Many experiments aiming at the investigation of nerve repair involve elaborate testing over a certain time period. Data arising from such experiments are often analyzed by time-point. Such a cross-sectional approach is often very inefficient. In this paper, we consider a case study in which repeated measurements of two response variables assumed to be Poisson distributed are obtained. We show how a repeated measures modeling approach, based on generalized linear models, can handle both responses in one model and improve the inference in the nerve repair experiments. The benefits of the model as well as problems that can occur are illustrated and discussed.