library(boot) # non-parametric bootstrap in MAIC and ML G-computation
library(copula) # simulating BC covariates from Gaussian copula
library(rstanarm) # fit outcome regression, draw outcomes in Bayesian G-computation
-library(ModStanR)Next, we load the data to use in the analysis. The data comes from a @@ -480,13 +480,13 @@
The {ModStanR} package has been written to be easy to
+
The {outstandR} package has been written to be easy to
use and essential consists of a single function,
-ModStanR(). This can be used to run all of the different
+outstandR(). This can be used to run all of the different
types of model, which we will call strategies. The first two
-arguments of ModStanR() are the individual and
+arguments of outstandR() are the individual and
aggregate-level data, respectively.
A strategy argument of ModStanR takes
+
A strategy argument of outstandR takes
functions called strategy_*(), where the wildcard
* is replaced by the name of the particular method
required, e.g. strategy_maic() for MAIC. Each specific
@@ -507,10 +507,10 @@
formula object passed as an argument to the strategy
function.
-
-The returned object is of class ModStanR.
The returned object is of class outstandR.
ModStanR_maic
+outstandR_maic
#> $contrasts
#> $contrasts$AB
#> [1] 0.5008915
@@ -549,11 +549,11 @@ MAIC
#> attr(,"CI")
#> [1] 0.95
#> attr(,"class")
-#> [1] "ModStanR" "list"
+#> [1] "outstandR" "list"We see that this is a list object with 3 parts, each containing
statistics between each pair of treatments. These are the mean
contrasts, variances and confidence intervals (CI), respectively. The
-default CI is for 95% but can be altered in ModStanR with
+default CI is for 95% but can be altered in outstandR with
the CI argument.
\[
y = X_3 + X_4 + \beta_t(X_1 - \bar{X_1}) + \beta_t(X_2 - \bar{X_2})
-\] However, ModStanR() knows how to handle this so
+\] However, outstandR() knows how to handle this so
we can simply pass the same (uncentred) formula as before.
ModStanR_stc <- ModStanR(AC.IPD, BC.ALD, strategy = strategy_stc(formula = lin_form))
-ModStanR_stc
+outstandR_stc <- outstandR(AC.IPD, BC.ALD, strategy = strategy_stc(formula = lin_form))
+outstandR_stc
#> $contrasts
#> $contrasts$AB
#> trt
@@ -619,7 +619,7 @@ STC
#> attr(,"CI")
#> [1] 0.95
#> attr(,"class")
-#> [1] "ModStanR" "list"
+#> [1] "outstandR" "list"For the last two approaches, we perform G-computation firstly with a frequentist MLE approach and then a Bayesian approach.
\[ \hat{\Delta}^{(2)}_{10} = g(\hat{\mu}_1) - g(\hat{\mu}_0) \]
- +ModStanR_gcomp_stan <- ModStanR(AC.IPD, BC.ALD, strategy = strategy_gcomp_stan(formula = lin_form))
+outstandR_gcomp_stan <- outstandR(AC.IPD, BC.ALD, strategy = strategy_gcomp_stan(formula = lin_form))
#>
#> SAMPLING FOR MODEL 'bernoulli' NOW (CHAIN 1).
#> Chain 1:
@@ -745,7 +745,7 @@ Bayesian G-computation with MCMC
#> Chain 2: 0.746 seconds (Sampling)
#> Chain 2: 1.422 seconds (Total)
#> Chain 2:
-