ICT4AWE 2026 Abstracts

Area 1 - Ageing Well – Social and Human Sciences Perspective

Full Papers

Paper Nr:	32
Title:	Promoting Longevity in the AI Era
Authors:	Marina Buzzi and Alice Marchetti
Abstract:	Health is a universal right. Artificial intelligence and ICT technologies are increasingly embedded in our daily lives, influencing study, work, social interactions, leisure, and health. This study explores the relationship between lifestyle and longevity, focusing on how emerging technologies can support individuals in pursuing healthier and more sustainable life paths. We aim to highlight how technology can contribute to extending healthspan and, more importantly, empower individuals to take control over their own health. Artificial intelligence applications, wearable devices, persuasive and monitoring apps and systems have been analyzed to deliver an overview of the current state of the art, in a narrative way. Enhancing health status and promoting active longevity in older adults has substantial societal implications, as it strengthens individual functional capacity while contributing to a reduction in long-term healthcare and assistance costs.
Download

Short Papers

Paper Nr:	12
Title:	Longitudinal Analysis of Sleep Patterns in Older Adults Using Wearable Technology: Insights from 12,651 Nightly Sleep Records
Authors:	Ekaterina Mashina, Jakob Kasbauer, Sebastian Wilhelm, Amir Baniassadi and Florian Wahl
Abstract:	This study utilizes wearable technology to conduct a longitudinal analysis of sleep patterns in older adults, examining a dataset of 12,651 nightly records from 54 participants. We assess sleep duration, quality, and the prevalence of social jetlag. Using the Oura Ring, comprehensive sleep data including duration, phases, and heart rate variability were collected. Average sleep duration closely aligned with recommended guidelines, and the sleep midpoint were stable, indicating consistent sleep timing among participants. Minimal social jetlag was observed, suggesting sleep schedules were influenced less by social engagements due to the senior housing facility environment. Furthermore, it was recorded that on weekends, participants slept longer than on weekdays. This is confirmed by the extension of the sleep duration range from 7.8 to 8 hours on weekdays to 7.8 to 9.2 hours on weekends. This aspect underlines the unique sleep patterns among the elderly and may indicate the importance of providing additional time for rest and recovery on weekends.
Download

Paper Nr:	14
Title:	Integrated Forecasting of Population Ageing and Earthquakes Impact for Healthcare Resources Demand Estimation
Authors:	Konstantinos Ntatis and Kostas Kolomvatsos
Abstract:	Population ageing and earthquake activity represent two different types of pressures on healthcare systems. Ageing is a predictable demographic process that gradually increases healthcare demand over time, whereas earthquakes constitute unpredictable natural events that may suddenly generate emergency healthcare needs. In this paper, we propose an integrated forecasting framework that examines demographic projections, statistical modeling of earthquake occurrence, energy-adjusted magnitude estimation and Artificial Intelligence (AI) based injury prediction to quantify future healthcare pressures. Demographic forecast data are used to estimate the expected growth in healthcare demand due to the increase of the population aged over 65. Seismic risk is modelled using a Poisson-based frequency estimator for annual earthquake rates, a customized magnitude prediction function based on energy-adjusted relations and a log-linear AI model that estimates human impact in terms of predicted injuries. Two hypothetical scenarios, mild and extreme, are generated to capture both normal and high-intensity seismic activity. By analyzing demographic ageing and earthquake-related impacts separately, the study highlights how healthcare systems must address both predictable long-term demand growth and emergency pressures. The proposed framework is applied to the case of Greece, illustrating how demographic trends and disaster risk can influence future healthcare resource requirements and emphasizing the importance of proactive planning and geographically targeted resilience strategies.
Download

Paper Nr:	15
Title:	Playing to Be Heard: Rethinking Participation in Digital Ageing through Governance-Oriented Gamification
Authors:	Cassie Xi Wang, Cath Conn, Julie Trafford, Fuwen Yang and Wuqi Qiu
Abstract:	As health and social care becomes increasingly digital-first, “participation” for older adults is often treated as a simple question: did they sign up, can they use it, and do they follow the rules? This paper argues that such a view misses what happens in everyday life. Many older people are not simply “non-compliant” or “low literacy”; they are trying to make sense of confusing systems, relying on family or community support to get things done, and sometimes stepping back when digital services feel risky, unfair, or hard to trust. If these experiences matter, then methods are needed that do more than measure uptake or performance. Older adults need ways to explain frustration, dependency, workaround, and refusal as part of living with digital-first systems. To address this gap, the paper proposes gamification not as a behaviour-change tool, but as a structured, low-pressure participatory format for sense-making and collective reflection around digital governance. Used in this way, gamification may help make visible where people get stuck, how support networks shape access, and what kinds of responsibility and control digital systems silently place on them. The paper develops this as a conceptual and methodological proposition, outlining a preliminary workflow, key design principles, and potential application scenarios rather than reporting a tested intervention or prototype. In doing so, it argues that digital participation should be understood not only as use, but as a lived and negotiated relationship with digital ageing governance.
Download

