turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

Find a Community

- Home
- /
- Analytics
- /
- Stat Procs
- /
- MCP-Mod using SAS

Topic Options

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

03-01-2017 04:36 PM

Hi all,

I'm trying to replicate an MCP-Mod procedure using SAS and my objective is to have the same results produced by R code. I'm following the code in chapter 7 of the book "Moden Approaches to Clinical Trial Using SAS" (S. Menon, R. C. Zink). The example in the book is based on biom dataset from the R package MCPMod, but I want to apply this code on a different data. The difference is that in my dataset there is a covariate, called base (it's a baseline variable). The SAS code print out the same optimal contrasts that I achieve with R. The best model between the candidate set of model is an Emax model. I know that in the MCPMod package covariates have an additive effect and also I know that usually in an Emax model covariates can have an effect on the parameter. In SAS I really don't know how can I add a covariate in the Emax model.

This is the code that I'm using:

proc nlmixed data=simpredict;

parms e0=0.3 emax=1 ed50=1 sigma=1;

bounds 0.001 < ed50 < 1.5;

mn=e0 + b1*base + emax*dose/(dose + ed50)+base;

model resp ~ normal(mn, sigma**2);

predict emax*dosepred/(dosepred + ed50) out=predout;

estimate 'TD' &deltadiff * ed50/(emax - &deltadiff);

run;

But the TD estimate from this procedure is different from the one that I have using R.

Am I putting in this covariate in the right way?

Thank you to all

CP

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

05-18-2017 10:17 AM

Hi Cristina,

I am doing something similar but I am unable to get the optimal contrasts in R and SAS to match. Do you mind sharing your code that got you the same contrasts?

Thanks

JS