Package 'ega'

Clarke and Parkes (Consensus) error grid analysis

Description

Clarke and Parkes (Consensus) error grid analysis

---

Assign Clarke error grid zones to paired glucose values

Description

referenceVals and testVals are assumed to contain paired glucose values from a reference method and a test method, respectively. unit contains info on the unit of measurement. Two options exist: "gram" for mg/dL and "mol" for mmol/l with "gram" applied by default. The discrepancy between the two values is used to place the pair into a Clarke error grid zone according to the criteria described in the original paper by Clarke et. al. (see reference below).

Usage

getClarkeZones(referenceVals, testVals, unit = "gram")

Arguments

referenceVals	A vector of glucose values obtained via the reference method.
testVals	A vector of glucose values obtained via a non-reference method (e.g. a new meter). The values in this vector are paired with those in referenceVals, so the length should be the same.
unit	A string specifying the units of measurement. This should be either "gram" (the default) for mg/dl or "mol" for mmol/l.

Value

A character vector is returned, with each element being one of "A", "B", "C", "D", or "E".

References

Clarke, W. L., D. Cox, L. A. Gonder-Frederick, W. Carter, and S. L. Pohl. "Evaluating Clinical Accuracy of Systems for Self-Monitoring of Blood Glucose." Diabetes Care 10, no. 5 (September 1, 1987): 622-28.

Examples

zones <- getClarkeZones (glucose_data$ref / 18, glucose_data$test / 18,
unit="mol")

# counts
table(zones)

# percentages
round (table (zones) / length (zones) * 100, digits=2)

---

Assign Parkes (Consensus) error grid zones to paired glucose values

Description

referenceVals and testVals are assumed to contain paired glucose values from a reference method and a test method, respectively. The discrepancy between the two values, as well as the type of error grid desired (Type 1 or Type 2 diabetes), is used to place the pair into a Parkes (Consensus) error grid zone, according to the criteria described in the second reference below. unit contains info on the unit of measurement. Two options exist: "gram" for mg/dL and "mol" for mmol/l with "gram" applied by default.

Usage

getParkesZones(referenceVals, testVals, type = 1, unit = "gram")

Arguments

referenceVals	A vector of glucose values obtained via the reference method.
testVals	A vector of glucose values obtained via a non-reference method (e.g. a new meter). The values in this vector are paired with those in referenceVals, so the length should be the same.
type	An integer (1 or 2) specifying whether to obtain zones for Type 1 or Type 2 diabetes. Defaults to 1.
unit	A string specifying the units of measurement. This should be either "gram" (the default) for mg/dl or "mol" for mmol/l.

Value

A character vector is returned, with each element being one of "A", "B", "C", "D", or "E".

References

Parkes, J. L., S. L. Slatin, S. Pardo, and B.H. Ginsberg. "A New Consensus Error Grid to Evaluate the Clinical Significance of Inaccuracies in the Measurement of Blood Glucose." Diabetes Care 23, no. 8 (August 2000): 1143-48

Pfutzner, Andreas, David C. Klonoff, Scott Pardo, and Joan L. Parkes. "Technical Aspects of the Parkes Error Grid." Journal of Diabetes Science and Technology 7, no. 5 (September 2013): 1275-81

Examples

zones <- getParkesZones(glucose_data$ref, glucose_data$test)

# counts
table(zones)

# percentages
round (table (zones) / length (zones) * 100, digits=2)

---

5072 paired reference and test glucose values.

Description

A dataset containing 5072 paired reference method and test method glucose values (in mg/dL).

Usage

glucose_data

Format

A data frame with 5072 rows and 2 variables:

ref

Reference method glucose value, in mg/dL

test

Test method glucose value, in mg/dL

Source

The data is from a modified clinical dataset.

---

Plot a Clarke Error Grid

Description

The function uses ggplot to draw the Clarke error grid lines according to the criteria described in the original publication by Clarke et. al. (see reference below). If zones have not already been assigned via the zones parameter, the function getClarkeZones is called first. The values in referenceVals and testVals are then superimposed as a scatter plot. Some basic plot parameters can be specified as arguments, but the return value can also be stored and modified further before plotting (see examples and vignette).

Usage

plotClarkeGrid(referenceVals, testVals, title = "Clarke Error Grid",
  xlab = "", ylab = "", linesize = 0.5, linetype = "solid",
  linecolor = "black", linealpha = 0.6, pointsize = 2, pointalpha = 1,
  zones = NA, unit = "gram")

Arguments

referenceVals	A vector of glucose values obtained via the reference method.
testVals	A vector of glucose values obtained via a non-reference method (e.g. a new meter). The values in this vector are paired with those in referenceVals, so the length should be the same.
title	The main plot title. Defaults to "Clarke Error Grid".
xlab	The x-axis label. Defaults to "Reference Glucose Concentration (mg/dL)".
ylab	The y-axis label. Defaults to "Test Glucose Concentration (mg/dL)".
linesize	The size to be used when drawing the zone lines. The acceptable values are the same as for geom_segment. The default is 0.5.
linetype	The type of line to be used when drawing the zone lines. The acceptable values are the same as for geom_segment. The default is "solid".
linecolor	The color of the zone lines. The acceptable values are the same as for geom_segment. The default is "black".
linealpha	The alpha (transparency) level to be used when drawing the zone lines. The acceptable values are the same as for geom_segment. The default is 0.6.
pointsize	The size to be used when plotting the glucose data points. The acceptable values are the same as for geom_point. The default is 2.
pointalpha	The alpha (transparency) level to be used when plotting the glucose data points. The acceptable values are the same as for geom_point. The default is 1.
zones	An optional character vector specifying the Clarke zones for each paired value. If this is not supplied, getClarkeZones will be called to generate zone labels.
unit	A string specifying the units of measurement. This should be either "gram" (the default) for mg/dl or "mol" for mmol/l.

