A care plan is a sequence of medical interventions formulated for curing a specified disease. Doctors usually craft different care plans for different patients based on their knowledge and experience as well as following the clinical guidelines. An intuitive summarization of previous successful care plans not only provides doctors a reference of successful treatment guidelines, but also helps with the detection of anomalous treatments and the improvement of existing care plans. However, producing such kind of summarization is challenging due to the complexity of the data, i.e., large, temporal oriented, and multidimensional. In this paper, we propose a Sankey diagram-based visualization design to visually summarize care plans based on our medical collaborators’ requirements. We apply our tool to a medical dataset of pneumonia patients collected from a Children’s hospital in Shanghai, China. Based on the visualization results, doctors detected many interesting findings, which will be discussed in the paper.