create_utility_function: Create Utility Function For The Normal Model

Description Usage Arguments Details Value Author(s)

Description

Create a utility function to be used together with a normal simulation model object created using create.normal.model or create.normal.model.from.file.

Usage

1
2
create.utility.function(model, n.min, sig.level, safety.max, cE, cS, p,
fixed.cost, cost.per.sample)

Arguments

model

A model object created using create.normal.model or create.normal.model.from.file.

n.min

The minimum group sample size for each phase III group trial required by the regulatory authority in order to consider market approval.

sig.level

The significance level used by the regulatory authority when performing the one-sided hypothesis tests for acceptable efficacy and safety levels in the phase III trial.

safety.max

A parameter defining the maximum safety threshold in the significance test for an acceptable safety level.

cE

A constant defining the utility gain per unit of efficacy.

cS

A constant defining the utility gain per unit of safety. The absolute value of this number defines the utility loss, and hence cS should typically be less than or equal to zero.

p

A number between 0 and 1 which weighs the relative contribution of the observed responses and the true population means to the utility upon regulatory approval. A value of 1 corresponds to no contribution made by the population means.

fixed.cost

The fixed cost of performing the phase III trials.

cost.per.sample

The cost per observation in the phase III trials.

Details

The utility function has the form:

RA.decision * gain - trial.cost

where

gain = p * (cE * mean(YE) + cS * mean(YS)) + (1 - p) * (cE * mean(muE) + cS * mean(muS))

trail.cost = fixed.cost + cost.per.sample * k.III * n.III

Value

An R function to be used together with model when calling eval.on.grid, fit.gpr, fit.loess and optimise.eu.

Author(s)

Sebastian Jobj<f6>rnsson jobjorns@chalmers.se


bdpopt documentation built on May 2, 2019, 9:18 a.m.