Typically, the laboratory results are converted into categorical values: either as normal/abnormal or as low/normal/high. Converted categorical values at baseline are then compared to each subsequent visit, resulting inrepeatedcategoricalmeasurements.However,becauseofthe inherent dependenceamongrepeatedobservationsmadeonthesame experimental unit and the inability to control the circumstances when obtaining measurements, an inferential analysis that compares the entire trends of shifts across treatment groups or populations poses a great challenge to statisticians. This type of laboratory shift analyses constitutes a significant portion of clinical trials research. This article proposes a method to analyze such trends of shifts in laboratoryvaluesusingthecontinuous-timeMarkov chain. An important advantage of this proposed model is that transition probabilities across each subject�s repeated categorical measures are modeled as a function of continuous intensity parameters that account for the differences in transition times. After having to account for the differences in such intensity of shifts in laboratory values, the comparison of trends across treatment groups or populations should be lessbiasedandmore precise than methods that ignore such information.

homogeneous continuous-time Markov chain, repeated categorical measurements, laboratory shift analyses, clinical trials.