The Project

What is the problem PARADISE wants to address?

In Europe, 30 million people suffer from an autoimmune disease; it is the third largest cause of morbidity and mortality, after cancer and heart disease, in industrialised countries. At the individual level, the impact of suffering can be immense, at the societal level, a significant health and economic burden. 

Autoimmune disease affects 10% of adults, most of whom are women, and two of the top five medications with the highest cost globally are used to keep these recurring conditions in remission. 

These medications suppress the immune system, leaving the patient exposed to increased infection and cancer risk. The general requirement for such treatments, and their side effects, has been raised as a key target for research by the PARADISE consortium patient groups. 

What is the aim of PARADISE?

PARADISE delivers a practical response to this challenge.

We aim to develop a personalised prediction tool that accurately defines the patient’s risk of disease recurrence so that medication doses can be tailored and, in some cases, stopped safely. We use systemic vasculitis as a typical autoimmune disease, bringing together clinical, biomarker and smartphone derived wellbeing data to inform predictive algorithms underpinning a physician tool. 

What exactly will PARADISE do?

PARADISE will combine and analyse multi-source heterogeneous data, create a prediction model, and implement it as a physician-facing tool to inform ANCA-associated vasculitis care, ready for a future clinical trial. Such artificial intelligence (AI) applications are coming under intense EU scrutiny, so we will co-develop an “AI transparency notice”, which will make explicit and explainable the PARADISE tool clinical outputs.

(RKD: Rare Kidney Disease Cohort, AI: Artificial Intelligence, IDIBELL: Bellvitge Biomedical Research Institute)

What is the pathway of this project?

As shown in the diagram below:

  • ⦁ PARADISE commences by collating existing data and bio-samples from biobanks that have ongoing patient sampling regimes
  • ⦁ This data is supplemented by autoantibody and transcriptomic data
  • ⦁ Semantic Modelling, data integration and uplift to RDF (Resource Description Framework) takes place next
  • ⦁ Once a robust statistical analysis plan, statistical modelling and machine learning are in place – the data integration starts and following this, prediction tool development starts
  • Patient and Public Involvement (PPI) is at the heart of PARADISE, patient engagements, workshops and explainable AI work will take place throughout the project.
  • Communication and dissemination, and project management are ongoing tasks.

Partager :