loamr is an R package for performing agreement analysis on
continuous measurements made by multiple observers. The package provides
functions for making agreement plots and for calculating the estimate
and CI for the limits of agreement with the mean proposed by Christensen
et al. (2020).
loamr can be installed using the following command:
devtools::install_github("HaemAalborg/loamr")The package includes a function to simulate data from the two-way random effects model described in Christensen et al (2020):
sim <- simMD(subjects = 80, observers = 10, mu = 5)
head(sim)
#> # A tibble: 6 x 3
#> subject observer value
#> <int> <int> <dbl>
#> 1 1 1 4.99
#> 2 1 2 4.51
#> 3 1 3 6.52
#> 4 1 4 5.97
#> 5 1 5 4.47
#> 6 1 6 4.20Estimate and CI for the limits of agreements with the mean:
LOAM(sim)
#> Limits of agreement with the mean for multiple observers
#>
#> The data has 800 observations from 80 individuals by 10 observers with 1 measurements
#>
#> LOAM: +/- 1.438 (1.363, 1.616)
#>
#> sigmaA: 0.264 (0.190, 0.337)
#> sigmaB: 0.235 (0.113, 0.357)
#> sigmaE: 0.737 (0.700, 0.777)
#> ICC(A,1): 0.104 (0.057, 0.170)
#>
#> Coverage probability for the above CIs: 95%The S3 class includes a generic plotting function made with ggplot2
for making an agreement plot with indication of estimate and CI for the
limits of agreement with the mean:
plot(LOAM(sim))
Elements of the plot is easily changed using functionalities from
ggplot2. For example, changing the title:
plot(LOAM(sim)) + labs(title = "Simulated Data")
- Christensen, H. S., Borgbjerg, J., Børty, L., and Bøgsted, M. (2020) “On Jones et al.’s method for extending Bland-Altman plots to limits of agreement with the mean for multiple observers”. BMC Medical Research Methodology. https://doi.org/10.1186/s12874-020-01182-w