AHFE Open AccessHealthcare and Medical Devices
Editors: Jay Kalra
Topics: Healthcare and Medical Devices
ISBN: 978-1-964867-87-8
DOI: 10.54941/ahfe1007230
Table of Contents
Assessing Digital Readiness in Diagnostic and Clinical Pathology: A Human Factors Approach
The practice of diagnostic and clinical pathology (DCP) is rapidly changing given recent advancements in imaging technologies and the application of diagnostic algorithms through artificial intelligence (AI). Proficiency in digital literacy is critical for both laboratorians and diagnosticians to safely and effectively interact with these new technologies as diagnostic pathology enters the digital era. This transformation raises many important questions regarding the human factors and ergonomic considerations including workload, situation awareness, usability, and cognitive load. To date, compared with algorithm-development studies, relatively little formal research has focused on the interface between the diagnostician and AI-augmented diagnostic systems within routine practice. This paper proposes a human factors-informed digital readiness framework for diagnostic and clinical pathology that maps the interfaces among hardware, software, and end users, with particular attention to cognitive workload, usability, and alignment with existing clinical workflow practices. One critical gap that continues to emerge is that no assessment tool to determine the digital readiness of existing diagnostic pathology workflows is currently available. In the absence of such standards, this emerging field risks fragmentation and inconsistent implementation. We advocate for the development of best-practice frameworks for digital readiness that are explicitly grounded in human factors and ergonomics principles applicable to urban centres, and extensible to rural and remote healthcare environments. This framework represents a foundational model intended for prospective validation and implementation across diverse diagnostic medicine practice environments.
Jay Kalra, Bryan Johnston
Open Access
Article
Conference Proceedings
Human Factors in AI-Driven Antimicrobial Stewardship: Clinician Decision-Making, Automation Bias, and Patient Safety Risks
This research explores how human behavior and psychology impact the effectiveness of artificial intelligence within hospital programs designed to manage antibiotic use. While these digital tools aim to combat antimicrobial resistance, their success often depends on how doctors interpret and trust the technology's suggestions. The study identifies significant obstacles such as alert fatigue and automation bias, which occur when clinicians either ignore warnings or follow computer guidance too blindly. Findings suggest that making AI logic more transparent and improving the way alerts are delivered can foster better professional engagement. Ultimately, the authors argue that human-centered design is essential to ensure these technological advancements actually lead to safer prescribing habits and better patient recovery. To achieve long-term success, medical systems must prioritize the interaction between clinicians and software during both the development and implementation phases.
Mahdi Marzi, Ayşe Karacalı Tunç, Şebnem MARZİ
Open Access
Article
Conference Proceedings
Assessing Hospital Patient Nutrient Intake with an AI-Powered Food Recognition System – A Feasibility Study of the FlavoriaFlex solution
Adequate dietary intake is essential for positive clinical outcomes of hospitalized patients, yet monitoring food intake is labor-intensive and often subjective. AI-based food recognition could automate monitoring and assessment, but evidence in real-world hospital settings is limited. This study evaluated an AI-powered food recognition system, FlavoriaFlex, to assess its detection performance, deployment feasibility, and acceptability among dietitians. Previously validated in restaurant (F1 0.75, weight MAE 23.6 g, energy MAE 235 kcal), the system was deployed in a hospital ward for six days. A total of 133 meals were recorded; 102 had paired leftover images (235 total images). Manual annotation of 483 food segments provided ground truth for evaluating food recognition and menu mapping. Semi-structured interviews with dietitians assessed usability, perceived benefits, and clinical value. FlavoriaFlex enabled automatic estimation of item- and meal-level consumption, including weights and energy- and macronutrient contents. Overall food recognition accuracy was 94% (F1 0.76), remaining high for served meals (96.5%, F1 0.85) and robust for visually complex leftovers (89.5%, F1 0.71). Unknown/non-food segments were minimal (2.4% of leftovers; 0.27% of weight). A web dashboard delivered real-time visualizations, including energy and nutrient intake. Dietitians reported reduced cognitive burden, more objective assessment, and improved observability into patient dietary intake, while emphasizing the need for further validation and integration for clinical use. These findings demonstrate that FlavoriaFlex could be integrated into hospital workflows to provide accurate, clinically meaningful intake estimates, with AI-assisted food recognition offering an efficient, reliable approach to improving nutritional monitoring at scale.
