Unified Modeling Language (UML) for Hospital-based Cancer Registration Processes


Objective: Hospital-based cancer registry involves complex processing steps that span across multipledepartments. In addition, management techniques and registration procedures differ depending on each medicalfacility. Establishing processes for hospital-based cancer registry requires clarifying specific functions andlabor needed. In recent years, the business modeling technique, in which management evaluation is done byclearly spelling out processes and functions, has been applied to business process analysis. However, there arefew analytical reports describing the applications of these concepts to medical-related work. In this study, weinitially sought to model hospital-based cancer registration processes using the Unified Modeling Language(UML), to clarify functions.
Methods: The object of this study was the cancer registry of Osaka UniversityHospital. We organized the hospital-based cancer registration processes based on interview and observationalsurveys, and produced an As-Is model using activity, use-case, and class diagrams. After drafting every UMLmodel, it was fed-back to practitioners to check its validity and improved.
Results: We were able to define theworkflow for each department using activity diagrams. In addition, by using use-case diagrams we were able toclassify each department within the hospital as a system, and thereby specify the core processes and staff thatwere responsible for each department. The class diagrams were effective in systematically organizing theinformation to be used for hospital-based cancer registries. Using UML modeling, hospital-based cancerregistration processes were broadly classified into three separate processes, namely, registration tasks, qualitycontrol, and filing data. An additional 14 functions were also extracted. Many tasks take place within thehospital-based cancer registry office, but the process of providing information spans across multiple departments.Moreover, additional tasks were required in comparison to using a standardized system because the hospitalbasedcancer registration system was constructed with the pre-existing computer system in Osaka UniversityHospital. Difficulty of utilization of useful information for cancer registration processes was shown to increasethe task workload.
Conclusion: By using UML, we were able to clarify functions and extract the typical processesfor a hospital-based cancer registry. Modeling can provide a basis of process analysis for establishment ofeficient hospital-based cancer registration processes in each institute.