I am wondering if it would be possible to add an additional option (e.g, userDefined) for the argument effectMeasure in getSimulationEnrichmentMeans/Rates/Survival, to allow a user defined function based on effect measures other than effectEstimate and testStatistic for selectPopulationsFunction. Examples for such alternatives are population-specific predictive probabilities and subgroup-specific posterior probabilities[^1]. This might be implemented by providing additional arguments for the customized function selectPopulationsFunction to facilitate the calculation for both type of probabilities.
Just a thought for consideration: it may be convenient to borrow some arguments from calcSubjectsFunction (customized function for sample size re-calculation), but to provide for populations as well as subgroups. To be specific, we may need
thetaH1, overallEffects, and stDevH1 for continuous endpoints.
piTreatmentH1, piControlH1, overallRatesTreatment, and overallRatesControl for binary endpoints.
thetaH1 and overallEffects for survival endpoints.
[^1]: Brannath, W., Zuber, E., Branson, M., Bretz, F., Gallo, P., Posch, M., and Racine-Poon, A. (2009), “Confirmatory adaptive designs with Bayesian decision tools for a targeted therapy in oncology,” Statistics in Medicine, 28, 1445–1463.
fpahlke
commented
5 months ago
Originally opened by xinzhn
I am wondering if it would be possible to add an additional option (e.g,
userDefined) for the argumenteffectMeasureingetSimulationEnrichmentMeans/Rates/Survival, to allow a user defined function based on effect measures other thaneffectEstimateandtestStatisticforselectPopulationsFunction. Examples for such alternatives are population-specific predictive probabilities and subgroup-specific posterior probabilities[^1]. This might be implemented by providing additional arguments for the customized functionselectPopulationsFunctionto facilitate the calculation for both type of probabilities.Just a thought for consideration: it may be convenient to borrow some arguments from
calcSubjectsFunction(customized function for sample size re-calculation), but to provide for populations as well as subgroups. To be specific, we may need[^1]: Brannath, W., Zuber, E., Branson, M., Bretz, F., Gallo, P., Posch, M., and Racine-Poon, A. (2009), “Confirmatory adaptive designs with Bayesian decision tools for a targeted therapy in oncology,” Statistics in Medicine, 28, 1445–1463.