Rehan Khalil, Sanna Koskimäki, Hanna Lähde, Shyam Bhetuwal, Lauri Koivunen, Veera Houttu, Kirsi Laitinen, Tuomas Mäkilä
Open Access
Article
Conference Proceedings
When One in a Million Matters: Developing Metrics for Human-AI Collaboration in Rare Disease Diagnosis
Rare diseases affect approximately 300 million people worldwide, yet physicians rarely encounter individual conditions, creating significant risk of delayed or missed diagnosis. AI diagnostic tools offer potential to reduce this uncertainty. However, metrics for assessing Human-AI Collaboration (HAIC) quality in this context remain underexplored — existing frameworks lack empirical operationalisation for the Human-Centric collaboration mode, where the physician retains full decision-making authority. This study aims to operationalise quality metrics for Human-Centric HAIC within rare neuromuscular disease diagnosis by exploring neurologists’ experiences with a conversational AI diagnostic assistant. An exploratory qualitative design employing thematic analysis is planned. Semi-structured interviews following a critical incident technique protocol will be conducted with 10–12 neurologists. Questions address ten collaboration quality dimensions: clarity of communication, ease of use, user satisfaction, feedback frequency, teaching efficiency, error reduction rate, task completion time, confidence, trust score, and safety incidents. Thematic analysis will identify context-specific subdimensions of each metric. The study will lay the groundwork for a domain-specific HAIC assessment instrument, provide design recommendations for clinical AI systems, and establish a basis for future psychometric validation and research on AI adoption in healthcare.
Anna Bilyk
Open Access
Article
Conference Proceedings
NecKorrect: Customisation Ergonomic Interventions for Cervical Spine Health
Pillow design plays a critical role in spinal alignment and sleep quality. This study introduces NecKorrect, a data-driven ergonomic intervention designed to optimise cervical alignment through personalised curvature. Using a hybrid 3D and 2D anthropometric approach, we developed customised pillows and evaluated them through subjective comfort assessments and objective MRI-based biomechanical analysis. Results from five participants indicate that pillows maintaining a cervical Cobb angle of 7°–9° were associated with the highest comfort ratings for most users, though individual anatomical variations necessitated tailored interventions. A nonlinear relationship was observed between pillow slope and spinal correction, revealing a mechanical saturation threshold beyond which overcorrection may induce compensatory flexion. This research reveals the importance of personalised ergonomic solutions and provides a scalable framework for combining biomechanical correction with user subjective experience in pillow design.
Qin Du, Junjian Chen, Yuqian Wang, Linlin Feng, Yan Luximon
Open Access
Article
Conference Proceedings
Modular Organ Aging Framework in the Real World: Cost, Frequency, Equity, and a Patient-Facing Calculator
Chronological age reflects time since birth, whereas biological ageing captures heterogeneous decline across cells, tissues, and organs. Composite biological age integrates biomarkers across organ systems, yet a feasibility-first pathway to translate organ-age theory into real-world monitoring and communication remains lacking. We present an implementation-oriented toolkit that classifies candidate biomarkers across cardiovascular, neurologic, pulmonary, renal, hepatic, musculoskeletal, immune, endocrine, and integumentary domains into functional ageing tests (e.g., grip strength, gait speed, reaction time, sleep architecture) and physiological markers (e.g., MRI-based brain volume, IGF-1, cystatin C). To enhance interpretability, we introduce a fast- versus slow-ageing organ framework based on deviations from age-referenced norms. Crucially, all measures are organized into feasibility tiers: Tier 0 (low-cost, at-home tests), Tier 1 (routine, low-cost laboratory tests), Tier 2 (specialist-administered assessments), and Tier 3 (send-out or referral laboratory tests). We summarize accessibility, cost, recommended testing intervals, expected biological drift to reduce over-interpretation, and population-specific reference ranges. The proposed application enables individuals to input routine laboratory and functional data to visualize organ-specific trajectories and composite biological age for self-tracking and research purposes only, supported by human factors such as clear design principles and explicit clinical disclaimers.
