Periodic Reporting for period 1 - CoDiet (COMBATTING DIET RELATED NON-COMMUNICABLE DISEASE THROUGH ENHANCED SURVEILLANCE)
Periodo di rendicontazione: 2023-01-01 al 2024-06-30
Unhealthy diets are associated with metabolic changes and increased risk of non-communicable diseases (NCDs). However, little is known about the dietary mechanisms that actually drive NCDs, and the tools used to collect dietary information are still inaccurate. There is also a lack of data among vulnerable groups, where NCDs are often over-represented. The EU-funded CoDiet project will address the knowledge gaps and develop a tool that will assess diet-induced NCD risk. Specifically, it will develop an enhanced method of dietary assessment using AI technologies. It will also develop a diet–NCD monitoring tool that will enable change in NCDs in response to diet to be monitored at the population level. The overall goal is to promote the uptake of an NCD-protective diet at a population level.
Our current understanding of the relationship between diet and the development of non-communicable disease (NCD) is limited by a number of factors. These include a lack of understanding of dietary mechanisms that drive NCD, inaccurate tools to collect dietary information, a nascent understanding of the role of personalised nutrition, and the lack of data in vulnerable groups where NCDs are often over-represented. The overarching aim of CoDiet is to develop a series of tools (through nine work packages) which will address the current gaps in our knowledge and lead to the development of a tool that will assess dietary-induced NCD risk.
We will achieve this through the six objectives which will answer the challenges of the work programme:
1: Development of AI-driven literature searching tools - bring clear understanding of large global literature in the field of physiological and metabolic links between diet and NCD.
2: Enhance the understanding of NCD risk factors - we will bring a series of beyond the state of the technics to gain mechanistic insight.
3: Understanding of the importance individual variation in response to diet to risk of NCD - this will give insight into the targeting of dietary NCD advice.
4: Develop an enhanced method of dietary assessment using machine learning technologies - solving a fundamental problem in nutrition of lack of an accurate dietary tool.
5: Develop an enhance diet-NCD monitoring tool - enabling change in NCD in response to diet to be monitored at the population level.
6: Develop a dynamic interface between diet and NCD risk factor monitoring and policy - Ensuring CoDiet is applicable at a population level.
The investigation of these objectives and the answers they provide will open a pathway to enhancing the uptake of NCD protective diet at a population level.
Our current understanding of the relationship between diet and the development of non-communicable disease (NCD) is limited by a number of factors. These include a lack of understanding of dietary mechanisms that drive NCD, inaccurate tools to collect dietary information, a nascent understanding of the role of personalised nutrition, and the lack of data in vulnerable groups where NCDs are often over-represented. The overarching aim of CoDiet is to develop a series of tools (through nine work packages) which will address the current gaps in our knowledge and lead to the development of a tool that will assess dietary-induced NCD risk.
We will achieve this through the six objectives which will answer the challenges of the work programme:
1: Development of AI-driven literature searching tools - bring clear understanding of large global literature in the field of physiological and metabolic links between diet and NCD.
2: Enhance the understanding of NCD risk factors - we will bring a series of beyond the state of the technics to gain mechanistic insight.
3: Understanding of the importance individual variation in response to diet to risk of NCD - this will give insight into the targeting of dietary NCD advice.
4: Develop an enhanced method of dietary assessment using machine learning technologies - solving a fundamental problem in nutrition of lack of an accurate dietary tool.
5: Develop an enhance diet-NCD monitoring tool - enabling change in NCD in response to diet to be monitored at the population level.
6: Develop a dynamic interface between diet and NCD risk factor monitoring and policy - Ensuring CoDiet is applicable at a population level.
The investigation of these objectives and the answers they provide will open a pathway to enhancing the uptake of NCD protective diet at a population level.
The CoDiet project started in January 2023 with a successful kick-off meeting organized in London with the 17 consortium partners and representatives from the European Commission.
An External Ethical Advisory Board (EEAB) was established during the proposal phase, and two meetings were held in the first 18 months to discuss work package plans, AI developments, and the impact of dietary policy changes on non-communicable disease (NCD) risk.
Significant progress was made across multiple work packages. CoDiet created the largest collection of biomedical data, verified by both AI and human experts, including unique categories of information. This data has been used to train specialized AI models for reviewing medical literature, helping us understand the connections between diet, disease risk factors, and biological mechanisms, particularly metabolic syndrome.
