Evaluation in Barcelona on began on 20th January. Accenture met the local team the Hospital Clínic de Barcelona to exchange ideas and plan the next few day’s work. Dr. Jesus Blanco, resident endocrinologist, arranged meetings with colleagues and patients with appointments scheduled during the visit.
Patients were asked for their consent and were shown how the UCY Flexpass system worked. They were then given the opportunity to interact with it. Interactions with the tool were guided, with close communication with the interviewer only provided if requested by the patient. This minimized any potential influence on patient opinion over the evaluation of the Flexpass system. Patients evaluated their interaction with the system via a questionnaire.
Concurrent interviews took place with hospital IT and assistance personnel and the Accenture and FCRB teams discussed data security and the servers that hold the patient data. Feedback was very positive and validated the decision to enact a two-phase authentication.
Further interviews with healthcare professionals were conducted, focusing on their opinion of how patients would interact with the Authentications system developed by the University of Cyprus.
A final consultation clarified the conclusions and feedback ascertained over the previous two days in relation to the interaction of the different group ages, their opinion and, the trust heavily invested in the pre-existing security measures at the Hospital Clínic de Barcelona.
Fabricated medical data is required to predict patient response to a given treatment pathway. This data is of initially of unknown quality. It is necessary to verify and improve data fabrication quality to ensure the generation of representative models of real-world data. Verification of data fabrication quality is confounded by patient confidentiality and the resulting inability to access the real medical data. As a result, “distinguishers”, able to distinguish real and fabricated data, have been created using machine learning. These distinguishers identify the most real-world representative data and used repeatedly to improve the data fabrication. The machine algorithms use decision trees since data is produced in an understandable output and works effectively on this kind of data.
Machine learning implementations have been created that can run on any data set and can be run remotely by partners across the project. The University of St Andrews has successfully used them to run both real medical and fabricated data. The process of improving data fabrication is continually evolving as the project progresses.