Paper Nr:	22
Title:	From Devices to Facilitation: Institutional Gaps in Digital Inclusion in Estonian Care Homes
Authors:	Marianne Paimre, Sirje Virkus and Karmen Kruus
Abstract:	Digital technologies are increasingly embedded in health and social care systems, yet older adults living in long-term care institutions remain among the most digitally excluded groups. This paper reports findings from a mixed-methods study conducted in Estonian elder care institutions, examining where and why digitalisation breaks down at the institutional level. A nationwide survey of care workers (n = 83) indicates that while staff regularly use digital systems and report good digital competence, structured and sustained support for residents’ digital engagement is limited. Complementary interviews with residents reveal curiosity and motivation to use digital tools alongside fear of making mistakes, physical limitations, and a strong need for individualised guidance. Taken together, the findings suggest that digital exclusion in care homes is less a problem of access or willingness and more a matter of missing human facilitation within institutional routines. Rather than framing digital exclusion as an individual deficit, the study highlights organisational role boundaries and support structures that leave residents’ everyday digital participation largely unsupported. The paper argues that facilitation-based approaches could help bridge this gap by embedding digital participation into everyday care practices, thereby supporting autonomy, social participation, and ageing well in institutional settings.
Download

Paper Nr:	30
Title:	Operationalisation of Cultural Sensitivity in Nursing Classification Systems: A Position Paper from the Ongoing Research Project NUTRI-SENSE
Authors:	André Heitmann-Möller, Martina Hasseler, Michael Feldhaus, Monika Schlegel, Sandra Hellmers, Andreas Hein, Steffen Busse, Julia Berndt, Mareike Förster, Rebecca Diekmann and Tobias Krahn
Abstract:	The nursing process is of high importance for the conduct of nursing services and their management. A backbone for an adequate application of the nursing process are nursing classification systems. It is unclear if these systems are cultural sensible, especially in case of the nutrition and fluid intake of care dependent persons with migration biographies like nursing home residents. To obtain data about this topic an exploratory scoping review was conducted. As result only one publication has been included, which address this issue. It shows that classification systems do not operationalize cultural sensitivity and diversity. Further research in regard of concept and content analysis are needed.
Download

Paper Nr:	34
Title:	A Novel Wellness Auriculotherapy Device for Perimenopausal Symptoms: A Web-Based User Satisfaction Survey
Authors:	Mariem Abid, Lamia Guellif and Ines Guellif
Abstract:	Perimenopause affects millions of women worldwide, often causing debilitating symptoms such as sleep disturbances, chronic stress, and musculoskeletal pain, which significantly impact quality of life. Auriculotherapy, a non-invasive neuromodulation approach, has shown promise in managing these symptoms. This study aims to evaluate user satisfaction, comfort, and the perceived efficacy of a novel wellness auriculotherapy device designed for perimenopausal women. An online satisfaction survey was conducted among an initial cohort of users to assess longitudinal physical tolerance. Data collected included adherence rates, ease of use, self-reported symptom relief across key domains (e.g., sleep quality, stress levels, and pain intensity), and user requirements for future digital integration. Preliminary findings from 42 participants indicate high user adherence and favorable ratings for device comfort and discretion. Participants reported notable improvements in anxiety, stress management, and sleep quality, along with moderate relief of physical and pelvic symptoms. Overall satisfaction scores demonstrate strong positive acceptance of the technology. Furthermore, user feedback revealed a significant demand for integrating Information and Communication Technologies (ICT), specifically requesting a companion mobile application for physiological tracking and personalized usage prompts. These results highlight the potential of the proposed wellness device as a complementary, user-friendly hardware solution for perimenopausal symptom management. Ultimately, it establishes a robust foundation for a future interconnected digital health ecosystem, contributing to improved daily functioning, autonomy, and overall well-being.
Download

