agreement, percentages of "positive" answers and kappa coefficients (k) were used for comparing data. RESULTS: Of the 1645 replies (44.3 %), the Register already had data on 324 patients for comparison. Response rate for most variables was better from patients (ranging from 76-100 %) compared with health care professionals (33-100 %). For 17 of 25 assessable variables including diabetes duration, height, weight, HbA1c, drug treatment and several diabetes complications, agreement was substantial or better with kappa >0.60. Data on family history of premature heart disease (k-0.59), foot examination (k = 0.26), foot ulcer (k = 0.32) and arterial surgery (k = 0.24) seemed to be difficult to answer by patients, whereas data on physical activity and self-monitoring of glucose seemed to be better when reported by patients. CONCLUSIONS: Patient response rate was acceptable, and data had good concordance with data from health care professionals for most variables. However, registers using patient questionnaires should compare questionnaire data with data from professionals at regular intervals.

A checklist for critical appraisal of studies of biological variation.

Data on biological variation are used for many purposes in laboratory medicine but concern exists over the validity of the data reported in some studies. A critical appraisal checklist has been produced by a working group established by the European Federation of Clinical Chemistry and Laboratory Medicine (EFLM) to enable standardised assessment of existing and future publications of biological variation data. The checklist identifies key elements to be reported in studies to enable safe accurate and effective transport of biological variation data sets across healthcare systems. The checklist is mapped to the domains of a minimum data set required to enable this process.

Point-of-care urine albumin in general practice offices: effect of participation in an external quality assurance scheme.

BACKGROUND: The Norwegian Quality Improvement of Primary Care Laboratories (Noklus) offers external quality assurance (EQA) schemes (EQASs) for urine albumin (UA) annually. This study analyzed the EQA results to determine how the analytical quality of UA analysis in general practice (GP) offices developed between 1998 (n=473) and 2012 (n=1160). METHODS: Two EQA urine samples were distributed yearly to the participants by mail. The participants measured the UA of each sample and returned the results together with information about their instrument, the profession and number of employees at the office, frequency of internal quality control (IQC), and number of analyses per month. In the feedback report, they received an assessment of their analytical performance. RESULTS: The number of years that the GP office had participated in Noklus was inversely related to the percentage of "poor" results for quantitative but not semiquantitative instruments. The analytical quality improved for participants using quantitative instruments who received an initial assessment of "poor" and who subsequently changed their instrument. Participants using reagents that had expired or were within 3 months of the expiration date performed worse than those using reagents that were expiring in more than 3 months. CONCLUSIONS: Continuous participation in the Noklus program improved the performance of quantitative UA analyses at GP offices. This is probably in part attributable to the complete Noklus quality system, whereby in addition to participating in EQAS, participants are visited by laboratory consultants who examine their procedures and provide practical advice and education regarding the use of different instruments.

Benchmarking by HbA1c in a national diabetes quality register--does measurement bias matter?

BACKGROUND: Bias in HbA1c measurement could give a wrong impression of the standard of care when benchmarking diabetes care. The aim of this study was to evaluate how measurement bias in HbA1c results may influence the benchmarking process performed by a national diabetes register. METHODS: Using data from 2012 from the Norwegian Diabetes Register for Adults, we included HbA1c results from 3584 patients with type 1 diabetes attending 13 hospital clinics, and 1366 patients with type 2 diabetes attending 18 GP offices. Correction factors for HbA1c were obtained by comparing the results of the hospital laboratories'/GP offices' external quality assurance scheme with the target value from a reference method. RESULTS: Compared with the uncorrected yearly median HbA1c values for hospital clinics and GP offices, EQA corrected HbA1c values were within ±0.2% (2 mmol/mol) for all but one hospital clinic whose value was reduced by 0.4% (4 mmol/mol). Three hospital clinics reduced the proportion of patients with poor glycemic control, one by 9% and two by 4%. CONCLUSIONS: For most participants in our study, correcting for measurement bias had little effect on the yearly median HbA1c value or the percentage of patients achieving glycemic goals. However, at three hospital clinics correcting for measurement bias had an important effect on HbA1c benchmarking results especially with regard to percentages of patients achieving glycemic targets. The analytical quality of HbA1c should be taken into account when comparing benchmarking results.