This use case is designed to examine how environmental exposures influence non-communicable diseases (NCDs) in Belgium through two complementary research phases.
The retrospective phase, planned to cover health data from 2012–2021 for myocardial infarction and 2016–2024 for 41 other NCDs, will involve an estimated 300,000–500,000 patients. Its objective is to understand how environmental stressors, such as air pollution, noise, temperature, and green space, relate to both acute and chronic presentations of NCDs. Before any data can be processed, the research team must obtain all required ethical, privacy, and security approvals. Once authorized, the retrospective phase will link anonymized hospitalization data, socioeconomic information, environmental exposure models, and seasonal infection data to build an AI-based risk model capable of identifying how environmental factors contribute to acute clinical events.
To learn more about the study’s methodology, data sources, privacy safeguards, and scientific rationale, please click here to access the detailed description.
Presentation of the research team
- J.F. Argacha, Department of Cardiology, UZB, VUB*
- De Potter, Department of Cardiology, UZB, VUB*
- Sahli, Electronics and Informatics (ETRO) department, VUB*
- Deligiannis, Electronics and Informatics (ETRO) department, VUB*
- Rodrigo Bonet, Electronics and Informatics (ETRO) department, VUB*
Examples of questions that the ENACT study should answer
- How does exposure to air pollution, food pollution, light pollution, noise pollution, temperature, access to green spaces influence the risk of taking medication for a chronic NCD (e.g. arterial hypertension or asthma), or being hospitalized for an acute NCD (e.g., stroke or pancreatic cancer)?
- What combination of environmental and socio-economic factors contributes most to acute health events such as heart attacks or strokes?
- Are certain populations, such as the elderly or lower-income groups, more vulnerable to environmental stressors?
- Can artificial intelligence improve the prediction of hospital admissions due to environmental exposures?
- How do different environmental stressors interact—do they amplify each other’s negative effects on health?
- How do short-term environmental factors (heatwaves, air pollution spikes) and long-term environmental factors (chronic exposure to pollutants) impact health differently?
- Are there seasonal variations in environmental health risks, and how should public health responses adapt?
- Can an exposomic risk score help individuals and healthcare professionals take preventive action against non-communicable diseases?
- Is an exposomic risk score able to detect the disease at its earliest stages before the symptoms start?
- How can cities and policymakers use ENACT findings to design healthier living environments?
- Can personalized environmental risk assessments guide public health campaigns and preventive treatments?
Methodology
The ENACT project aims to understand how environmental factors affect both acute and chronic presentations of NCD. The study follows several key steps to ensure accuracy and privacy.
First, researchers will collect past hospital records and use of specific medication from Belgian social security databases to analyse cases of 42 different non-communicable diseases (NCDs). They will define which diseases to study, the time of analysis, and how to securely handle the data in compliance with privacy laws. Approval from ethics and security committees will be obtained before proceeding.
Next, environmental data will be integrated into the study. This includes pollution levels, temperature, noise, green spaces, and other environmental factors. The data will be linked to both individuals’ living locations and broader geographic zones to examine exposure levels at different scales.
Then, researchers will develop a model to assess the risk of hospital admissions based on environmental exposures. Initially, the model will focus on heart-related conditions like heart attacks, using data from 2012 to 2019. Later, it will be expanded to include all other diseases.
In the next step, the model will be tested and validated using more recent hospital data from 2020 to 2024. This step ensures the findings remain accurate over time and can help predict hospitalizations due to environmental risks. The goal is to create a tool that helps healthcare providers and policymakers reduce health risks linked to environmental exposures.
Finally, the ability of the model to detect the earliest sign of a disease before the symptoms start, as some changes in biomarkers of cardiovascular risk, like heart rate variability, will be tested in a longitudinal study following 1250 volunteers for 4 years.
General description of the different categories of data processed in the ENACT study
Hospitalization and Medication use Data: Tracking Trends in Acute Diseases and chronic NCD
We will examine hospital admissions and specific medication use in Belgium between 2016 and 2024 for 40 specific acute and chronic NCDs. This data comes from the Belgian Health Insurance records (FOD-RIZIV dataset) and the Intermutuality Agency (IMA datasets). We will focus on cases where the primary diagnosis matches specific medical codes (ICD-10) and estimate how many patients were hospitalized for each condition.
Socioeconomic Data: Understanding the Impact of Social Factors
Socioeconomic status can influence a person’s health and their risk of hospitalization. To account for this, we will use data from the Belgian social security database (KSZ dataset), which provides general information such as marital status, job category, and income range.
