proportionalRecovery.txt

model{
    for (nn in 1:NMeasurements){
        y[nn] ~ dnorm(ym[nn],yp) # Model fit to data
        ym[nn] <- alpha[id[nn]] + r[g[id[nn]]]*(66-alpha[id[nn]]) * (1-exp(-t[nn]/tau[g[id[nn]]])) # Model estimate
    }
    for (ss in 1:NSubjects) {
        alpha[ss] <- alphaT[ss] * 66 # Patient specific intercept in range 0-66
        logit(alphaT[ss]) <- alphaL[ss] # Patient specific intercept in range 0-1
        alphaL[ss] ~ dnorm(alpham[g[ss]],alphap[g[ss]]) # Patient specific intercept in range -inf +inf
        g[ss] ~ dcat(gp[]) # Group assignment for a patient
    }
    for (gg in 1:NGroups) {
        rL[gg] ~ dnorm(0,1.0E-3) # Recovery coefficient for group g in range -inf +inf
        logit(r[gg]) <- rL[gg] # Recovery coefficient for group g in range 0-1
        tau[gg] <- tauShift[gg] + 1/7
        tauShift[gg] ~ dgamma(1.0E-3,1.0E-3) # Time constant for group g
        alpham[gg] ~ dnorm(0,1.0E-3) # Group initial distribution mean in logistic space
        alphap[gg] ~ dgamma(1.0E-3,1.0E-3) # Group initial distribution precision in logistic space

    }
    gp[1:NGroups] ~ ddirch(gpalpha[1:NGroups]) # Group assignment probability
    yp ~ dgamma(1.0E-3,1.0E-3) # Model precision
}