Periodic Reporting for period 3 - WARIFA (Watching the risk factors: Artificial intelligence and the prevention of chronic conditions)
Período documentado: 2024-01-01 hasta 2025-06-30
The United Nations has specifically addressed the need for preventing noncommunicable diseases in their "2030 agenda for sustainable development". While modern public health care originally emerged in response to infectious disease epidemics, public health now includes noncommunicable diseases. World Health Organization (WHO) defines noncommunicable diseases as Chronic Conditions (CCs) caused by a combination of genetic, physiological, environmental, and behavioural factors which can affect all age groups and countries. The main types of CCs include cardiovascular diseases (CVD), cancers, chronic respiratory diseases, and diabetes.
The WARIFA project developed and validated a prototype of a combined early risk-assessment tool that provides individual citizens with personalised recommendations for the prevention of CCs - such as CVD, melanoma, chronic obstructive pulmonary disease (COPD), and diabetes – which represent the leading causes of mortality in the European Union. The WARIFA app was made available to citizens via a user-friendly smartphone app. The project demonstrated that AI-based combined early risk assessment empowers citizens to adopt healthier habits by delivering personalised guidance to modify risk behaviours.
At the individual level, citizens will be supported in improving by at least 20% each risk factor by increasing the level of physical activity; reducing sun exposure; and for people with type 1 diabetes - reducing the number of hypoglycaemic events and consequently of the related acute admissions in the health care system; or monitor the blood HbA1c level.
At the healthcare level, WARIFA may support earlier identification of CC risk by enabling citizens to self-assess and track risk behaviours, potentially increasing the likelihood of timely prevention and diagnosis. Regarding the organisational structures, WARIFA could help clinicians improve efficiency and resource utilization by reducing the number of consultations for risk assessment and by encouraging prevention-oriented pathways.
WARIFAs interdisciplinary collaboration enabled the WARIFA prototype to reach TRL 7 by the end of the project. The development of the WARIFA app followed an iterative cycle of design, development, testing, and feedback. Throughout 2024 and 2025 the consortium made significant progress, regarding data collection and app development. By M54, all six project milestones were accomplished, and each of the 52 planned deliverables were successfully submitted. The project is now at the end of its planned duration, with 958.71 PMs reported, representing 100% of the total PMs planned.
The WARIFA project developed and validated a prototype of a combined early risk-assessment tool that provides individual citizens with personalised recommendations for the prevention of CCs - such as CVD, melanoma, chronic obstructive pulmonary disease (COPD), and diabetes – which represent the leading causes of mortality in the European Union. The WARIFA app was made available to citizens via a user-friendly smartphone app. The project demonstrated that AI-based combined early risk assessment empowers citizens to adopt healthier habits by delivering personalised guidance to modify risk behaviours.
At the individual level, citizens will be supported in improving by at least 20% each risk factor by increasing the level of physical activity; reducing sun exposure; and for people with type 1 diabetes - reducing the number of hypoglycaemic events and consequently of the related acute admissions in the health care system; or monitor the blood HbA1c level.
At the healthcare level, WARIFA may support earlier identification of CC risk by enabling citizens to self-assess and track risk behaviours, potentially increasing the likelihood of timely prevention and diagnosis. Regarding the organisational structures, WARIFA could help clinicians improve efficiency and resource utilization by reducing the number of consultations for risk assessment and by encouraging prevention-oriented pathways.
WARIFAs interdisciplinary collaboration enabled the WARIFA prototype to reach TRL 7 by the end of the project. The development of the WARIFA app followed an iterative cycle of design, development, testing, and feedback. Throughout 2024 and 2025 the consortium made significant progress, regarding data collection and app development. By M54, all six project milestones were accomplished, and each of the 52 planned deliverables were successfully submitted. The project is now at the end of its planned duration, with 958.71 PMs reported, representing 100% of the total PMs planned.
WARIFA started on the 1st of January 2021 and has a duration of 54 months.
The consortium gathered from the 17th to the 19th of June 2024 for the final consortium wide meeting in Tromsø, Norway.
The following main results were achieved in the projects final reporting period:
1. Periodic activity and management report - update 2
2. Periodic activity and management report - update 3
3. Periodic risk and ethic management report – update 1
4. Report of the tests made of the Prototype version of the WARIFA app
5. AI prediction engine integrated in the WARIFA system
6. Context engine that enables the automatic extraction of context information and identification of dynamic rules from the data generated in WP4
7. Semantic and structured data base for the storage of context data and rules
8. Simulation technique for the generation of big data
9. Report on procedure performance
10. Fine-tuning and validation report
11. Dissemination – activities report – update
12. Business & Commercialisation Plan - update
13. Policy recommendations
In the last reporting period, the dissemination of the project results was supported through the realization of the following D&C actions:
Participation in national/international events, with 11 events attended, and 1 project workshop organized.
Regular updates released on the WARIFA project website, with 20 news published.
Extensive activity on social media platforms, i.e. LinkedIn, Twitter, and Facebook., with 144 posts published.
Realization of 1 project video, spread across all project channels.
Networking with related EU funded projects
The following milestones were reached by June 2025:
Validation - Report of the tests made of the Prototype version of the WARIFA app Co-creation report (30 Jun 2025)
Novel framework and new pathways (30 Jun 2025)
The consortium gathered from the 17th to the 19th of June 2024 for the final consortium wide meeting in Tromsø, Norway.
