refernce period bayesian estimates "carried-forward"

Dear Users, I works on a model describes IVIVC using the approach described by Gaynor et al 2011. The input data come from a cross-over study with a reference period (imediate release). I would like to use the KA bayesian posthoc estimates of this refernece period for all other periods unchanged. My question is, how can I separate this period from the rest ones to ensure that nonmem does not fit this parameter other periods, but the reference? Thanks, István István Németh PhD Developmental Drug Metabolism & Pharmacokinetics Pharmacological and Drug Safety Research Gedeon Richter Plc Tel: 00 36 1 505 7082 Fax: 00 36 1 889 8526