Paper Nr:	42
Title:	Relationship between Staff Quality of Work, Safety Culture and Pressure Injury in Nursing Homes: A Path Analysis
Authors:	Rista Fauziningtyas, Mei-Chan Chong and Maw Pin Tan
Abstract:	Purpose: This study examines how staff quality of work and safety culture are associated with pressure injury incidence in nursing homes, providing evidence to support safer care for older adults. Methods: A cross-sectional study was conducted among 261 staff members from eleven nursing homes in Indonesia. Data were collected on staff characteristics, quality of work (QoW), safety culture, and the incidence of reported pressure injuries within the previous three months. Path analysis was used to examine direct and indirect relationships among variables, including the potential mediating role of safety culture in the relationship between staff-related factors and patient safety outcomes. Findings: Quality of work was strongly associated with safety culture. Length of employment was significantly associated with QoW, indicating that more experienced staff reported better-perceived work conditions. Safety culture was negatively associated with pressure injury incidence. Unexpectedly, QoW was positively associated with the incidence of pressure injuries. After model respecification, model fit indices reached acceptable levels. These findings should be interpreted with caution due to the cross-sectional design and potential reporting differences across facilities. Conclusion: The study clarifies how staff work environment factors are associated with patient safety outcomes in nursing homes and highlights the importance of strengthening safety culture to reduce the risk of pressure injuries.
Download

Area 2 - Digital Health

Full Papers

Paper Nr:	19
Title:	Explainable Artificial Intelligence for Identifying an Optimal IMU Sensor Configuration for Automated Mobility Assessment in Older Adults
Authors:	Kaoutar El Ghabi, Frédéric Bousefsaf, Yann Morère, Clint Hansen, Walter Maetzler and Robbin Romijnders
Abstract:	Automated mobility assessment is essential for monitoring functional decline in older adults, particularly in the presence of neurological or musculoskeletal pathologies. While inertial measurement units (IMUs) combined with deep learning enable objective mobility evaluation, sensor placement is often guided by convention rather than systematic analysis. This study proposes an eXplainable Artificial Intelligence (XAI) framework to identify an optimal and clinically applicable IMU sensor configuration for fully automated mobility assessment. A recurrent neural network (BiLSTM) was initially trained using data from 15 IMUs acquired during the Timed Up and Go test. XAI techniques were applied to quantify sensor importance and derive reduced sensor configurations. Results show that an optimal configuration composed of six sensors achieves higher accuracy, improved stability, and better generalization compared to the full configuration, while remaining clinically feasible. These findings highlight the value of XAI for principled sensor selection and support reliable automated mobility assessment in heterogeneous older adult populations.
Download

Paper Nr:	25
Title:	EmoBridge: Bridging Emotion Comprehension from Avatars to Human Faces for Individuals with Autism Spectrum Disorder
Authors:	Jingyun Wang, Adam Wynn and Rebecca Wood
Abstract:	It is challenging for autistic people to interpret and respond appropriately to social scenarios due to differences in emotion understanding and recognition abilities. This paper aims to design, implement, and evaluate an adaptive, gamified tool to support emotion recognition skill development in children with ASD. First, we propose a novel adaptive algorithm that facilitates a personalized progression from avatar to human faces according to the user’s real-time performance accuracy. Second, we present the implementation of EmoBridge, a mobile application that integrates this algorithm within a gamified learning session and includes a browse section for users to view emotion-categorized images and upload their own. Third, we report findings from a preliminary experiment involving 18 children aged 7-12, including both neurotypical and ASD-diagnosed participants, and their parents. The results demonstrate EmoBridge’s potential to support emotion recognition, particularly for verbal children with ASD, while also revealing specific limitations for non-verbal users. The feedback collected provides valuable insights that will guide future iterations and research, contributing to the development of more inclusive and adaptive assistive technologies.
Download

Paper Nr:	35
Title:	"AI Counsellor": LLM-Based Counselling Chatbot for University Students
Authors:	Saul Hewes, Sunčica Hadžidedić and Mengyisong Zhao
Abstract:	The growing demand for mental health services has led to the rapid emergence of digital mental health applications. Simultaneously, recent advances in artificial intelligence (AI) have enabled the development of high-performance chatbots powered by large language models (LLMs). Our paper explored the potential of an LLM-based mental health counselling chatbot for university students. We developed the chatbot using the QLoRA fine-tuning approach of a lightweight open-source LLM on curated counselling transcript data. We evaluated the system in a user study, with 30 university students, and found that a single chatbot counselling session reduced negative emotions significantly. Overall, our results suggested that LLM-based counselling chatbots present promising opportunities for AI-driven mental health interventions.
Download

