Set bcea() Parameters: Constructor and Setters"
In BCEA: Bayesian Cost Effectiveness Analysis

knitr::opts_chunk$set(
collapse = TRUE,
comment = "#>",
fig.width = 6,
fig.height = 4
)

library(BCEA)

There are several arguments passed to bcea() to specify the form of the analysis. These are

bcea(eff, cost,
     ref = 1,
     interventions = NULL,
     .comparison = NULL,
     Kmax = 50000,
     wtp = NULL,
     plot = FALSE)

Those of interest here are:

ref is the reference intervention group to compare against the other groups.
.comparisons are the groups to compare against ref. The default is all of the non-ref groups. This is a new argument in the latest release of BCEA to make it more flexible and consistent with other functions. A preceding dot is used to keep it back-compatible with previous versions of BCEA. Argument c is partially matched with both c and comparison otherwise throwing an error.
Kmax is the maximum value of the willingness-to-pay to calculate statistics for.

During an analysis we may want to explore changing some of these parameters and keeping all of the others the same. We can do with with package setter functions.

Changing Reference Group

Load cost-effectiveness data.

data(Vaccine)

We first create bcea object using the constructor function for 2 different reference groups.

he_ref1 <- bcea(eff, cost,
                ref = 1,           
                interventions = treats,
                Kmax = 50000)  
str(he_ref1)

ceplane.plot(he_ref1)

he_ref2 <- bcea(eff, cost,
                ref = 2,           
                interventions = treats,
                Kmax = 50000)  

str(he_ref2[c("n_comparators", "ICER", "ref", "comp")])

Alternatively, we can do the same by modifying the first output.

setReferenceGroup(he_ref1) <- 2

str(he_ref1[c("n_comparators", "ICER", "ref", "comp")])

Changing Kmax

In the same way as above we can change Kmax in 2 equivalent ways.

he_Kmax1 <- bcea(eff, cost,
                 ref = 1,           
                 interventions = treats,
                 Kmax = 50000)  

str(he_Kmax1[c("n_comparators", "ICER", "ref", "comp", "Kmax")])

he_Kmax2 <- bcea(eff, cost,
                 ref = 2,           
                 interventions = treats,
                 Kmax = 2000)  

str(he_Kmax2[c("n_comparators", "ICER", "ref", "comp", "Kmax")])

setKmax(he_Kmax1) <- 2000

str(he_Kmax1[c("n_comparators", "ICER", "ref", "comp", "Kmax")])

Change Comparison Groups

Lets load some data with more than two groups.

data(Smoking)

Defaults is all other groups which in this case is 2, 3 and 4.

he_comp234 <- bcea(eff, cost,
                   ref = 1,           
                   interventions = treats,
                   Kmax = 50000)  

str(he_comp234[c("n_comparators", "ICER", "ref", "comp")])

ceplane.plot(he_comp234, wtp = 2000)

Let us compare against only groups 2.

he_comp2 <- bcea(eff, cost,
                 ref = 1,
                 .comparison = 2,
                 interventions = treats,
                 Kmax = 2000)  

str(he_comp2[c("n_comparators", "ICER", "ref", "comp")])

ceplane.plot(he_comp2, wtp = 2000)

We can achieve the same thing using the appropriate setter.

setComparisons(he_comp234) <- 2

str(he_comp234[c("n_comparators", "ICER", "ref", "comp")])

ceplane.plot(he_comp234, wtp = 2000)

We can select multiple comparison groups too. Let us compare against only groups 2 and 4.

he_comp24 <- bcea(eff, cost,
                  ref = 1,
                  .comparison = c(2,4),
                  interventions = treats,
                  Kmax = 2000)  
str(he_comp24[c("n_comparators", "ICER", "ref", "comp")])

ceplane.plot(he_comp24, wtp = 2000)

setComparisons(he_comp234) <- c(2,4)

str(he_comp234[c("n_comparators", "ICER", "ref", "comp")])

ceplane.plot(he_comp24, wtp = 2000)

Further, a bcea object with all comparison groups can be passed to other functions such as ceplane.plot and ceac.plot with a comparison argument, which will do the modifications using these functions internally instead.

ceplane.plot(he_comp234, comparison = 2, wtp = 2000)
ceplane.plot(he_comp234, comparison = c(2,4), wtp = 2000)

# create output docs
# rmarkdown::render(input = "vignettes/Set_bcea_parameters.Rmd", output_format = "pdf_document", output_dir = "vignettes")
# rmarkdown::render(input = "vignettes/Set_bcea_parameters.Rmd", output_format = "html_document", output_dir = "vignettes")