SMARTAR-tutorial

In SMARTAR: Sequential Multiple Assignment Randomized Trial and Adaptive Randomization

SMARTAR package is for primary data analysis for sequential multiple assignment randomization trial (SMART) and are calibration tools for clinical trial planning purposes. This is a simple illustration of SMARTAR package. It only contains five main functions which are seqmeans, atsmeans, smartest, smartsize and getncp. In addition, it also contains one dataset codiacs.

The use of these functions and dataset is:

load package and dataset

library(SMARTAR)
data(codiacs)

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

seqmeans

Exports treatment sequence, summarizes all the sequence-specific descriptive statistics and graphs, and provides design diagram of SMART.

seqmeans(data=codiacs ,family="gaussian",plot="d", digits = 2,xlab = "SMART design")
seqmeans(data=codiacs ,plot="s",color = "lightblue",xlab = "SEQ",family="gaussian")

atsmeans

Exports all the ATS embedded in SMART design and gives estimated strategy values and the variance-covariance matrix of estimated values.

atsmeans(data=codiacs,conf=TRUE, alpha=0.05,plot=TRUE,digits = 2,pch=18,xlab="Treatment sequence")
atsmeans(data=codiacs,conf=TRUE, alpha=0.05,digits = 2,pch=18,xlab="abc")

smartest

Exports results of statistical tests of comparing adaptive treatment strategies based on both global and pairwise tests.

smartest(data=codiacs,method="IPW",adjust="Bon")

getncp

Return the value of non-centralized parameter for the chi-square distribution.

getncp(df=5, alpha = 0.05, beta = 0.2, d = 1e-04, start = 5)

smartsize

Exports estimated strategy-specified means and their confidence interval, as well as the asymptotic variance-covariance matrix for these estimates.

smartsize(delta=0.0435,df=5,global=TRUE,alpha=0.05,beta=0.20)

SEQ <- 1:8
A1 <- c(rep(0,4),rep(1,4))
PI1 <- rep(0.5,8)
O2 <- rep(c(0,0,1,1),2)
P2 <- c(0.7,0.7,0.3,0.3,0.6,0.6,0.4,0.4)
A2 <- rep(c(0,1),4)
PI2 <- rep(0.5,8)
MEAN <- 1:8
SD <- rep(10,8)
SIMatrix <- as.data.frame(cbind(SEQ,A1,PI1,O2,P2,A2,PI2,MEAN,SD))

smartsize(SIMatrix,global=TRUE,alpha=0.05,beta=0.20)

SMARTAR documentation built on July 31, 2020, 1:06 a.m.