Paper Nr:	39
Title:	Toward Breaking the Recovery Plateau in Chronic Rehabilitation: An Integrative Framework of Biodiversity Exposure, Lifestyle Intervention, and ICT-Based Monitoring
Authors:	Ryota Sakayama, Tatsuya Kawaoka and Masatoshi Funabashi
Abstract:	Functional recovery in chronic-phase rehabilitation is frequently characterised by a plateau, representing a major challenge in long-term care. While nature-based and lifestyle interventions have shown promising associations with physiological improvements, their integration with ICT-based monitoring and their applicability in resource-limited clinical settings remain underexplored. In this study, we propose an ICT-integrated framework that combines biodiversity exposure, personalised lifestyle intervention, and continuous behavioural monitoring. We report observational findings from a chronic-phase cerebrovascular accident (CVA) cohort, where improvements in motor function and physiological indicators were observed following an integrated intervention. Notably, the observed improvement in gait speed corresponds to a large effect size in a context where recovery is typically considered to plateau. To address implementation barriers such as spatial constraints and delayed biomarker feedback, we developed a streamlined model integrating compact ecological units (Syneco Portal) and an Adaptive Health Self-Assessment System (AHSAS). This framework enabled continuous, low-burden monitoring and daily ecological exposure across diverse care contexts. Rather than isolating the effects of individual components, the findings highlight measurable system-level changes emerging from the interaction of ecological, behavioural, and ICT-based processes. These results support the feasibility of an integrated ecological–behavioural–ICT framework as a practical infrastructure for adaptive rehabilitation in complex and resource-constrained environments.
Download

Short Papers

Paper Nr:	23
Title:	Predictive Modelling for Enhanced Management of Urgent and Emergency Care Services During Winter in England
Authors:	Yasin Yusuf, Abdallah Naser, Zoheir Ezziane and Ahmad Lotfi
Abstract:	This paper examines the use of predictive modelling to improve winter Urgent and Emergency Care (UEC) service management in England, a time when services experience significant operational pressure. Using historical data from major hospital emergency departments, referred to as Type 1 Accident and Emergency (A&E), patient attendances and subsequent admissions are analysed to develop forecasting models over multiple time horizons. Well-established techniques including Auto-Regressive Integrated Moving Average (ARIMA) and linear regression are employed to identify demand trends. Results demonstrate a strong and statistically significant relationship between A&E attendances and admissions, supporting the use of predictive models for resource planning and service optimisation. The feasibility of operational deployment within the National Health Service (NHS) is discussed, highlighting the importance of continuous model updating to support winter preparedness.
Download

Paper Nr:	27
Title:	HIT to Escape the Lava: A Serious Game for High-Intensity Training in Patients with Chronic Low Back Pain
Authors:	Ayesha Tahir, Rosa Lilia Segundo Díaz, Wim Ramakers, Kristof Kempeneers, Jonas Verbrugghe, Annick Timmermans and Karin Coninx
Abstract:	Chronic low back pain (CLBP) is a disabling condition that significantly limits overall functioning and quality of life. Although exercise therapy (ET) is a widely recommended treatment, patient adherence and motivation remain low, especially in home-based rehabilitation. This study presents the design, development, and initial evaluation of a gamified cycling exercise with high intensity interval training (HIIT) tailored for persons with CLBP. The game integrates cardiorespiratory interval training into a gamified environment to create a motivating and personalised exercise experience. Based on a formative expert evaluation, we conclude that the game concept provides a natural and motivating setting for a high-intensity cycling exercise. Some recommended design improvements to enhance the natural alignment of the game and the HIIT training have been identified. Further exploration of the game concept with patients is needed to determine its feasibility and effectiveness. The results of the current study will shape the design and development of digital health interventions that support motivating and enjoyable rehabilitation in and outside clinical settings.
Download