Environmental Data: Linking Pollution, Climate, and Health
The study will estimate environmental exposures at the statistical sector corresponding to the patients’ home addresses (Vt). A specialized agency (Vlaamse Instelling voor Technologisch Onderzoek, VITO) will use advanced mapping techniques to calculate air pollution, temperature, noise, green space, and other pollutant levels at each person’s home and work sector. This will help us assess how long-term and short-term environmental conditions may have influenced their risk of developing a NCD.
Considering Seasonal Illnesses Like Flu and COVID-19
Since seasonal flu and COVID-19 can increase hospital admissions, we will include data from Sciensano’s influenza tracking system . This will help us separate the effects of pollution from seasonal illnesses when analyzing hospitalization trends.
By combining these different sources of data, our study will help policymakers and healthcare providers better understand how environmental conditions impact public health and how to reduce preventable hospital admissions.
Recipients of personal data
The data collected by Healthdata will be processed by 5 different researchers :
- Dr. De Potter T, Cardiology Department, UZ Brussel, VUB
Medical Doctor and PhD researcher with experience in statistics and epidemiology
Function: Principal investigator of the study
Reason: Verification of data integrity, performance of descriptive statistics and inferential statistics
- Prof. Dr. Argacha JF, Cardiology Department, UZ Brussel, VUB
Medical doctor and professor in Cardiology with expertise in environmental cardiology
Function: Supervisor of the study
Reason: Verification of data integrity, performance of descriptive statistics
- Prof. Sahli H, Electronics and Informatics (ETRO) Department, VUB
Professor in applied computer science
Function: Supervisor of the study
Reason: Management of the study design and statistical analysis
- Prof. Deligiannis N, Electronics and Informatics (ETRO) Department, VUB
Professor in applied computer science
Function: Co-investigator of the study
Reason: Management of the study design and statistical analysis
- Dr Esther Rodrigo Bonet, Electronics and Informatics (ETRO) Department, VUB
Postdoctoral researcher in physics and applied computer science
Function: Co-investigator and senior statistician of the study
Reason: Performing statistics by artificial intelligence and machine learning
Privacy
The ENACT study follows the same meticulous data protection framework as a previous study led by the same research group. Privacy is a top priority, and over a year of preparatory work has ensured that data processing complies fully with the General Data Protection Regulation (EU GDPR). A secure and anonymized data flow has been established to protect patients’ rights and interests. The eHealth platform, a federal public institution, acts as a trusted third party to handle data anonymization before researchers access it. As a result, researchers working with the final dataset cannot trace individual patients or groups of patients. The accessible databases contain no sensitive personal data and are only available on a secure server managed by the Intermutualistic Agency (IMA), a non-profit organization founded by the health insurance funds.
External evaluation of the ENACT study
The ENACT study was approved by the ethics committee of a university hospital (UZB, EC-2025-118). This guarantees compliance with the rules of bioethics.
The secure processing of data will be evaluated by an independent committee. The report of their evaluation can be consulted at the following link: https://www.ksz-bcss.fgov.be/fr/deliberations
Finally, all researchers involved in this work have signed a confidentiality agreement with the data controller of the ENACT project.
Conflict of interest, origin of funding
The researchers involved in the ENACT project have no direct conflict of interest. Their declaration of conflict of interest can be consulted on the following link (see team presentation)
Funding is obtained through the HORIZON Europe program, the EU’s key research and innovation funding initiative.
Results
The site is updated regularly to present the main results
Requests for additional information
Principal investigator and data controller of the ENACT study:
- Dr De Potter T
Centrum voor Hart- en Vaatziekten,
Universitair Ziekenhuis Brussel
Laarbeeklaan 101
1090 Brussel
Email: tom.depotter@uzbrussel.be
- Prof Dr Argacha JF
Centrum voor Hart- en Vaatziekten,
Universitair Ziekenhuis Brussel
Laarbeeklaan 101
1090 Brussel
The prospective phase, planned for 2026–2030, will focus on approximately 1250 apparently healthy volunteers living in environmental “hotspots” identified from retrospective findings. The prospective study will investigate early, pre-clinical signs of cardiovascular dysregulation by combining high-resolution environmental monitoring with physiological and biological measurements.
Environmental exposures, including particulate matter, nitrogen dioxide, ozone, noise, temperature, greenness, and UV radiation, will be captured using ENACT wearable devices and residential geolocation. Physiological indicators such as heart rate variability and skin blood flow variability, as well as biomarkers of inflammation and oxidative stress, will help detect early alterations linked to exposomic stress. Lifestyle variables collected through questionnaires will complement these data. The overarching goal of this phase is to validate and refine the AI-based exposomic risk model developed in the retrospective phase, enabling earlier identification of at-risk individuals and informing public health strategies in environmentally vulnerable communities.