Program

WP 1 - University: TUD 
Faculty: Electrical Engineering, Mathematics and Computer Science

Integrated data science with focus on health

In WP1, we are developing a mathematical model that estimates the risk of complications after colorectal surgery. Since healthcare dataset may consists of complex dependencies between variables, we use copula-based model. Specifically, we use vine-copula for its flexibility to describe complex dependencies and its relatively computationally cheap for doing inferences. Before we apply the model, we investigate whether we use all of dataset, or more recent ones. This is in relation with the possibility of dataset shifts that may happen throughout the year. Lastly, we would like the model to be able to update itself continuously as new data becomes available.

Where has the WP1 developed over the past 2 years

Over the past two years, our team has been working on patient risk prediction in healthcare. We have fine-tuned predictive models across various domains, including colon cancer, weight-loss surgery, and bowel surgery, to improve accuracy and clarity in forecasting patient outcomes. Our efforts have identified which models best balance performance with ease of interpretation.

A key part of our work has been the rigorous selection of relevant risk factors. We can focus on the crucial elements that influence postoperative outcomes by filtering out unnecessary information. A first example of our progress is showcased in Gidius van de Kamp’s thesis, “Patient Level Predictions in Bowel Surgery: Comparing Variable Selections for Rare Outcome Modeling on Real Surgery Data.” His work explores statistical techniques to improve prediction accuracy, specifically for rare postoperative outcomes, using real data from our collaborator, Medisch Spectrum Twente (MST).

By working with MST data, we have also dedicated considerable effort to data cleaning and processing. This includes handling missing values through imputation strategies, ensuring data consistency across sources, removing noise and outliers, and standardizing data formats. These steps are essential to transform raw, real-world data into a reliable statistical analysis and modeling basis.

Another standout area in our work has been using vine copula models in patient risk profiling. These models are excellent for uncovering complex, non-linear relationships in healthcare data, relationships that traditional methods might overlook. With vine copulas, we aim to capture how different health indicators interact, providing deeper insights into patient risk.

Our current challenge is integrating diverse data types, from continuous measures to ordinal categories, into our predictive models. Ultimately, we aim to enhance predictive accuracy and make data-driven policy recommendations using these advanced methodologies. We can support more targeted and effective healthcare strategies by distinguishing high-risk patients from those with low or no risk.

What was the status before the WP1?

Prior to the start of RECENTRE, the team members were highly motivated by practical problems and engaged in multi-disciplinary projects (or internship) and driven to develop tailored statistical methods to applied settings. The team members were also building solid theoretical and methodological background relevant for the RECENTRE work. All three members had affinity, but were not limited to, health-related applications.

Current status

Currently, we have a PhD student, Victor Ryan, who has been on this project since October last year. At the start of the project, he has been studying the theory of copulas and vine-copulas. Furthermore, since the challenge is to integrate various data types into the model, he explored the limitations and implications of including discrete variables in copula-based models. The next step will involve detecting dataset shift that may occur throughout the year. We will begin by studying the types of dataset shifts and their implications for the dependence between the variables under each type of shift.

Collaboration

Our collaborative efforts with partners such as MST, Ziekenhuisgroep Twente (ZGT), and the Dutch ColoRectal Audit (DCRA) have been transformative for our research. By combining our strengths, clinical data, expertise, and insights, we have enriched our modeling efforts and improved the validity of our predictive systems.

WP 2 - University: UT-1 
Faculty: Electrical Engineering, Mathematics and Computer Science 

Dynamic modelling for Complications after Cancer and Obesity

WP2 focuses on the development of AI-driven dynamic predictive models to detect complications arising from cancer, obesity, and their treatments. The goal is to use continuous patient monitoring data to track clinical changes over time, enabling early identification of health deterioration. These models will provide clinicians with valuable insights to support timely interventions, improve post-treatment care, and enhance patient outcomes through personalized and proactive healthcare solutions.

Status Before WP2 & Current Status

Before WP2 began its activities, the work within the consortium was at the stage of conceptualizing the application of dynamic models for predicting post-treatment complications and identifying relevant data sources. We only had ideas, but didn’t know how to achieve this and which partners to collaborate with.

Since WP2 began its activities (developments in the last 2 years), we have carried out various steps to achieve the goals of the work package as described in the RECENTRE proposal.

