Hi all, My stat colleague was wondering if the Emax models in Phoenix can be used to model data in a crossover study (3 doses) and within subject correlations estimated. I am not that familiar with the Emax module so hoping somebody can comment. Thanks, Mike
Dear Mike I am not sure I understood the question perfectly but let me tell you what you can do with NLME. You can define a column called occasion where occasion =1 refers to the first period of the cross over, 2 to the second and 3 to the third. Then you can define Emax with both a fixed effect, a between subject random effect and also a between occasion random effect. When you do that, you can get the Bayesian estimates of Emax and retrieve for each subject their own Emax at each occasion. This enables you to calculation the correlations across the different occasions for Emax. Is it what you want to do? Best is to send me an example explaining what you are trying to do and I can make you a demo project. Note that all this applies to NLME (population approach) best Serge
Hi Mike, perhaps your stat colleague was thinking of using the BE wizard (or NCA & Toolbox =LME) with the results from individually modelling the individuals in the 3 periods with the EC50, EMax etc. instead of the more typical Cmax and AUClast; if so it seems feasible to me. Simon