Setting feeding status as covariate

Hi all, My question is if feeding status can be used as covariate of Kt or TLag in delayed absorption model (transit or Tlag model) It is a food effect study (19patients, crossover, no meal, low fat and high fat). There are 19 patients x 3 feeding status x 6 time points = 342 observations. The data are not rich enough to run the basic transit or Tlag population model, and can not support estimating IOV either. But there are correlations between feeding status (setting no meal=1, low fat=2, high fat=3) and individually estimated Kt or Tlag. I found the delayed absorption population models worked if I set feeding status as covariate of Kt or TLag. I also found few published paper using feeding status as covariate. So, I wonder if it is acceptable to add feeding status as covariate of Kt or TLag. Thanks in advance, Bo