Understanding Error Grid Analysis

Cox, Daniel J.; Gönder-Frederick, Linda; Kovatchev, Boris; Julian, Diana M; Clarke, William L.

doi:10.2337/diacare.20.6.911

Cited by 57 publications

(31 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Error grid analysis was developed by the research group of Cox and Clarke to determine the clinical accuracy of glucose results, taking into account both the difference between the reference and the test method values, and the pertinence of the treatment decision resulting from the test method value (15,43). According to this analysis, only a single HemoCue measurement (0.2%) could potentially have led to a dangerous failure to detect and treat hypoglycemia, and five (1.0%) measurements could have resulted in benign estimation errors.…”

Section: Discussionmentioning

confidence: 99%

Comparison of the accuracy of the hemocue glucose analyzer with the Yellow Springs Instrument glucose oxidase analyzer, particularly in hypoglycemia

et al. 2005

View full text Add to dashboard Cite

Objective: We aimed to assess the accuracy of the HemoCue Beta-glucose analyzer (HemoCue) and its correlation with the Yellow Springs Instrument (YSI 2300 STAT; YSI) glucose oxidase analyzer, in particular for hypoglycemic values. Design and methods: Samples were taken from 24 volunteers during hyperinsulinemic glucose clamp studies. Glucose concentrations were determined immediately with the HemoCue in whole blood and with the YSI in plasma from the same sample. After correction for the difference between whole blood and plasma, the paired plasma glucose concentrations were analyzed with various statistical methods.Results: A total of 500 paired glucose values were obtained, 209 of which were in the hypoglycemic range. Mean^S.E. values were 4.85^0.004 mmol/l for the HemoCue (range 1.87 -16.17) and 4.81^0.004 mmol/l for the YSI (range 1.88 -15.00; P ¼ 0.80). In the hypoglycemic region, values were 3.26^0.004 mmol/l for the HemoCue (range 1.87 -5.17) and 3.22^0.003 mmol/l for the YSI (range 1.88 -4.20; P ¼ 0.59). Regression analyses were HemoCue ¼ 1.019(YSI) 2 0.0577 mmol/l, with r ¼ 0.9787 for all values; for hypoglycemic values the HemoCue ¼ 1.1169(YSI) 2 0.3393 mmol/l, with r ¼ 0.8798. Using Altman's residual plot, the difference was 0.03^0.0009 mmol/l, with 18 (3.6%) paired values outside the 95% limits of agreement (20.82 to 0.89 mmol/l). In the hypoglycemic range, the difference was 0.04^0.001 mmol/l, with six (2.9%) values outside the 95% limits of agreement (2 0.71 to 0.79 mmol/l). In error grid analysis, one value was in zone D (0.2%) and five values (1%) were in zone B; 98.8% were within zone A. Conclusions: Determination of glucose with the HemoCue system had very good correlation with the YSI system in a broad range of glycemia and also for hypoglycemic values. We believe that these methods can be used interchangeably for research and clinical purposes in adults.European Journal of Endocrinology 153 275-281

show abstract

Section: Discussionmentioning

confidence: 99%

Comparison of the accuracy of the hemocue glucose analyzer with the Yellow Springs Instrument glucose oxidase analyzer, particularly in hypoglycemia

et al. 2005

View full text Add to dashboard Cite

show abstract

“…Spearman's rank correlation coefficient was used to identify significant correlations between measurements, and Spearman's correlation coefficient (r) was obtained using GraphPad Prism 6 (GraphPad Software, Inc., San Diego, CA). PBGM data were evaluated using the clinical relevant error grid analysis (Clarke et al 1987, Cox et al 1997. The error grid compares values measured by reference glucose values (x-axis) versus the PBGM values (y-axis) in 5 zones associated with the following 5 risk levels: zone A, clinically accurate; zone B, altered clinical action but no or minimal effect on clinical outcome; zone C, altered clinical action with a likely effect on clinical outcome; zone D, altered clinical action with considerable medical risk; and zone E, altered clinical action with dangerous consequences (Johnson et al 2009).…”

Section: Methodsmentioning

confidence: 99%

Evaluation of portable blood glucose meters using canine and feline pooled blood samples