Pramath Kakodkar, Nooshin Shekari, Jay Kalra, Tareq Ahram
Open Access
Article
Conference Proceedings
Automated Hemostasis in Limb Trauma: FEM Insights for Tourniquet Optimisation
Haemorrhage is the leading cause of preventable death in military operations, often resulting from explosions, gunshots, stabbings, or cuts. Limbs are particularly vulnerable, accounting for nearly 55% of combat-related injuries, and rapid blood loss can lead to over half of deaths within minutes and almost one-third within hours. These statistics highlight the urgent need for rapid and reliable haemorrhage control, particularly in situations where immediate human intervention may be delayed or impossible.To address this critical problem, an automated primary hemostasis system was developed to arrest bleeding without human assistance. The design of its signal-triggering mechanism was informed by finite element method (FEM) simulations, which provided a quantitative understanding of tissue mechanics under tourniquet application. Human soft tissues exhibit complex behaviour due to their multilayered structure, heterogeneity, anisotropy, and viscoelasticity, necessitating nonlinear constitutive models. In this work, skin, subcutaneous adipose tissue, and muscle were modelled as nearly incompressible, isotropic hyperelastic materials using a Neo-Hookean formulation, whereas bones were considered rigid. The Neo-Hookean model captures tissue response through two key parameters: the shear modulus, which governs deformation under shear, and the bulk modulus, which defines resistance to volumetric compression.A detailed three-dimensional model of the upper limb was reconstructed from biomedical imaging data, including skin, adipose tissue, muscle, and humeral bone. The discretised model comprised over 850,000 tetrahedral elements, with material properties assigned according to tissue type and scaled to anatomical geometry. Boundary conditions permitted free displacement of soft tissues, while bone nodes were constrained axially. Tourniquet application was simulated by imposing prescribed displacement fields derived from three-dimensional indentation maps. To ensure numerical stability in the presence of nonlinear tissue behaviour, compression was applied incrementally using a conservative predictor strategy.Simulation results revealed striking patterns of tissue deformation. Maximum displacement occurred at the tourniquet centre, predominantly in the radial direction, while axial displacement dominated regions outside the compressed zone. Muscle surface strain increased progressively with compression intensity and exhibited a diffuse spatial distribution, whereas stress concentrations were localised at tissue interfaces. These mechanical patterns were interpreted in the context of physiological blood flow, both at rest and during intense exertion. Elevated arterial pressure and increased flow rates accelerate haemorrhage, shortening the window for effective intervention and emphasising the importance of timely hemostatic control.Integrating FEM-based tissue mechanics with hemodynamic considerations provides critical insights into the efficiency of haemorrhage control strategies. This approach not only informs the design and optimisation of tourniquets but also guides the development of automated hemostasis systems capable of rapid deployment in high-risk scenarios. By linking quantitative tissue deformation with clinical outcomes, the study offers a powerful framework for improving survival rates in military and civilian trauma situations, highlighting how biomechanical modelling can directly influence life-saving interventions.
Emilia Visileanu, Adrian Salistean, Alexandra Gabriela Ene, Radu Herzog, Razvan Scarlat
Open Access
Article
Conference Proceedings
Individual Performance Analytics in a Virtual Reality Simulation for Medication and Medical Supply Storage: An Experience Report
Safe storage of medications and medical supplies is a core requirement of patient safety. At the Sanitätsakademie der Bundeswehr (Bundeswehr Medical Academy), a deliberately error-seeded “small material distribution unit” (skill lab) is used to illustrate theoretically taught content through practical scenarios. These scenarios address issues such as expired products, improper storage, damaged packaging, and inadequate access control. Inspections are typically conducted in groups, which limits insight into individual competence. A realistic virtual reality (VR) simulation of the storage environment was therefore developed to capture individual performance analytics. The simulation reproduces the skill lab environment in detail and randomises material-related defects and room-level safety checks per run. Its design is informed by findings indicating that high realism in immersive VR enhances learning. An instructor-facing launcher allows training staff to define the difficulty level and configure error types and quantities. At the beginning of a six-week training programme, twelve participants completed a run in the Easy mode, and nine of them returned near the end of the training for a run in the Hard mode. At the end of each run, a report was generated which documented the difficulty level, total number of errors, number and proportion of corrected errors, results of room-configuration checks, and for each item the status “correctly disposed”, “incorrectly disposed”, or “overlooked”. Based on this data, descriptive metrics and illustrative individual performance profiles were derived. This experience report indicates that a VR simulation can meaningfully complement group-based training by providing graded scenarios and individual analytics.