• Collaboration & Data Access: Ongoing collaboration with multiple institutions to get access to relevant patient datasets for model development. Currently we have established collaboration with Prinses Maxima Centre for obtaining data on late effects after childhood cancer and with ZiekenhuisGroep Twente for access to their bariatric surgery follow-up database.
• Systematic Review: Conducting a systematic review on dynamic models for predicting complications after cancer or cancer treatment to existing approaches and guide future model development.
• Data collection: Small pilot on collecting vital signs and activity data using the Philips healthdot sensor in a cohort of 30 patients who underwent major abdominal oncologic surgery to monitor for complications. This was carried out in collaboration with Medisch Spectrum Twente.
• Preprocessing & Model Development: Initial work on processing patient data and preliminary development of machine learning models to predict early signs of complications in patients ongoing major abdominal oncological surgery.
• Setting up a new study to monitor symptoms in cancer survivors based on data obtained from wearables (in collaboration with Helen Dowling Institute and WP5).
• Pilot research into how digital phenotyping using smartphones can be used to detect the onset of late-effects in cancer survivors.
• Presentations at various conferences on the work carried out in WP2.

WP 3 - University: TU/e 
Faculty: Electrical Engineering 

THz and Photonic health sensors  for monitoring

WP3 focus on implementing terahertz and photonics sensor for monitoring obesity, diabetes and cancer patients. As monitoring patients becomes more holistic, it is essential to introduce different metrics from different sensors, thus terahertz and photonic sensors, which were used to identify biomarker, can be used in complement with other conventional health sensors. We hope to develop a sensor solution that is continuous, cost-effective, unobtrusive, rapid, and personalized for interventions. 

Achieved and current status 

WP3 began in February 2024. The first year has primarily focused on preparation, which includes the procurement of equipment and participation in several courses and training activities. These activities include the Terahertz Summer School, technical courses and attendance at the IEEE Photonics Benelux Annual Symposium 2024. Since the PhD student has no previous experience in terahertz, this valuable experience allows him to learn essential experimental technique, such as alignment, data processing. Additionally, we have been collaborating with WP4 from Wageningen university on the project EWUU-HOBO, where we explore whether terahertz waves can be used to measure body water content and the concentration of creatinine in blood. 

We are working on several projects in parallel: 

• Systematic Review: The goal of this literature review is to examine previous studies to discuss possible opportunities for implementing terahertz sensors for monitoring obesity, diabetes, and cancer patients. This review will help us to understand how terahertz wave could be used in these applications. 
• Skin Optical Phantom: We are developing an optical phantom that mimics the optical properties of human skin at terahertz frequencies. We are testing this phantom with a 300 GHz setup designed and implemented at TU/e. With this phantom, we will be able to conduct experiments under more controlled conditions and gain insight for future skin study in vivo. 
• Moisture Evaluation of Bitumen: This work is to evaluate the moisture level in bitumen, a construction material used on road pavements. This experiment helps us to understand the underlying damage mechanism caused by water in bitumen.  We used tuneable continuous wave spectroscopy system and 2.52 THz continuous wave terahertz system to perform comprehensive analytic study.  

Future plan 

To implement terahertz and photonics sensor, what we do? Will you to use TDS and CW THz system, if yes, could you describe how shortly? Or will you design different sensor for monitoring?   

To demonstrate terahertz sensor potential in monitoring application, we will modify a terahertz time-domain spectroscopy system specifically for monitoring. For instance, we aim to investigate skin hydration level in obese patients after performing bariatric surgery, using terahertz wave. To achieve this, we will use a robotic arm to mount and position the probe to measure at different body locations, by analysing the time domain signal and other related parameters such as BMI, waist circumference, we will gain better insight on obesity skin hydration development after the surgery. Additionally, we plan to measure creatinine concentration in blood using terahertz wave, we will use a microfluidic chip that collects the extracted blood sample, by analysing the absorption spectrum in the chip, the corresponding concentration can be determined. 

WP 4 - University: WUR-1 
Faculty: Agrotechnology and Food Sciences 

Measuring the unmeasurable: identifying how disease status impacts health of multiple organ systems over the life course.  