Paper Nr:	31
Title:	SOVIA: A Social Voice Assistant Architecture Based on a Large Language Model to Reduce Loneliness among Older Adults for a Participatory Field Trial
Authors:	Jakob Kasbauer, Bence Szalma, Alexander Seidl, Armin Gerl, Dietmar Jakob and Florian Wahl
Abstract:	Voice assistants are increasingly investigated as accessible technologies to support ageing in place and mitigate loneliness among older adults. However, commercial voice assistants remain passive, respond only to commands, lack conversational depth, and are unsuitable for sustained biographical dialogue. This short paper presents the technical architecture of SOVIA, a social virtual assistant that combines LLM-based dialogue with a voice interface and biographically structured conversation prompts to support ageing well at home. Based on requirements and design goals, we propose a modular architecture comprising: (i) a speech pipeline (ASR/TTS), (ii) an LLM-agnostic dialog core, (iii) a conversation orchestration service for proactive and thematic prompts, (iv) a privacy-aware user memory and summarization module, and (v) a telemetry and configuration backend. We evaluate a key conversational component, the TTS, for German-speaking older adults 65+. Finally, we describe a tablet-based deployment model for in-home use and a logging strategy to capture interaction logs for an upcoming field trial and mixed-methods evaluation.
Download

Paper Nr:	36
Title:	Agentic Conversational Agents for Mental Health: Designing a Multilingual CBT Framework for Anxiety Management
Authors:	Mouna Abdel Kefi, Onsa Lazzez, Yassine Aribi, Najla Halouani, Jihen Aloulou and Adel M. Alimi
Abstract:	Anxiety disorders remain a major global mental health concern, exacerbated by limited access to culturally and linguistically adaptive therapeutic interventions. Cognitive Behavioral Therapy (CBT)–based chatbots offer scalable solutions for mental health support; however, existing systems often lack multilingual capability, contextual awareness, and adaptive therapeutic reasoning. This paper proposes a multilingual CBT framework for anxiety management based on Agentic Retrieval-Augmented Generation (Agentic RAG). The framework integrates agentic planning, contextual retrieval, and goal-oriented dialogue generation to enable dynamic and personalized therapeutic interactions across languages. By embedding agentic autonomy within the chatbot architecture, the system demonstrates enhanced adaptability, coherence, and cross-linguistic relevance. Experimental evaluation highlights the effectiveness of the proposed approach in improving engagement and therapeutic outcomes for multilingual users. The findings underscore the potential of agentic architectures in advancing AI-driven mental health interventions, particularly for anxiety management.
Download

Paper Nr:	43
Title:	Machine Learning Regression for Time-to-First-Fall Prediction in Parkinson’s Disease
Authors:	Nouha Ed-Daoui, Younes Jabrane, Matthieu Bereau, Amir Hajjam El Hassani and Maxime Desmarets
Abstract:	Falls are a major source of morbidity and loss of autonomy in Parkinson’s disease (PD), yet predicting when the first fall will occur remains challenging. Most previous studies have focused on binary fall classification, whereas estimating the timing of fall onset may provide more informative support for preventive clinical decision-making. In this study, we investigated time-to-first-fall prediction using baseline multidimensional assessments from the Parkinson’s Progression Markers Initiative (PPMI). Only patients with sporadic PD were included, and the outcome was defined as the time interval, in months, between baseline and the first documented fall during follow-up. A regression benchmarking framework was implemented to compare several machine learning models, including tree-based ensembles, gradient boosting methods, neural networks, Gaussian processes, kernel-based regression, and regularized linear models. Model performance was evaluated using mean absolute error (MAE), median absolute error (MedAE), root mean squared error (RMSE), coefficient of determination (R2), and explained variance on the test set, with additional assessment using five-fold cross-validation on the training set. Tree-based ensemble models achieved the best results, with Random Forest providing the strongest performance (MAE = 13.94 months, RMSE = 18.29 months, R2 = 0.844). These findings suggest that baseline clinical and biomarker features contain valuable prognostic information for estimating fall timing and support the use of ensemble regression models for anticipatory fall-risk management in PD.
Download

