targetD: Compute the MLE (and its SE) of the dose achieving a specified target improvement from placebo.

Description

The MLE (se) of the dose required to achieve a targetted improvement from placebo. The fit can be from a 3- or 4- parameter Emax model or output from function emaxalt, or an object of class emaxsimobj. The Emax model is on the logit scale for binary data.

Usage

targetD (fit,
	target,
	modType=4,
	binary=FALSE)

Value

Returns a vector with two elements:

targetDose: The MLE of the dose achieving the target.
seTD: SE for target.dose

Arguments

fit: Output of nls fit to a 3- or 4-parameter Emax model. The order of the parameters in the fit must be (log(ed50),emax,e0) or (log(ed50),lambda,emax,e0). fit can also be a list with the first element the coefficient vector, and the second element the variance-covariance matrix. Alternatively, fit may be of class emaxalt or emaxsimobj, and the target dose is based on the fitted model.
target: Targetted change from placebo (positive or negative).
modType: Value is 3 or 4 for the 3 or 4-parameter Emax model output from nls with parameters in the order (ed50,emax,e0) or (ed50,lambda,emax,e0). modType is ignored if fit is from emaxalt or emaxsimobj.
binary: When TRUE, the fit is assumed to be for binary data on the logistic scale. target is input as a risk difference, and transformed internally. When the fit is of class emaxalt or emaxsimobj, the binary status is taken from the object and binary is ignored.

Author

Neal Thomas

Examples

Run this code

    if (FALSE) {

		## emaxsim changes the random number seed
    doselev<-c(0,5,25,50,100,250)
    n<-c(78,81,81,81,77,80)
    dose<-rep(doselev,n)

    ### population parameters for simulation
    e0<-2.465375 
    ed50<-67.481113 
    
   	dtarget<-100
		diftarget<-9.032497
		emax<-solveEmax(diftarget,dtarget,log(ed50),1,e0)

    sdy<-7.967897
    pop<-c(led50=log(ed50),emax=emax,e0=e0)    
    meanresp<-emaxfun(dose,pop)  
    y<-rnorm(sum(n),meanresp,sdy)
    nls.fit<-nls(y ~ e0 + (emax * dose)/(dose + exp(led50)), 
                              start = pop, control = nls.control(
                              maxiter = 100),trace=TRUE,na.action=na.omit)

    targetD(nls.fit,10,modType=3)

    ###
    ### apply targetD to an emaxsim object
    ###
    nsim<-50
    sdy<-25
    gen.parm<-FixedMean(n,doselev,emaxfun(doselev,pop),sdy)  
    D4 <- emaxsim(nsim,gen.parm,modType=4)
    summary(D4,testalph=0.05)

    out<-NULL
    for(i in 1:nsim){
        out<-rbind(out,targetD(D4[i],target=4))
    }
}
    # \dontshow{

		## emaxsim changes the random number seed
    doselev<-c(0,5,25,50,100,250)
    n<-c(78,81,81,81,77,80)
    dose<-rep(doselev,n)

    ### population parameters for simulation
    e0<-2.465375 
    ed50<-67.481113 
    
 		dtarget<-100
		diftarget<-9.032497
		emax<-solveEmax(diftarget,dtarget,log(ed50),1,e0)

    sdy<-7.967897
    pop<-c(led50=log(ed50),emax=emax,e0=e0)    
    meanresp<-emaxfun(dose,pop)  
    y<-rnorm(sum(n),meanresp,sdy)
    nls.fit<-nls(y ~ e0 + (emax * dose)/(dose + exp(led50)), 
                              start = pop, control = nls.control(
                              maxiter = 100),trace=TRUE,na.action=na.omit)

    targetD(nls.fit,10,modType=3)
# }