ContMon_getbound: Get boundaries for continuous trial monitoring

In IDDI-BE/PhII_Bayes: Bayesian tools for PhII clinical trial design

Description

This function calculates boundaries for every sample size

Usage

ContMon_getbound(
  SS,
  target,
  Beta_dis,
  P_target = NULL,
  maxtox = NULL,
  prt1 = 0,
  tox_ass = seq(0.1, 0.5, 0.05),
  sim = 0,
  nsim,
  out
)

Arguments

SS	total sample size, e.g. calculated for primary efficacy endpoint
target	target toxicity rate
Beta_dis	two parameters Beta(alpha,beta) of the prior Beta distribution
P_target	P(toxicity rate>target|data)>=P_target as decision threshold
maxtox	maximum allowable observed toxicity rate, where boundary not crossed
prt1	value for number of patients in starting cohort where alternative safety rule is followed, e.g. 3+3
tox_ass	vector of rates for calculating Bayesian probabilities P(toxicity rate<=tox_ass[]|boundary crossed) and frequentist operating characteristics
sim	binary indicator: simulations for frequentist operating characteristics (0=NO; 1=YES)
nsim	number of simulated datasets
out	path for storing csv output files

Details

Two parameters should be specified: either 'target' and 'maxtox' either 'target' and 'P_target'. If the Bayesian posterior probability, given the data is higher than 'P_target', i.e. P(toxicity rate>target|c+1,n)>=P_target, the boundary is crossed. For every 'n', the boundary 'c' is given, which corresponds with the maximum allowed toxicity. So the boundary is crossed at 'c+1'

P_Bayes= Bayesian posterior probability P(toxicity rate>target|c,n)

P_obs= c/n or 'allowable proportion': if number of events <=c then continuation, if c+1: boundary crossed

'Bayes_result' gives the Bayesian posterior probability P(toxicity rate<=target|c+1,n), so given that this boundary 'c+1' is exceeded, for assumptions, specified by the 'tox_ass' parameter

Prob(X>n|n,tox_true=...) gives the Prob(X>=c+1|different assumptions ('toxass'))

Cum Prob(...) gives the cumulative probability of at least 1x/2x/3x and 4x crossing the boundary by n patients

Examples

## Not run: 
ContMon_getbound(SS=19,target=0.2,Beta_dis=c(1,1),P_target=NULL,maxtox=0.33,
    tox_ass=seq(0.10,0.5,0.05),sim=1,nsim=100,prt1=6,out=NULL)

## End(Not run)

IDDI-BE/PhII_Bayes documentation built on May 19, 2021, 3:04 p.m.