This Master’s Thesis at the Politecnico di Milano is the first formal academic output linked to the CORE-MD project. This pilot study focused on the development of an IT tool for assisting medical devices post-market surveillance. Interestingly, this tool aims to retrieve and aggregate public information relevant to safety notices from the official website of the Italian competent authority. As a matter of fact, it could help the expert panels (EXPAMEDs) make decisions about reviewing new high-risk medical devices.
The final product of this thesis is a graphical user interface for clustering and visualising the medical devices notices. It offers different functionalities such as filtering the devices by on or more manufacturers or by one or more CND
categories. By accurately selecting and visualising the portion of the dataset of interest, expert panel members are helped in the time-consuming initial MD screening process. Through the available graphs and statistics they are
supported in the process of studying the phenomenon.
With the developed tool, it is possible to obtain a level of detail of over 90%. Expert panels are therefore facilitated in their research as they could reach almost complete information in a very short time. However, this work certainly has limitations:
- the reduced size of Dataset 2 did not allow us to use machine learning methods.
- notices were not categorized in terms of safety concerns for the patients.
Starting from this last aspect, as future works, it is clear that extrapolating and analysing the contents of the pdf files describing the individual notifications would allow to categorise the notices by the type and severity of the device’s
technical problem.
Another aspect that could be explored in detail is the action taken by the manufacturer after the incident, e.g., the device may be recalled, undergo a modification or require a software update.
Read the full Executive summary