Rohan Saxena, Andrei Darii, Andrei Florea, Marian Sorin Nistor, Angela Huße, Lars Schneidereit, Stefan Pickl
Open Access
Article
Conference Proceedings
An Ergonomic Perspective on Cortisol, Cardiovascular Risk, and Anxiety in Full- Time Faculty Workers
Human Systems Integration (HSI) and Organizational Ergonomics provide a robust framework to analyze the modern university as a complex socio-technical system. This paper expands upon the findings of a baseline study involving 90 university faculty members in Cartagena, Colombia (Alayón et al., 2025), shifting the focus from individual pathology to systemic ergonomic feedback. The data revealed that state anxiety is significantly associated with a failure in the physiological decline of cortisol at 4:00 PM, as well as with elevated triglyceride levels. From an HSI perspective, these biomarkers are identified as proactive ergonomic indicators of "systemic strain" and a lack of adequate "recovery windows" within the organizational interface. The persistence of evening cortisol is discussed as a failure in psychological detachment processes, likely resulting from high administrative workloads and the "porosity" of contemporary academic roles. Based on these indicators, the paper proposes systemic interventions grounded in the SEIPS 2.0 and Compensatory Control models, specifically the implementation of circadian-aligned work scheduling and the reduction of "cognitive friction" in bureaucratic procedures to mitigate long-term allostatic load. The study concludes that the integration of biological and psychological markers into organizational monitoring allows institutions to identify misalignments between system demands and human biological constraints. This proactive approach ensures the long-term sustainability of the academic system by protecting its most critical component: the human operator.
Alicia Alayón, Nohora Ochoa Arizal, Jose López Toro, Manuel Noreña Correa, Francisco Hernández Rojas
Open Access
Article
Conference Proceedings
Exploring Patient Safety Awareness and Risk Perception Among Clinical Staff and Inpatients
Patient safety is a critical component of healthcare quality, emphasizing the prevention of avoidable human- and system-level errors throughout the care continuum. This study employed a structured questionnaire to evaluate patient safety awareness and risk perception among healthcare professionals, patients, and family members. A total of 125 valid responses from healthcare professionals and 161 responses from patients and family members were included in the analysis. Healthcare professionals most frequently reported patient safety incidents involving patient falls (58.4%), tube dislodgement (53.6%), and medication errors (40.8%). Workload assessments indicated that temporal demand and effort were rated highest, implying that increased multitasking requirements and elevated workload intensity may heighten vulnerability to error. In contrast, patients and family members demonstrated heightened awareness of risks related to falls and infections but expressed substantially lower concern regarding medication errors. This discrepancy highlights a noteworthy perceptual divergence between frontline clinical staff and care recipients concerning patient safety priorities. Overall, the findings underscore that patient safety is shaped by the interaction of human factors and organizational systems. Incorporating user-centered interface design, strengthening patient engagement strategies, and integrating ergonomic principles into clinical workflows may contribute to reducing preventable medication errors and fostering a more robust, collaborative culture of safety within healthcare organizations.