Paper Nr:	45
Title:	Digital Technologies for Assessing CVD Risk in Menopausal Women
Authors:	Marco Manso, Barbara Guerra, Laura Patrícia Manso, Flavia Wahl, Hossam Haick, Joseph Muallem, Cristian Robledo Lete, Leopoldo Pla Sempere, Xabier Mugica Iturbe, Irati Arenzana Irazu, Cristina Martín, Adrián Otamendi and Ivan Macia
Abstract:	CVD is a leading cause of mortality among women worldwide, and the CVD risk is further exacerbated during menopausal transition. This work addresses CVD risk in menopausal women, integrating monitoring and diagnostic technologies for personal and clinical use. Conducted as part of the CARAMEL project, this work is instrumental for the development of a novel, flexible and individualised approach to CVD risk assessment, reflecting the complexity of women’s cardiovascular health across their life course. In this context, CARAMEL employs several innovative technologies to collect health and wellbeing data. Operated by clinicians, new diagnostics technologies are complemented with mobile health monitoring applications and wearable technologies, used by women on a continuous basis. The information generated by these technologies increase the quantity, scale and longitudinal depth of existing standardised high-quality data typically used in clinical studies. Further, it is in the combination of multiple data sources that lies the development, integration and validation of new personalised CVD-RA models.
Download

Paper Nr:	51
Title:	AI-Generated Health Communication to Understand How People Interpret Client Narratives: Interpretive Variability and Its Implications for Trust and Biopsychosocial Care Prioritisation
Authors:	Amanda Kearns, Anne Moorhead, Maurice Mulvenna and Raymond Bond
Abstract:	Little is known about how health communication and client narratives are interpreted across different user groups. This study examines how AI-generated videos, used to emulate client health communication, are interpreted and evaluated, with a focus on variability in care prioritisation and perceptions of AI-mediated communication. A cross-sectional online survey incorporating AI-generated, one-way video case studies was conducted (n=276; 93 health professionals and 183 members of the public). Participants were randomly assigned to one of four AI-generated video scenarios featuring a standardised core client narrative with variations in sociocultural context. Participants ranked care priorities, indicated preferences across lifestyle factors, and evaluated perceived impacts of AI-mediated communication. Across both groups, psychological and lifestyle support were consistently prioritised over medical interventions, reflecting a shared biopsychosocial orientation to care. However, interpretation was not uniform. Members of the public demonstrated greater variability in prioritisation across the different videos, suggesting sensitivity to sociocultural cues despite consistent clinical content, whereas health professionals showed more consistent response patterns. Perceptions of AI-mediated communication were characterised by ambivalence rather than rejection, with concerns centred on loss of human touch, potential miscommunication, and the need for human oversight. These findings indicate that the interpretation of standardised AI-generated videos of client narratives is not consistent across users. Instead, responses vary according to user group and contextual cues. Digital mental health systems should therefore be designed to support interpretive variability, preserve human oversight, and maintain relational aspects of care in sensitive clinical contexts.
Download

Area 3 - Telemedicine and Independent Living

Full Papers

Paper Nr:	24
Title:	Beyond Weight Loss: Exploring Wellbeing and Lifestyle Outcomes in Digital Weight Management Interventions
Authors:	Mengyisong Zhao, Suncica Hadzidedic, Jingyun Wang, Victor Elijah Adeyemo, Shauna Concannon, A. B. Sirin-Ayva and Grant Westermann
Abstract:	Obesity has been a major public health concern for decades. Digital weight management interventions (DWMIs) offer promising solutions, yet prior work has rarely assessed broader changes in wellbeing and lifestyle outcomes alongside weight management. We conducted a within-subject field study to examine changes in users’ emotional, psychological, and behavioural states across six key outcomes after 6 of the 12-week real-world weight management intervention (WMI). Results showed significant improvements in mental wellbeing at the post-study time point (SWEMWBS, W = 45.5, p = .003), while eating disorder risk also changed (SCOFF, W = 40, p = .035). In exploratory demographic analyses, negative affect reduced among participants without disabilities than those with (U = 170.5, p = .014), and participants living with someone showed a trend towards greater improvements in health promoting lifestyle (U = 128.5, p = .064). The user-centred evaluation of our platform yielded promising results for feature satisfaction and intention to reuse. These findings reflect the psychological complexity of obesity and internalised weight stigma, suggesting a nuanced pattern in which mental wellbeing improvements may occur alongside emotional challenges and behavioural changes during a WMI. Our findings further imply that next-generation DWMIs should more explicitly account for wellbeing and lifestyle factors and harness the potential of AI-based systems to better support diverse user needs and improve effectiveness, accessibility and user experience.
Download

