model_gemtc: Bayesian net-meta model for MD, HR, and RR This function is...

In meta2020/net.meta: The Functions for Network Meta Analysis

Description

Bayesian net-meta model for MD, HR, and RR This function is the warpper of mtc.network, mtc.model, mtc.run, and relative.effect from gemtc package. R packgae gemtc and rjags are required.

Usage

model_gemtc(
  long.data,
  id.treatments,
  reference,
  outcome = c("MD", "HR", "RR"),
  mtc.n.adapt,
  mtc.n.iter,
  mtc.thin
)

Arguments

long.data	data.frame to be analyzed should be formatted in long format
id.treatments	data.frame to specify the id and treatments
reference	the referential id in the net-meta
outcome	the outcome should be MD-mean difference, HR-hazard ratio, and RR-risk ratio
mtc.n.adapt	the number of adaptation (or tuning, burn-in) iterations, default is 5000, which means to discard 1-5000 of the iterations.
mtc.n.iter	the number of simulation iteration, default is 10000, which means to perform 10000 simulations
mtc.thin	default is 20, which means to extract 20th value; details in mtc.run from gemtc package

Value

Summary list of the results

Examples

LDT1 <- read.csv(system.file("extdata", "HR_SH_D.csv", package = "net.meta"))
trt1 <- read.table(system.file("extdata", "HR_SH_D.txt", package = "net.meta"),
    header=TRUE,quote = '"', stringsAsFactors=FALSE)

trt1$description <- factor(trt1$description, trt1$description)
LDT1$study <- factor(LDT1$study, unique(LDT1$study))

set.seed(1)
bmt1 <- model_gemtc(
long.data=LDT1,
id.treatments=trt1,
reference="A",
outcome="HR",
mtc.n.adapt = 500, mtc.n.iter = 1000, mtc.thin = 20)

meta2020/net.meta documentation built on March 30, 2021, 7:31 p.m.