Dao-yuan Wang, Yen-Hui Lin
Open Access
Article
Conference Proceedings
VR Games as a Complementary Tool for Upper Limb Rehabilitation: A Biomechanical and Usability Analysis
This study presents an exploratory analysis of the effectiveness and applicability of a virtual reality (VR) exercise system as a complementary tool for shoulder physiotherapy rehabilitation, aiming to evaluate its usability and potential integration into clinical protocols. The investigation involves four healthy participants performing a 4-minute "easy" level protocol on the XRWorkout platform, using an IMU-based motion capture suit for detailed kinematic analysis. A biomechanical evaluation was conducted by comparing participants’ joint angles with established theoretical reference ranges. Results showed high adherence to prescribed motor trajectories, with Time Out of Range (TOR) consistently below 0.3%, indicating that movements remained predominantly within safe therapeutic boundaries. While kinematic profiles were regular, recorded range of motion was generally below theoretical maximums and inter-shoulder asymmetry was noted. It is concluded that the system demonstrates significant potential as a complementary and engagement-enhancing tool for conventional shoulder physiotherapy, providing gamified, quantifiable exercises during stages of functional recovery.
Renata Vieira Do Rosário, Alexandre Anibal Campos Bonilla, Flávio Anthero Nunes Vianna dos Santos
Open Access
Article
Conference Proceedings
Therapeutic Applications of Remote Aviation (T.A.R.A.): A Neuroergonomic Framework for Aerially Mediated Cognitive-Affective Modulation
Recent advances in Neuroergonomics, Affective Computing, and Remote Aviation have begun to reveal how technologically mediated sensorimotor engagement can reshape cognitive–affective regulation. Building on these convergences, this paper introduces T.A.R.A. (Therapeutic Applications of Remote Aviation) as a neuroergonomic therapeutic framework that reconceptualizes small unmanned aerial systems (sUAS) into interfaces for cognitive–emotional modulation. While prior work has primarily examined drone operation in the context of performance optimization, comparatively little attention has been given to their therapeutic potential, particularly their capacity to facilitate emotional reframing through visuospatial control.T.A.R.A. is a multi-layered system architecture that establishes an adaptive cognitive–emotional environment in which drones serve as “distance-regulated surrogates,” enabling the externalization of emotional state. A central feature of the system is the Aerial Biofeedback Loop, which infers autonomic and cognitive states through physiological indices such as lightweight EEG signals, and dynamically modulates flight parameters to guide users toward psychophysiological coherence. Rather than directly intervening at the level of the operator, T.A.R.A. achieves regulation indirectly by shaping the interaction environment itself. As such, T.A.R.A. positions itself not as a discrete intervention, but as a novel, testable paradigm that enables therapeutic transformation through the repurposing of Human machine interactions.
Suvipra Singh
Open Access
Article
Conference Proceedings
Effects of an Electric Drive Wheel on Hand Force, Body Posture and Perceived Exertion During Hospital Bed Transport by Nursing
Studies show a high prevalence of back, shoulder and neck pain among nurses. Moving hospital beds is one of the most demanding tasks for healthcare workers, especially due to heavy weights and long distances. To reduce the physical burden on healthcare workers during bed transport, a novel electric drive wheel has been developed. This electric drive wheel replaces one of the four outer castors of a hospital bed. Existing hospital beds can be retrofitted using a plug-and-play approach. The aim of this study was to investigate its impact on the action force, body posture and perceived exertion when moving hospital beds. 13 nurses (5 male, 8 female) moved a standardised hospital bed through a realistic course that included ramps, curves, corridors and an elevator. The bed was moved with and without the drive wheel. Hand action forces were measured using 3D force measurement grips. Body posture was recorded using a Xsens motion analysis system. The subjective perception of strain was analysed using the Borg CR10 scale. In addition, t-tests and Wilcoxon tests were performed to analyse significant differences in peak and mean values between the bed configurations (manual, assist).The drive wheel significantly reduced physical load, lowering hand forces by 22% and trunk inclination by 18%. The highest reductions up to 45% occurred during start phase and when pushing the beds on ramps. Test participants rated the physical workload as “severe” to “very severe”. With the powered castor the physical workload decreased by 69% as “slight”. Hospital beds equipped with electric drive wheels can help reduce physical strain in everyday clinical care during pushing and pulling, especially in high-strain situations such as navigating ramps. However, not all recommended ergonomic limits were met, even with driving assistance.