The following main results were achieved in the projects final reporting period:
1. Periodic activity and management report - update 2
2. Periodic activity and management report - update 3
3. Periodic risk and ethic management report – update 1
4. Report of the tests made of the Prototype version of the WARIFA app
5. AI prediction engine integrated in the WARIFA system
6. Context engine that enables the automatic extraction of context information and identification of dynamic rules from the data generated in WP4
7. Semantic and structured data base for the storage of context data and rules
8. Simulation technique for the generation of big data
9. Report on procedure performance
10. Fine-tuning and validation report
11. Dissemination – activities report – update
12. Business & Commercialisation Plan - update
13. Policy recommendations
In the last reporting period, the dissemination of the project results was supported through the realization of the following D&C actions:
Participation in national/international events, with 11 events attended, and 1 project workshop organized.
Regular updates released on the WARIFA project website, with 20 news published.
Extensive activity on social media platforms, i.e. LinkedIn, Twitter, and Facebook., with 144 posts published.
Realization of 1 project video, spread across all project channels.
Networking with related EU funded projects
The following milestones were reached by June 2025:
Validation - Report of the tests made of the Prototype version of the WARIFA app Co-creation report (30 Jun 2025)
Novel framework and new pathways (30 Jun 2025)
Health apps are already being used in the prevention and management of CCs. However, the issue of long-term commitment continues to be a major challenge. By improving the understanding of what simple changes in the users` behaviour can do for their health, the motivation to make lifestyle-related changes can improve substantially. The willingness to engage in their own health, may be further increased by providing users with one single app that can manage multiple CCs. In addition to better self-management for citizens and patients, the Warifa app may also help healthcare personnel. Although the Warifa app is primarily designed for citizens and patients, information on the advice given by the app, may also inform decisions to be taken by the user`s doctor. As environmental conditions, such as air and water quality, is recorded by the app and directly can influence disease outcomes, health personnel may find the information from the Warifa app useful.
WARIFA developed a prototype of an AI-based system that aims to help prevent chronic conditions for all citizens. To achieve this objective, it was necessary to combine ubiquitous data and personal user-generated data, and interdisciplinary efforts from clinical, technical, and sociological backgrounds.
By combining ubiquitous data from the user`s environment with user-generated data, AI algorithms can process the most relevant data in the appropriate context and then provide the tools for personalized advice resulting in more specific preventive interventions.
The main results from the work conducted are available in the following scientific papers:
Prediction of hypoglycaemia in subjects with type 1 diabetes during physical activity
Interpretable and multimodal fusion methodology to predict severe hypoglycemia in adults with type 1 diabetes
A Bayesian Belief Network model for the estimation of risk of cardiovascular events in subjects with type 1 diabetes, Computers in Biology and Medicine,
Making “inclusion” more than a buzzword: A critical interpretive synthesis of literature about recruiting seldom-heard groups in health research
Incorporating Uncertainty Estimation and Interpretability to Personalised Glucose Prediction Using the Temporal Fusion Transformer.
An AI-based module for interstitial glucose forecasting enabling a “Do-It-Yourself” application for people with type 1 diabetes.
Personalized glucose forecasting for people with type 1 diabetes using large language models.
Evaluating Time Series Classification Models for Nocturnal Hypoglycemia: From Predictive Performance to Environmental Impact.
Transfer learning for a tabular-to-image approach: A case study for cardiovascular disease prediction. Journal of Biomedical
The developed technical prototype was tested in a pilot trial. The aim of the testing was to get a proof of concept that a single app can provide lifestyle recommendations relevant for multiple CCs while taking into account the context of the user.
WARIFA developed a prototype of an AI-based system that aims to help prevent chronic conditions for all citizens. To achieve this objective, it was necessary to combine ubiquitous data and personal user-generated data, and interdisciplinary efforts from clinical, technical, and sociological backgrounds.
By combining ubiquitous data from the user`s environment with user-generated data, AI algorithms can process the most relevant data in the appropriate context and then provide the tools for personalized advice resulting in more specific preventive interventions.
The main results from the work conducted are available in the following scientific papers:
Prediction of hypoglycaemia in subjects with type 1 diabetes during physical activity
Interpretable and multimodal fusion methodology to predict severe hypoglycemia in adults with type 1 diabetes
A Bayesian Belief Network model for the estimation of risk of cardiovascular events in subjects with type 1 diabetes, Computers in Biology and Medicine,
Making “inclusion” more than a buzzword: A critical interpretive synthesis of literature about recruiting seldom-heard groups in health research
Incorporating Uncertainty Estimation and Interpretability to Personalised Glucose Prediction Using the Temporal Fusion Transformer.
An AI-based module for interstitial glucose forecasting enabling a “Do-It-Yourself” application for people with type 1 diabetes.
Personalized glucose forecasting for people with type 1 diabetes using large language models.
Evaluating Time Series Classification Models for Nocturnal Hypoglycemia: From Predictive Performance to Environmental Impact.
Transfer learning for a tabular-to-image approach: A case study for cardiovascular disease prediction. Journal of Biomedical
The developed technical prototype was tested in a pilot trial. The aim of the testing was to get a proof of concept that a single app can provide lifestyle recommendations relevant for multiple CCs while taking into account the context of the user.