Lognormal PK Estimates

From: David Collins Date: February 18, 1994 technical Source: phor.com
Using the first order method, is it possible to use lognormal distribution for pk estimates. If so the case, how would a typical control stream look for say clearance and a cv variance model? We are working with a PC version of NONMEM with the Microsoft FORTRAN Powerstation. Getting back to the problem at hand, or do you have to use the CE method if you believe your parameters are distributed lognormal? If so, what then would the control stream look like. Could you please give a brief explanation of the theory if possible.