Mori¹,

Oda²,

Onozawa³

et al. 2016

Polish Journal of Veterinary Sciences

View full text Add to dashboard Cite

This study evaluated the accuracy and reproducibility of a human portable blood glucose meter (PBGM) for canine and feline whole blood. Reference plasma glucose values (RPGV) were concurrently measured using glucose oxidation methods. Fifteen healthy dogs and 6 healthy cats were used for blood sampling. Blood glucose concentrations and hematocrits were adjusted using pooled blood samples for our targeted values. A positive correlation between the PBGM and RPGV was found for both dogs (y = 0.877, x = -24.38, r = 0.9982, n = 73) and cats (y = 1.048, x = -27.06, r = 0.9984, n = 69). Acceptable results were obtained in error grid analysis between PBGM and RPGV in both dogs and cats; 100% of these results were within zones A and B. Following ISO recommendations, a PBGM is considered accurate if 95% of the measurements are within ± 15 mg/dl of the RPGV when the glucose concentration is <100 mg/dl and within ±15% when it is ≥100 mg/dl; however, small numbers of samples were observed inside the acceptable limits for both dogs (11%, 8 of 73 dogs) and cats (39%, 27 of 69 cats). Blood samples with high hematocrits induced lower whole blood glucose values measured by the PBGM than RPGV under hypoglycemic, normoglycemic, and hyperglycemic conditions in both dogs and cats. Therefore, this device is not clinically useful in dogs and cats. New PBGMs which automatically compensate for the hematocrit should be developed in veterinary practice.

show abstract

“…These and other weaknesses have been discussed in the literature (10,11). To address such concerns, we have assembled a new EG based on a survey of diabetes specialists.…”

Section: Figure 1-egs For Type 1 Diabetes (A) and Type 2 Diabetes (B)mentioning

confidence: 99%

“…Nevertheless, the Clarke EG can be criticized on the basis of the placement of its risk boundaries (10,11), some of which skip risk categories. It was created 15 years ago by a small number of clinicians, and advances in the field since then justify a re-examination of the subject.…”

mentioning

confidence: 99%

A new consensus error grid to evaluate the clinical significance of inaccuracies in the measurement of blood glucose.

et al. 2000

View full text Add to dashboard Cite

OBJECTIVE -The objectives of this study were 1) to construct new error grids (EGs) for blood glucose (BG) self-monitoring by using the expertise of a large panel of clinicians and 2) to use the new EGs to evaluate the accuracy of BG measurements made by patients. RESEARCH DESIGN AND METHODS -To construct newEGs for type 1 and type 2 diabetic patients, a total of 100 experts of diabetes were asked to assign any error in BG measurement to 1 of 5 risk categories. We used these EGs to evaluate the accuracy of self-monitoring of blood glucose (SMBG) levels in 152 diabetic patients. The SMBG data were used to compare the new type 1 diabetes EG with a traditional EG.RESULTS -Both the type 1 and type 2 diabetes EGs divide the risk plane into 8 concentric zones with no discontinuities. The new EGs are similar to each other, but they differ from the traditional EG in several significant ways. When used to evaluate a data set of measurements made by a sample of patients experienced in SMBG, the new type 1 diabetes EG rated 98.6% of their measurements as clinically acceptable, compared with 95% for the traditional EG. CONCLUSIONS -The consensus EGs furnish a new tool for evaluating errors in the mea E m e r g i n g T r e a t m e n t s a n d T e c h n o l o g i e s 1144DIABETES CARE, VOLUME 23, NUMBER 8, AUGUST 2000 New error grid for blood glucose 152 patients who routinely monitor their own BG. The study was performed in a diabetes clinic with patients using their own meters. The distribution of errors in our sample is considered in light of both EGs and the latest recommendations of the American Diabetes Association (ADA). RESEARCH DESIGN AND METHODS Consensus EGWe surveyed 100 physicians at the 1994 ADA Annual Meeting to construct an unbiased tool to analyze the clinical significance of SMBG measurement errors. All of the respondents were clinicians who treated diabetic patients. Pursuant to constructing an EG in the fashion of Clarke et al. (6), each doctor was asked to assign any plausible error in BG measurement to 1 of 5 risk categories. The risk categories, in order of increasing severity, were defined as follows: A: no effect on clinical action; B: altered clinical action or little or no effect on clinical outcome; C: altered clinical action-likely to effect clinical outcome; D: altered clinical action-could have significant medical risk; and E: altered clinical action-could have dangerous consequences.The above definitions were intended to correspond to the definitions of the risk zones in the Clarke EG while allowing the respondents maximal freedom to set their own boundaries. For example, zone A of the Clarke EG is defined as Ͻ20% deviation or having both reference and measured BGs Ͻ70 mg/dl. In addition, the UVA authors (6) stated that "values falling within this range are clinically accurate in that they would lead to clinically correct treatment decisions." Our definition of zone A asks each respondent to define his or her own range of "clinically accurate measurements," which is clarified as having "no effe...

show abstract

Understanding Error Grid Analysis

Cited by 57 publications

References 20 publications

Comparison of the accuracy of the hemocue glucose analyzer with the Yellow Springs Instrument glucose oxidase analyzer, particularly in hypoglycemia

Comparison of the accuracy of the hemocue glucose analyzer with the Yellow Springs Instrument glucose oxidase analyzer, particularly in hypoglycemia

Evaluation of portable blood glucose meters using canine and feline pooled blood samples

A new consensus error grid to evaluate the clinical significance of inaccuracies in the measurement of blood glucose.

Contact Info

Product

Resources

About