futilityOC: Calculates the operating characteristics of futility rules

In alasorous/futilityOC: Evaluation of the operating characteristics of futility stopping rules

Description

Calculates the operating characteristics of futility rules

Usage

futilityOC(
  txtype = "norm",
  pow.type = "PREDICTIVE",
  p.stop = 0.15,
  ind = 1,
  epts.r = 1,
  inF = 0.5,
  H1 = TV,
  TV = H1,
  benef = NA,
  ctr.r = 0.5,
  NB.k = 0.5,
  over.dp = 1.5,
  alpha = 0.1,
  power = 0.8,
  n = 0,
  mat = 1,
  sd = 1,
  LRV = "",
  delta = "",
  ran.r = 1,
  pr.tv = 0.1,
  pr.lrv = 0.8,
  print = T
)

Arguments

txtype	The type of treatment (tx) effect "norm", HR","NB","PR","QP","RD", "RR". Default="norm" HR = hazard ratio; NB = Negative binomial; PR = Poisson regression QP = Quasi Poisson; RD = Risk difference; RR = risk ratio
pow.type	The type of power for futility rule "PREDICTIVE" or "CONDITIONAL" power. Default="PREDICTIVE"
p.stop	power boundary to stop for futility. Default = 0.15
ind	value 1, 2 or 3. 1 = power for statistical significance, 2 = power to meet Go criteria, 3 = power to meet better than NoGo boundary. Default = 1
epts.r	Assumed correlation between the endpoints used at the futility analysis and the final. Default = 1
inF	information fraction at interim. Default = 0.5
H1	The treatment effect size assumed under the alternative hypothesis. Default = TV
TV	Target value for Go No Go decision. Default = H1
benef	A logical value. TRUE indicates a larger effect is beneficial. FALSE indicates a small effect is beneficial. Default values depend on the type of treatment effect: = TRUE for effect types of norm and RD = FALSE for effect types of HR, NB, PR, QP and RR
ctr.r	Assumed event rate for the control arm, which is necessary to be specified for effect types of NB, PR, QP, RD and RR. Default = 0.5
NB.k	The shape parameter of negative binomial distribution, which is necessary when effect type is NB. Default = 0.5.
over.dp	The over-dispersion estimate for treatment effect type QP. Default = 1.5
alpha	Alpha value that is used for determining sample size n, if n is not specified. Default = 0.1 one-sided.
power	Statistical power that is used for determining sample size if n is not specified. Default = 0.8.
n	The sample size of the trial. When it is not specified (=0), its value is estimated. Note: n represents the number of events when treatment effect is in HR with time-to-event endpoint and the default mat = 1 is not altered.
mat	Data maturity rate of time-to-event endpoint. Default = 1
sd	The standard deviation specified for endpoints following a normal distribution. Default = 1
LRV	The lower reference value specified for its use in Go/NoGo decision rules. Default = 2/3 * TV
delta	A vector of assumed/hypothesized treatment effects for which the operating characteristics of futility analysis are to be evaluated. Default = a vector of 4 with values = (TV, LRV, TV/4, 0).
ran.r	The randomization ratio. Default = 1
pr.tv	Standard GNG rules. Probability that the treatment effect is equal to or greater than TV. Default = 0.1
pr.lrv	Standard GNG rules. Probability that the treatment effect is equal to or greater than LRV. Default = 0.8
print	TRUE prints the output as well as returning dataframe

Value

dataframe with Design.specifications, Go.NoGo.parameters, Prob.stop.under.Hypotheses, Futility.stopping.boundary, Conditional.power.at.stopping.boundary, Predictive.power.at.stopping.boundary, Critical.values.for.Go.NoGo.based.on.Interim.data

Examples

  futilityOC(txtype = "RD", pow.type = "PREDICTIVE", p.stop = 0.2,
  ind = 1, inF = 0.5, TV = 0.2, # target difference
  benef = T, ctr.r = 0.12, alpha = 0.025, n = 72)

alasorous/futilityOC documentation built on Sept. 7, 2020, 8:55 a.m.