Alina Bola, Niels Hinricher, Chris Schröer, Claus Backhaus
Open Access
Article
Conference Proceedings
Comparing Human- and Machine-Guided Virtual Reality Training: The Role of Physiological Stress in Learning Outcomes
Virtual reality (VR) is increasingly used in medical education due to its capacity to provide immersive, standardized, and scalable training environments, yet direct comparisons between human-guided and machine-guided VR instruction remain limited, particularly regarding the role of learners’ physiological responses. The present study compared human-guided (IG) and machine-guided (MG) VR-based Basic Life Support (BLS) training and examined whether physiological stress responses during training and examination phases moderated learning outcomes. Fifty-five undergraduate students completed a VR BLS session under either IG (N = 25) or MG (N = 30), followed by a VR-based exam assessing learning outcomes. Electrodermal activity (EDA) was recorded continuously as an index of physiological stress, with mean EDA values computed separately for the training and exam phases, and participants also reported their sense of presence in the virtual environment. Independent samples t-tests indicated no significant group differences in physiological stress during either the training or exam phases, suggesting comparable levels of physiological activation across instructional modalities. In contrast, participants in IG reported a significantly higher sense of presence than those in MG. An ANCOVA controlling for presence and stress levels during both phases revealed a significant main effect of instructional group on exam performance, with participants in IG achieving higher scores than those in MG, while none of the covariates significantly predicted performance. These findings indicate that the benefits of human-guided VR instruction extend beyond differences in average physiological arousal or subjective presence.
Dilan Çabuk Çolak, Mehmet Emin Aksoy, Dilek Kitapcioglu, Tuba Usseli, Hayrettin Can Südor
Open Access
Article
Conference Proceedings
From Design to Trial: Understanding lived experience and the role of SHAPES in Home-Based Therapy for Post-Stroke Elbow Spasticity
This research explores the lived experience of using the ShefStim APS device and a new form of transcutaneous electrical stimulation (TENS) known as Sheffield Adaptive Patterned Electrical Stimulation (SHAPES) for post stroke elbow spasticity (PSES). Spasticity is a common outcome following a stroke, leading to stiffness, discomfort, fatigue, and reduced upper limb mobility. There is a need for early, accessible, and cost-effective treatments. The ShefStim APS is a small, wearable, battery powered stimulator secured to the upper arm using a bespoke sleeve, designed to activate sensory nerves through TENS and SHAPES and reduce PSES. A partially double blind Randomised Controlled Trial (RCT) is underway to evaluate efficacy and cost effectiveness. As part of the study, experiential semi-structured interviews have been conducted to understand ease of use and acceptance of the device in real world home settings. Thematic analysis of 15 interviews indicates broadly positive user experiences, with participants reporting comfortable stimulation sensations, straightforward device operation, and acceptable wearability of the sleeve. Participants insights suggest areas for further refinement, particularly improving donning and doffing for one handed use, offering greater variation in sleeve sizing, and optimising the hydrogel interface to support easier placement and removal. If the final RCT outcomes reinforce these findings, addressing these ergonomic considerations should enhance independent use, reduce reliance on caregivers, and improve overall user experience and adoption.
Louise Moody, Mark Reeves, Avril D. McCarthy, Jamie Healey, Ali Ali, Wendy Tindale, Krishnan Padmakumari Sivaraman Nair
Open Access
Article
Conference Proceedings
Improving outcomes in patients' refractory to standard therapies: Dual-Mechanism to Mitigate Antigen Escape for Bridge-to-Transplant Immunotherapy
Refractory patients who fail standard therapies have limited options, and NK92-, CD16+NK92-, and CAR-NK92–based approaches are being explored to prolong survival and serve as bridging therapy to transplantation. Because NK92 cells are infused after irradiation, their in vivo persistence and activity are short-lived. We will generate CD16-expressing NK92 cells to enhance cytotoxic potency and broaden applicability by enabling tumor-associated antigen (TAA)–directed, monoclonal antibody–dependent cellular cytotoxicity (ADCC). In parallel, we will engineer “ASSASSIN” cells (CD16+IL12+NK92) and introduce BCMA-specific CARs to create CAR ASSASSIN (CAR+CD16+IL12+NK92) cells capable of dual/bispecific killing via CAR recognition and CD16-driven ADCC. This combinatorial targeting is designed to reduce antigen-escape–mediated loss of efficacy, a key limitation of conventional CAR therapies. To promote more durable immune control despite the transient lifespan of irradiated NK92 cells, IL-12 secretion is incorporated to stimulate the host immune system, aiming to sustain antitumor activity after infused cells disappear. Overall, this study proposes a ready-to-use, IL-12–secreting NK92 platform that can be redirected to multiple tumor antigens within a bridging-therapy framework and may support the development of standardized, flexible, and clinically actionable treatment approaches for time-sensitive therapeutic settings.
