BLQ data in OptimalDesign?

How-to
1

Hello!

I have a quick question about OptimalDesign in Pumas: is there a built-in way to handle BLQ data in design evaluation?

I have tried

@derived begin
    conc_model := @. Central 
    CONC_normal := @. Normal(Central, sqrt(σ_add²))
    CONC ~ @. censored_latent(CONC_normal, LLOQ, Inf)
end

vs

@derived begin
    conc_model := @. Central 
    CONC ~ @. Normal(Central, sqrt(σ_add²))
end

but the results are the same

Thanks in advance!

Lucie

2

Hi Lucie,

Unfortunately, the Fisher information matrix implemented in Pumas assumes a first order approximation and Gaussian error model. There are a few ways we can generalize that to handle different error models but none are implemented.

Regards,

Mohamed

1 Like
3

Hi Mohamed,

That’s what I thought, thank you very much for confirming!

Best regards

Lucie