Prediction in health requires the digital analysis of massive data. In practice, however, medical information is centralised, not shared and not sufficiently structured, which is a barrier to the application of artificial intelligence. Faced with this observation, the objective was to study the interest of e-health in the sharing and structuring of medical data and then to propose a new, decentralised and more suitable technological solution. After a review of the scientific literature, a multicentre case-control study was first conducted using infrared micro-spectroscopy applied to the tumour parts of a retrospective cohort of 100 patients operated on for kidney cancer and followed for 5 years.
This thesis shows that e-health and new digital technologies can enable more efficient sharing and better structuring of health data for predictive medicine.
This work opens the way to a paradigm shift, a revolution in the medical world, with the possibility, among other things, of prospective exploitation of real-life medical data and continuous feeding of the new University Medical Departments (DMU), while putting the patient at the centre of a connected and better coordinated care pathway.