Bland-Altman Limits of Agreement

- The difference between two variables are plotted against their averages.

- On x axis, the averages of the two measurements are displayed. On y axis, the difference between two variables are displayed.

- This is a method for graphical comparison of two technicals or measurements or two variables of interest. If the differences are small and mean of the differences is near zero, the test is considered reliable

Example in R

Untitled.utf8

Generate Bland-Altman Limits of Agreement Plot using the same distribution. Observe that the mean of difference is closer to 0, with no visible bias.

set.seed(1234)
library("blandr")
measurement1 <- rnorm(100)
measurement2 <- rnorm(100)
statistics.results <- blandr.statistics( measurement1 , measurement2 )
blandr.plot.ggplot( statistics.results )
## Warning: Use of `plot.data$x.axis` is discouraged. Use `x.axis` instead.
## Warning: Use of `plot.data$y.axis` is discouraged. Use `y.axis` instead.

Generate Bland-Altman Limits of Agreement Plot using distribution with different means and add bias. Observe that the mean of difference is closer to -5.

measurement1 <- rnorm(100)
measurement2 <- rnorm(100,4,1) + rt(100,4,1)
statistics.results <- blandr.statistics( measurement1 , measurement2 )
blandr.plot.ggplot( statistics.results )
## Warning: Use of `plot.data$x.axis` is discouraged. Use `x.axis` instead.
## Warning: Use of `plot.data$y.axis` is discouraged. Use `y.axis` instead.

Live Example: Try this yourself!

References:

Hopkins, W. G. (2000). Measures of Reliability in Sports Medicine and Science. Sports Medicine, 30(1), 1-15. http://dx.doi.org/10.2165/00007256-200030010-00001