To develop tailored risk-based recommendations for high-risk populations it is essential to collect data on physiological and behaviour outcomes through monitoring. However, if monitoring becomes too burdensome and interferes with the daily life activities of a patient this can negatively impact a patient's quality of life. Furthermore, this can reduce adherence to the monitoring activities. For effective monitoring it is, thus, important to find a balance between amount of data that needs to be collected and the burden of monitoring for the patient. 

Wearable sensors and home-based monitoring provide opportunities to reduce the burden of monitoring for the patient while increasing or maintaining the amount of data collected. Within our workpackage, we aim to investigate the potential of wearable devices and home-based monitoring for monitoring the health of a particular target group:  women who have received surgical or pharmacological treatment for obesity. Though there is ample evidence that obesity treatments are effective in reduce long-term bodyweight and obesity-related comorbidities. This patient group is at risk of late negative effects of treatment due to the impaired nutritional status. The extend to which this impaired nutritional status impacts health on the longer term, however, is largely unknown. Contributing to this current focus of monitoring on body weight and comorbidities, but also the high rate of loss to follow-up. Within two years 30 to 60% of the patients stop attending follow-up visits. In this patient group there is thus a need to increase monitoring of long-term outcomes without increasing the burden for the patients.

Over the last two years, our research team, in collaboration with the Vitalys, has established a prospective observation cohort study for which recruitment started in January 2025. The aim of this study, called the MONUCO study, is to monitor the health of 1150 women for up to 10 years after obesity treatment. The results of this study will be used to investigate association between lifestyle behaviours, nutritional status and health outcomes. Additionally, the validity and feasibility of a selection of wearable sensors and other home-based monitoring methods will be assessed in this study populations. The research team has started an investigation into the feasibility and validity of using commercially available wearables sensors for monitoring patients after obesity treatment. This investigation was kicked off with a scoping review on the influence of user characteristics, such as sex, age and BMI, on wearable validity in collaboration with other members of the RECENTRE program. Furthermore, together with the RECENTRE members from Eindhoven and other collaborators for the University of Utrecht and University Medical Center Utrecht, the research team has started exploring opportunities for improving the feasibility and accuracy of body composition measurement in patients with obesity. 

WP 5 - University: UT-2 
Faculty: Behavioural, Management and Social Sciences 

High engagement of at-risk populations with unobtrusive high-tech monitoring and intervening in daily-life by using experimental adaptive future-self approaches  

Development of eHealth monitoring and coaching systems starts with including participants needs and wishes to determine the scope and acceptance of technology. This "peoples wishes first approach” is an important step in developing meaningful and effective health technology. The downside is that users can only express needs which they are currently able to imagine. In work package 5, we are using virtual reality to visualize potential future sensor technology to create a more vivid and experiential needs assessment experience. Building on these needs, ways to proactively engage cancer- and obesity patients with innovative, tailored, and non-obtrusive sensor technology for health monitoring will be investigated. This will be done by letting them experiment with sensor-related future-self representations in the form of digital twins, and the sensor system in an adaptive virtual environment which simulates daily living in a future home. By using virtual reality as a visualization tool for formative evaluation in eHealth development, we aim to develop systems that contribute optimally to meaningful lifestyle change of high-risk populations to ensure successful and sustainable implementation of the system. 

Using virtual reality for needs assessment purposes is an innovative approach, so not much was done in this regard before the project, especially in the context of eHealth. Virtual reality in itself is a relatively new technology and the concept of future selves and digital twins is also not extensively researched. However, some information on virtual reality- and future-self interventions in the healthcare context is already available through previous studies. Therefore, literature reviews on these two topics were initiated. Both reviews will soon be finished, with the future-self review being in the final stages of preparation for journal submission, and the virtual reality review currently going through the process of peer review.

Lessons learned along the way were used to design the first virtual reality experiments consisting of four future home variations which differ in the level of sensor obtrusiveness and the framing of feedback messages. Experimental conditions with high sensor obtrusiveness include wall-mounted sensors and an interactive smartwatch, whereas low sensor obtrusiveness conditions only include a non-interactive sensor patch. The feedback messages were either task-focused, giving only factual information, or person-focused, introducing compassionate and encouraging elements to the feedback.