Value

A ggplot object is returned. If the return value is not assigned, a plot is drawn.

References

Clarke, W. L., D. Cox, L. A. Gonder-Frederick, W. Carter, and S. L. Pohl. "Evaluating Clinical Accuracy of Systems for Self-Monitoring of Blood Glucose." Diabetes Care 10, no. 5 (September 1, 1987): 622-28.

See Also

getClarkeZones ggplot

Examples

library(ggplot2)

# default
plotClarkeGrid(glucose_data$ref, glucose_data$test)

# with options
plotClarkeGrid(glucose_data$ref, glucose_data$test,
              pointsize=1.5,
              pointalpha=0.6,
              linetype="dashed")

# store return value and modify
ceg <- plotClarkeGrid(glucose_data$ref, glucose_data$test)

ceg + theme_gray() +
   theme(plot.title = element_text(size = rel(2), colour = "blue"))

---

Plot a Parkes (Consensus) Error Grid

Description

The function uses ggplot to draw the Parkes (consensus) error grid lines according to the criteria described in the publications listed in the References section (see below). If zones have not already been assigned via the zones parameter, the function getParkesZones is called first. The values in referenceVals and testVals are then superimposed as a scatter plot. Some basic plot parameters can be specified as arguments, but the return value can also be stored and modified further before plotting (see examples and vignette).

Usage

plotParkesGrid(referenceVals, testVals, type = 1, title = "", xlab = "",
  ylab = "", linesize = 0.5, linetype = "solid", linecolor = "black",
  linealpha = 0.6, pointsize = 2, pointalpha = 1, zones = NA,
  unit = "gram")

Arguments

referenceVals	A vector of glucose values obtained via the reference method.
testVals	A vector of glucose values obtained via a non-reference method (e.g. a new meter). The values in this vector are paired with those in referenceVals, so the length should be the same.
type	An integer (1 or 2) specifying whether to plot the grid for Type 1 or Type 2 diabetes. Defaults to 1.
title	The main plot title. Defaults to "Parkes (Consensus) Error Grid for Type [type] Diabetes".
xlab	The x-axis label. Defaults to "Reference Glucose Concentration (mg/dL)".
ylab	The y-axis label. Defaults to "Test Glucose Concentration (mg/dL)".
linesize	The size to be used when drawing the zone lines. The acceptable values are the same as for geom_segment. The default is 0.5.
linetype	The type of line to be used when drawing the zone lines. The acceptable values are the same as for geom_segment. The default is "solid".
linecolor	The color of the zone lines. The acceptable values are the same as for geom_segment. The default is "black".
linealpha	The alpha (transparency) level to be used when drawing the zone lines. The acceptable values are the same as for geom_segment. The default is 0.6.
pointsize	The size to be used when plotting the glucose data points. The acceptable values are the same as for geom_point. The default is 2.
pointalpha	The alpha (transparency) level to be used when plotting the glucose data points. The acceptable values are the same as for geom_point. The default is 1.
zones	An optional character vector specifying the Clarke zones for each paired value. If this is not supplied, getClarkeZones will be called to generate zone labels.
unit	A string specifying the units of measurement. This should be either "gram" (the default) for mg/dl or "mol" for mmol/l.

Value

A ggplot object is returned. If the return value is not assigned, a plot is drawn.

References

Parkes, J. L., S. L. Slatin, S. Pardo, and B.H. Ginsberg. "A New Consensus Error Grid to Evaluate the Clinical Significance of Inaccuracies in the Measurement of Blood Glucose." Diabetes Care 23, no. 8 (August 2000): 1143-48

Pfutzner, Andreas, David C. Klonoff, Scott Pardo, and Joan L. Parkes. "Technical Aspects of the Parkes Error Grid." Journal of Diabetes Science and Technology 7, no. 5 (September 2013): 1275-81

See Also

getParkesZones ggplot

Examples

library(ggplot2)

# default
plotParkesGrid(glucose_data$ref, glucose_data$test)

# with options
plotParkesGrid(glucose_data$ref, glucose_data$test,
              pointsize=2,
              pointalpha=0.5,
              linesize=2,
              linealpha=0.3,
              linetype="dotdash")

# store return value and modify
peg <- plotParkesGrid(glucose_data$ref, glucose_data$test, type=2)

peg + theme_gray() +
   theme(plot.title = element_text(size = rel(2), colour = "red"))

---