Paper Nr:	37
Title:	Integrating Lived Experience into Footwear Innovation after Stroke: A Qualitative Study of a Multidisciplinary Workshop
Authors:	Ana Rita Pinheiro, Anabela Pinto, Ana Paula Ferreira, António Cavaleiro Henriques, Nuno Nogueira, Rita Pereira, Cátia Sampaio, Sara Faria, João Pinheiro, Patrícia Fernandes, Augusto de Sousa Coelho, Marta Peixoto and Joaquim Alvarelhão
Abstract:	Patient and public involvement has become increasingly recognized as essential to the development of effective and acceptable health technologies. In the context of stroke rehabilitation, footwear with integrated technology presents a complex design challenge due to highly individualized sensorimotor impairments and daily-life constraints. This qualitative study describes and analyzes a multidisciplinary workshop designed according to complexity-driven principles and inspired by the DREAM framework, aiming to translate lived experience into actionable footwear design requirements. The workshop brought together stroke survivors, footwear industry professionals, engineers, researchers, and healthcare practitioners in a structured environment combining theoretical input, experiential learning, practical gait and footwear analysis, and collaborative reflection. Data were collected from workshop transcripts and post-workshop questionnaires, which were analyzed using thematic analysis and semantic mapping with VOSviewer. Findings highlight the central role of comfort, safety, flexibility, thermal regulation, and adaptability in post-stroke footwear, as well as the value of early and structured user involvement in shaping realistic and meaningful design specifications. The study demonstrates that multidisciplinary, participatory workshops can support mutual learning, enhance professional empathy, and inform the development of technology-integrated footwear that better reflects the lived realities of stroke survivors.
Download

Short Papers

Paper Nr:	20
Title:	Deep Learning Models for Senior Activity Recognition Using Ambient Sensors
Authors:	Kaoutar El Ghabi, Amazigh Kridi, Frédéric Bousefsaf, Yann Morère and Olivier Habert
Abstract:	The ability to automatically recognize activities of daily living (ADLs) has significant applications in senior home assistance and safety monitoring. This paper explores deep learning techniques to identify ADL events using only ambient, non-intrusive sensors. A combination of one-dimensional convolutional neural networks (1D-CNNs) and long short-term memory (LSTM) networks with an attention mechanism is proposed. Experiments on two real-world datasets demonstrate that the model achieves state-of-the-art performance in multi-class activity classification from temporal sensor data streams. The results validate the feasibility of deep learning for unobtrusive senior activity recognition and could enable personalized assistive systems for elderly home care.
Download

Paper Nr:	21
Title:	An Explainable Artificial Intelligence Framework for Chronic Kidney Disease Prediction and Monitoring
Authors:	Khaoula Benabderrahim, Mounira Tarhouni, Salah Zidi and Najoua Bennaji
Abstract:	Robust feature selection is a critical requirement for interpretable predictive modeling in clinical decision-support systems, particularly in the presence of high-dimensional data with complex dependencies. To address this challenge, we propose a causal feature selection and prediction framework based on explicit causal structure learning. The proposed framework performs causal discovery using the LiNGAM algorithm and estimates variable importance through the analysis of causal relationships, thereby enabling the identification of clinically relevant predictors while reducing spurious associations. The selected features are subsequently used to train supervised predictive models, including multilayer perceptron, support vector machines, linear regression, and XGBoost, with hyperparameters optimized via grid search. The proposed methodology is evaluated on the task of chronic kidney disease (CKD) progression prediction. Experimental results demonstrate that the multilayer perceptron achieves the best overall performance, with an accuracy of 94.4% and an F1-score of 0.85, while XGBoost attains an area under the ROC curve of 0.88. Model interpretability is ensured through post-hoc explainability methods, including LIME and SHAP. These results indicate that causal feature selection can effectively improve both predictive performance and interpretability in clinical risk modeling.
Download