The virtual reality environment was created from scratch in collaboration with the BMS Lab at the University of Twente. The game engine Unity was used to create a living space in which a sensor-feedback system was integrated to show to the participants what sensor-assisted lifestyle change in daily life could look like in the future. Needs and wishes will be assessed by means of self-report questionnaires revolving around, e.g., behavioural stages of change, health literacy, and internal feedback styles, as well as a semi-structured interview. Data collection for the virtual reality needs assessment is about to start, both for participants with cancer and obesity. Soon after, the future-self review and results from the needs assessment will inform the next stage of the project in which digital twins will be embedded into the virtual reality experience.

WP 6 - University: WUR-2 
Faculty: Social Sciences Group 

Adaptive Lifestyle Interventions Empowering Patients

Many health challenges, such as cardiometabolic diseases, are complex, dynamic, and vary over time and individuals. To address this, our research focuses on developing adaptive lifestyle interventions tailored to patient and healthcare needs. These interventions move beyond a one-size-fits-all by delivering personalized behavior change strategies. This innovative approach combines sensing technology with personalized care to address patients’ needs and improve health outcomes.

Work Package Overview

Our goal is to explore how to transition from patient needs, different contexts and technology into actionable recommendations for adaptive lifestyle interventions. This involves:

• Understanding the mechanisms driving behavior change in complex conditions.
• Aligning the intervention with principles of an adaptive intervention, using data-driven approaches to deliver personalized care
• Developing decision rules to optimize the intensity, type, and delivery of treatment based on clinical decisions and continuous patient data

To achieve this, we use a participatory research approach where patients and healthcare practitioners are actively involved in the design and development process, as well as innovative trial methods such as micro-randomized trials, and n-of-1 trials. 

What We’ve Achieved So Far

Over the past two years, we’ve laid a solid foundation through two key phases:

1.      Contextual Exploration

o   Conducted a systematic realist review to articulate the mechanisms by which adaptive interventions are developed, focusing on how they work, for whom, and under what circumstances in patients with metabolic syndrome. For this we first gathered information from stakeholders in healthcare context.

• We built a network of patients and healthcare practitioners, fostering collaboration to align with real-world needs.
• Assembled a multidisciplinary research team to ensure diverse perspectives guide the intervention’s development.

o   Explored initial technical development by creating a reinforcement learning algorithm for self-monitoring physical activity.

2.      Understanding Value Drivers

o   Conducted and worked closely with patients and healthcare practitioners to identify  core needs, preferences, and values in an adaptive lifestyle intervention.

A Holistic Approach to Development and sustained Implementation

While designing the intervention, we’ve also prioritized its long-term feasibility by:

• Investigating the business model to support its scalability and sustainability over time
• Exploring platform opportunities to ensure seamless collaboration and integration into the healthcare ecosystem.

This dual focus on personalized care and practical implementation brings us closer to delivering an intervention that can truly make a difference in patients' lives.

Current Status

The project is now transitioning into the design and testing phases. We are refining the patient and healthcare practitioner values into design requirements for the intervention.

WP 7 - University: UT-3 
Faculty: Behavioural, Management and Social Sciences

WP7 supports RECENTRE with awareness and guidance on health technology innovation and implementation. RECENTRE integrates state-of-the-art theories and models for sustainable implementation of innovations in healthcare practice. These include stakeholder involvement, user acceptance, regulations, (early-)cost effectiveness analyses and business modelling. This is facilitated by an implementation specialist who is involved in each work package to increase the potential for sustainable implementation.

During the past 2 years WP7 has explored, together with all work packages, implementation definitions, assumptions, stakeholders, and support needs. Identified follow-up activities have been documented in an implementation research strategy plan. The plan includes 3 main strategies for 2024-2025: 1) contextual inquiry to better understand current practice, 2) stimulating interdisciplinary collaboration (internally and externally to RECENTRE), and 3) needs-based activities for awareness and guidance.

So far, other work packages have appreciated the support provided by WP7 in terms of stimulating the interdisciplinary connection and collaboration and in terms of retaining the focus on innovation, implementation, and socio-economic impact, especially since it can be difficult for researchers to find time taking this into account from the very start of a research program. As such, WP7 activities have been tightly integrated with the ongoing activities and plans of the other work packages. For 2025, WP7 is focusing on intensifying external stakeholder interactions based on the preliminary results of contextual inquiry, e.g. on patient journeys, market trends and reimbursement of lifestyle interventions and eHealth technologies.