A pilot study running in the UK, Greece, Ireland, and Spain has started focused on testing new technologies to monitor diet and NCD risk. We proposed using passive cameras and biomarkers from blood and urine for dietary assessment, and integrated genetic, metabolomics, and metagenomics data with non-invasive sensors for NCD risk monitoring. Methods and guidelines for a consumer perception study were also defined, integrating sociology and psychology. The clinical study results expected by October (M22) will contribute to our goals of understanding diet and NCD risk and developing AI models..
To support data storage and analysis, robust infrastructure for data engineering and machine learning was developed, resulting in several published research papers. Advanced machine learning methods and external data sources were identified to enhance AI tools for personalized dietary advice, improving compliance with NCD prevention guidelines.
Policy impacts on diet quality and NCD risk were analyzed, especially among vulnerable socio-economic groups.
A range of dissemination and communication materials was created in the first year for communication channels (CoDiet’s webpage and social media accounts). CODIET has participated and organized several public events. An Exploitation and IP Management Plan was also created, along with an Exploitation Supervisory Board to review key exploitable results.
These efforts collectively aim to understand NCD risk factors, improve diet-related health outcomes, and develop personalized AI models
An External Ethical Advisory Board (EEAB) was established during the proposal phase, and two meetings were held in the first 18 months to discuss work package plans, AI developments, and the impact of dietary policy changes on non-communicable disease (NCD) risk.
Significant progress was made across multiple work packages. CoDiet created the largest collection of biomedical data, verified by both AI and human experts, including unique categories of information. This data has been used to train specialized AI models for reviewing medical literature, helping us understand the connections between diet, disease risk factors, and biological mechanisms, particularly metabolic syndrome.
A pilot study running in the UK, Greece, Ireland, and Spain has started focused on testing new technologies to monitor diet and NCD risk. We proposed using passive cameras and biomarkers from blood and urine for dietary assessment, and integrated genetic, metabolomics, and metagenomics data with non-invasive sensors for NCD risk monitoring. Methods and guidelines for a consumer perception study were also defined, integrating sociology and psychology. The clinical study results expected by October (M22) will contribute to our goals of understanding diet and NCD risk and developing AI models..
To support data storage and analysis, robust infrastructure for data engineering and machine learning was developed, resulting in several published research papers. Advanced machine learning methods and external data sources were identified to enhance AI tools for personalized dietary advice, improving compliance with NCD prevention guidelines.
Policy impacts on diet quality and NCD risk were analyzed, especially among vulnerable socio-economic groups.
A range of dissemination and communication materials was created in the first year for communication channels (CoDiet’s webpage and social media accounts). CODIET has participated and organized several public events. An Exploitation and IP Management Plan was also created, along with an Exploitation Supervisory Board to review key exploitable results.
These efforts collectively aim to understand NCD risk factors, improve diet-related health outcomes, and develop personalized AI models
The Codiet project is in its early stage, therefore many of the results beyond the state of the art are work in progress. At the end of month 18, significant developments include:
- Creation of a dedicated cardiometabolic diseases corpus in standardised format of 1,000 Open Access articles with automated annotation of entities using existing NLP methods.
- Creation of a knowledge graph and ranking of article relevance for physiological processed associated with dietary (potentially modifiable) risk for CMDs for 1,000 OA articles.
- Novel Technologies (passive camera and non-invasive sensors) and pilot study flux trained in 4 sites
- Executing a clinical study with 200 participants for validation novel technology in 4 countries.
- Development of the appropriate infrastructure and informational capacity of the data gathered.
- Interactive sessions among WPs to understand the technical characteristics of the data.
- Report on the update to Food Environment Policy Index assessments describing policy scenarios.
- Creation of a dedicated cardiometabolic diseases corpus in standardised format of 1,000 Open Access articles with automated annotation of entities using existing NLP methods.
- Creation of a knowledge graph and ranking of article relevance for physiological processed associated with dietary (potentially modifiable) risk for CMDs for 1,000 OA articles.
- Novel Technologies (passive camera and non-invasive sensors) and pilot study flux trained in 4 sites
- Executing a clinical study with 200 participants for validation novel technology in 4 countries.
- Development of the appropriate infrastructure and informational capacity of the data gathered.
- Interactive sessions among WPs to understand the technical characteristics of the data.
- Report on the update to Food Environment Policy Index assessments describing policy scenarios.