Pharmacological studies often involve drug safety issues aimed at testing for a difference between experimental groups where the data are longitudinal in nature. Linear Mixed Effects model (LMM) has been used to analyze this kind of data. Frequently, treatment effect, if present, is not constant over time so that it is important to investigate the trend of a difference between experimental groups. The shape of the trend is usually unknown. Therefore, a more flexible approach is required to model the mean response curve. We use a semi-parametric mixed effects model to test a difference between experimental groups. We derive Bayesian confidence intervals and test statistics for general hypotheses in semi-parametric mixed effects models.