In a fast-changing healthcare environment, understanding the changes of medical behaviors in clinical pathways can help hospital managers improve the pathways and make better medical strategies for patient careflow. In this study we propose an approach to detect medical behavior changes between two time periods, by providing a change pattern detection algorithm dividing the discovered change patterns into four categories (i.e., perished patterns, added patterns, unexpected changes, and emerging patterns). The proposed approach is evaluated via real-world data sets extracted from Zhejiang Huzhou Central Hospital of China with regard to the clinical pathway of bronchial lung cancer in 2007-2009 and 2011. The experiment results include three categories of change patterns from the collected data-sets, making a relatively comprehensive cover on the significant changes in clinical pathways, which might be essential from the perspectives of clinical pathway analysis and improvement.