Paper Nr:	38
Title:	From Raw Sensor Streams to Clinically Meaningful ADLs: Hierarchical Semantic Tokenization and LLM-Based Interpretation for Privacy-Preserving Home Monitoring
Authors:	Yong-Hyun Lim, Homin Kang, Kangbok Seo and Soon-Ju Kang
Abstract:	ADL assessment remains central to clinical decision-making in ageing and dementia care, yet routine practice still depends largely on questionnaires and episodic observations. This creates a persistent gap between real-world function and what is documented in clinical records. At the same time, smart-home sensing and AI-based activity recognition face well-known barriers: privacy concerns, annotation cost, domain shift across homes, weak explainability, and poor long-term usability in cognitively impaired populations. This position paper argues for a pragmatic middle path between rule-based systems and end-to-end black-box classifiers: semantic tokenization of ambient sensor streams followed by large language model (LLM) interpretation. The proposed framework converts continuous sensor signals into interpretable semantic tokens, allowing an LLM to integrate spatiotemporal context and infer meaningful Basic and Instrumental ADL events. We present the conceptual architecture, implementation status (simulation-stage prototype), and a translational validation agenda. Our claim is not that LLMs should replace clinicians or caregivers, but that semantically structured AI can improve privacy-preserving home monitoring while increasing explainability and communication value in ambient assisted living.
Download

Paper Nr:	44
Title:	Mobilaines: Lessons from Co‑Designing an Age‑Friendly Trip‑Planning Tool
Authors:	Bessam Abdulrazak, Sahar Tahir, Sara Bahrampoor Givi, Mireille Gagnon-Roy and Véronique Provencher
Abstract:	Urban mobility plays a central role in supporting social participation and aging in place for older adults. However, most existing trip-planning tools remain insufficiently adapted to age-related mobility constraints, focusing primarily on travel time or distance while neglecting physical effort, environmental accessibility, and cognitive usability. This paper presents lessons learned from the Mobilaines project, a five-year Living Lab initiative conducted in Sherbrooke, Canada, aimed at developing an age-friendly trip‑planning platform through participatory co-design and algorithmic innovation. The project combined an interdisciplinary Living Lab methodology with a persona-based design framework to identify mobility needs and translate them into functional and algorithmic requirements. Technically, the platform introduces a multi-objective routing engine that integrates distance, travel time, and terrain slope. Usability evaluations demonstrated substantial acceptance and task performance, with the System Usability Scale increasing from 54.6 to 81.5 and task completion rates reaching 97%. Beyond the technical results, the project highlights key lessons regarding participatory design, data limitations in urban mobility infrastructures, and the importance of hybrid human-digital support models for inclusive mobility technologies.
Download

Paper Nr:	46
Title:	Designing Participatory Audiovisual Narratives for Disability Inclusion: A Practice-Based Qualitative Study on Adapted Sports
Authors:	Francisca Rocha Lourenço, Rita Oliveira, Oksana Tymoshchuk and Ana Patrícia Oliveira
Abstract:	This study examines how inclusive communication practices can be applied in the development of a participatory audiovisual campaign designed for digital dissemination to promote the inclusion of people with disabilities through adapted sport. Adopting a practice-based qualitative approach, the research was conducted in three main phases: i) qualitative content analysis of 15 international YouTube campaigns addressing disability inclusion and/or adapted sports; ii) fieldwork and data collection through semi-structured audio-video interviews and recording of adapted sailing and polybat activities; and iii) production of an audiovisual campaign informed by the practices identified. The analysis highlighted recurring practices such as first-person testimonials, participation-centred representation, emphasis on lived experience, clear contextualisation of initiatives, and accessibility-oriented structure, including subtitles and clear audiovisual structure. These practices informed the narrative and editing decisions of the final campaign, published on YouTube. As a preliminary validation step, the video was reviewed by three experts in disability inclusion, adapted sport, digital communication, and audiovisual production. The study links the analysis of existing digital audiovisual case studies with the practice-based development of an inclusive campaign and proposes a transferable framework for designing audiovisual content intended for digital dissemination that supports awareness, participation, and disability inclusion through adapted sport.
Download

Paper Nr:	16
Title:	Internet-of-Medical-Things Systems: A Review of Use Cases and Quality Requirements and Definition of a Multi-View Reference Architecture
Authors:	Violetta Fazzi and Claus Pahl
Abstract:	The Internet of Medical Things (IoMT) is a specialization of the Internet of Things (IoT), whose scope is using platforms, applications, and devices designed to support patients for medical purposes. IoMT is particularly used for use cases such as treating diseases, post-operative and elder care, or monitoring the patient in hospital and non-hospital settings. We provide a review of software architecture concerns of IoMT systems, looking into architectural concerns such as layers and distribution patterns and analyzing a number of medical use cases in terms of functional and quality needs. Identifying quality concerns is an objective leadin to the definition of a multi-view reference architecture incorporating quality concerns with architectural patterns.
Download