Digital Endpoint Adjudication

Important results from the project

The purpose and aim of the project, to create an application that based on information from a study can learn patterns and thus be able to detect clinical events that lie outside the expected in order to be able to be investigated closer in this way, has succeeded very well. The application is available and can be used via an interface by people without technical competence. By importing data from an Excel file, the application can learn from existing data to give suggestions on events to be controlled. The model is general and can include data from different types of studies.

Expected long term effects

The result of the project in the form of the application has been very successful. It works very well in its current form, although it needs further development to be used on a larger scale. It is above all the user interaction that needs to be developed more and also certain parts of the neural network that need development for it. The effects of the application´s work are judged to be very promising, but have not been able to be used to such an extent that the expected result in the form of an increased percentage of caught events in percentages has been calculated.

Approach and implementation

The layout of the project has worked very well. All steps have been carried out as planned. The challenge has been in the difficulty of accessing resources. Resources were not available for the project as planned, which has meant a delay of the project. Once resources were available for the project, the work has progressed according to plan and the application has been developed as it was intended from the beginning of the project. Some adaptation has been made of the application´s interface to speed up the project.