Three analysers (machines) for determination of HbA1c (glycosylated haemoglobin) were tested on samples from 38 individuals. Each had drawn a venous and capillary blood sample. These were analysed on five different days.
A data frame with 835 observations on the following 6 variables.
devType of machine used. A factor with levels
BR.V2, BR.VC and Tosoh.
typeType of
blood analysed (capillary or venous). A factor with levels Cap
Ven
itemPerson-id. A numeric vector
d.sampDay of sampling.
d.anaDay of laboratory analysis.
yThe measured value of HbA1c.
Bendix Carstensen, Steno Diabetes Center.
In the terminology of method comparison studies, methods is the
cross-classification of dev and type, and replicate is
d.ana. It may be of interest to look at the effect of time between
d.ana and d.samp, i.e. the time between sampling and analysis.
These data were analysed as example in: Carstensen: Comparing and predicting between several methods of measurement, Biostatistics 5, pp. 399–413, 2004.
data(hba1c)
str(hba1c)
#> 'data.frame': 835 obs. of 6 variables:
#> $ dev : Factor w/ 3 levels "BR.V2","BR.VC",..: 2 2 2 2 2 2 2 2 1 1 ...
#> $ type : Factor w/ 2 levels "Cap","Ven": 2 2 2 2 1 1 1 1 2 2 ...
#> $ item : num 12 12 12 12 12 12 12 12 12 12 ...
#> $ d.samp: num 1 1 1 1 1 1 1 1 1 1 ...
#> $ d.ana : num 2 3 4 5 2 3 4 5 2 3 ...
#> $ y : num 8.7 8.7 8.7 8.7 9.2 9 8.8 8.7 9.4 9.3 ...
hb1 <- with( hba1c,
Meth( meth = interaction(dev,type),
item = item,
repl = d.ana-d.samp,
y = y, print=TRUE ) )
#> #Replicates
#> Method 3 4 #Items #Obs: 835 Values: min med max
#> BR.V2.Cap 0 38 38 152 5.3 8.0 12.6
#> BR.VC.Cap 19 19 38 133 5.3 8.2 12.1
#> Tosoh.Cap 0 38 38 152 5.0 7.8 11.8
#> BR.V2.Ven 19 19 38 133 5.5 8.1 12.0
#> BR.VC.Ven 19 19 38 133 5.3 8.0 11.6
#> Tosoh.Ven 20 18 38 132 5.3 8.0 12.1