Analyze for efficacy using a beta prior to compute the posterior probability that experimental is better than standard of care.
AnalyzeUsingBetaBinomial.RdIn this version, the analysis for efficacy is to assume a beta prior to compute the posterior probability that experimental is better than standard of care Futility is based on a low posterior probability, eg it is unlikely that experimental is better than standard of care
Arguments
- SimData
Data frame which consists of data generated in current simulation.
- DesignParam
Input Parameters which user may need to compute test statistic and perform test. User should access the variables using names, for example, DesignParam$Alpha, and not order.
- LookInfo
List Input Parameters related to multiple looks which user may need to compute test statistic and perform test. User should access the variables using names, for example LookInfo$NumLooks and not order. Other important variables in group sequential designs are: LookInfo$NumLooks An integer value with the number of looks in the study LookInfo$CurrLookIndex An integer value with the current index look, starting from 1 LookInfo$CumEvents A vector of length LookInfo$NumLooks that contains the number of events at the look.
- UserParam
A list of user defined parameters in East or Solara. UserParam must be supplied and contain the following named elements:
- UserParam$dAlphaCtrl
Prior alpha parameter for control treatment. Equivalent to the prior number of treatment successes.
- UserParam$dBetaCtrl
Prior beta parameter for control treatment. Equivalent to the prior number of treatment failures.
- UserParam$dAlphaExp
Prior alpha parameter for experimental treatment. Equivalent to the prior number of treatment successes.
- UserParam$dBetaExp
Prior beta parameter for experimental treatment. Equivalent to the prior number of treatment failures.
- UserParam$dUpperCutoffEfficacy
A value (0,1) that specifies the upper cutoff for the efficacy check. Above this value will declare efficacy
- UserParam$dLowerCutoffForFutility
A value (0,1) that specified the lower cutoff for the futility check. Below this value will declare futility.
If user variables are not specified then a Beta( 1, 1 ) prior is utilized for both standard of care and experimental.
Value
The function must return a list in the return statement of the function. The information below lists elements of the list, if the element is required or optional and a description of the return values if needed.
- Decision
Required value. Integer Value with the following meaning:
- Decision = 0
when No boundary, futility or efficacy is crossed
- Decision = 1
when the Lower Efficacy Boundary Crossed
- Decision = 2
when the Upper Efficacy Boundary Crossed
- Decision = 3
when the Futility Boundary Crossed
- Decision = 4
when the Equivalence Boundary Crossed
- ErrorCode
Optional integer value
- ErrorCode = 0
No Error
- ErrorCode > 0
Non fatal error, current simulation is aborted but the next simulations will run
- ErrorCode < 0
Fatal error, no further simulation will be attempted
- Delta
Estimated different between experimental and standard of care
Note
When using simulation to obtain the frequentist Operating Characteristic (OC) of a Bayesian design, you should set dLowerCutoffForFutility = 0 when simulating under the null case in order to obtain the false-positive rate of the non-binding futility rule. When you set dLowerCutoffForFutility > 0, simulation will provide the OC of the binding futility rule because the rule is ALWAYS followed.
Helpful Hints: There is often info that East sends to R that are not shown in a given example. It can be very helpful to save the input objects and then load them into your R session and inspect them. This can be done with the following R code in your function.
saveRDS( SimData, "SimData.Rds") saveRDS( DesignParam, "DesignParam.Rds" ) saveRDS( LookInfo, "LookInfo.Rds" )
The above code will save each of the input objects to a file so they may be examined within R.