The need to understand users lies at the core of product design process, but traditional approaches of studying user behavior carry potential limitations. Therefore, making use of the data logged from the user interaction with the system, provides one with factual insights on user behavior to correctly identify and understand their needs and the actual use of the system. However, this data is often not logged for the purpose of studying user behavior, posing in this way a challenge for getting relevant insights. To help identification of the context, the different phases of an intervention and key events during these phases, in 2 hospitals a camera system is installed, collecting video material of cases. Also the x-ray images are collected from these cases to give information on the ongoing case. By mapping the system logs to both the video and x-ray image data, insights could potentially be generated about key events in the workflow, that can be used to predict upcoming phases in the the intervention and/or key events. By predicting upcoming events, the workflow of clinical cathlab users can be significantly supported by sending role-based and pro-active messages to users that otherwise would be idle or waiting for information. A simple example would be to inform the transporter of a patient 10 minutes before a procedure ends, to make sure there is a bed and an empty spot at the ward.
The focus will be on the definition of a method that can discover and predict the various phases and key events of interventional procedures from above described three sources of data. The method will produce an algorithm that is trained on these 3 sources, but will also work robustly for different labs, different users and when there is no video footage or x-ray information available. Logically, limitations will show itself when less input data is available, it is important to understand and document these limitations as these will be input to the sensor roadmap for the x-ray system.
As such, the research goal that is addressed is: Given a system event log, video footage and x-ray image data, how can the different (clinical) workflow phases in the lab be detected and predicted by identifying and predicting key workflow events.
For more information, contact Natalia Sidorova: N.Sidorova@TUE.nl