Derya Di̇lek Kançaği, Cihan Taştan, Selen Abanuz, Didem Çakırsoy, Gözde Sır Karakuş, Bulut Yurtsever, Utku Seyis, Raife Dilek Turan, Ömür Selin Günaydın, Gamze Tumentemur, Samed Özer, Cansu Hemşinlioğlu, Cavit Kerem Kayhan, İlham Gafarlı, Koray Yalçın, Ümit İnce, Siret RATİP, Ercüment Ovalı
Open Access
Article
Conference Proceedings
Design and Workflow in Single-Family NICU Rooms: Examining Flow Disruptions in Infant Care
Neonatal intensive care units (NICUs) are uniquely complex settings, where families and clinicians navigate workflows shaped by sophisticated medical technology and safety risks surrounding fragile infants. To enhance developmental care, many NICUs have transitioned toward the Single-Family Room (SFR) design model, treating patients in private rooms. Compared to open bay layouts, SFRs have shown increased infection control and family participation in care. Despite these benefits, relatively little is known about how SFR design characteristics introduce safety risks to infant care, especially considering the increased family involvement in care tasks such as Kangaroo care and infant feeding. These tasks can be cognitively and physically demanding, especially for families, and carry substantial risk of infection, traumatic injury and medical errors. While design may potentially minimize flow disruptions and risks in healthcare tasks, evidence is scarce from the built environment perspective, especially in the NICU. We conducted a study to identify how specific SFR features such as room size, layout, and equipment/furniture design can influence disruptions in the flow of infant care tasks jointly performed by families and staff. Using an exploratory approach, we conducted two online focus groups and a case study involving field observations and interviews, thematically analysing the data. Findings revealed 11 types of design-related flow disruptions in the SFR, categorized as high and low risk. Flow disruptions were linked to SFR layout, storage and surface organization, infant bed clearances, cord management, and furniture/equipment design. Findings also revealed physical mechanisms involved in these flow disruptions, laying foundation for future research.
Herminia Machry, Lorena Muzel Gomes, Marzia Chowdhury
Open Access
Article
Conference Proceedings
Designing for Therapy: An Ergonomic Hand Orthosis that Enhances Functional Recovery
This project presents the design and development of an ergonomic, user-centered hand, wrist, and finger orthosis to support rehabilitation in pediatric users and stroke patients with neurological impairments. Conditions such as stroke and pediatric motor disorders are often associated with altered muscle tone, spasticity, and impaired neuromuscular control, limiting voluntary hand opening, extension, and functional grasp. These patients frequently exhibit flexor dominance and reduced active extension, making reach, release, and object manipulation challenging. There is a growing need for pediatric-specific orthotic solutions that address differences in hand anatomy, growth, activity patterns, and therapy engagement. The project began with a review of literature on neurological spasticity, pediatric rehabilitation, and existing orthotic devices, supplemented by discussions with therapists to understand functional limitations, positioning requirements, and safety considerations. Key design requirements included ergonomic alignment, adjustability, lightweight construction, comfort, and ease of application. Multiple concepts were developed through sketching and CAD modeling to ensure anatomical alignment of the wrist and metacarpophalangeal joints. Prototypes were fabricated using lightweight, skin-friendly materials and evaluated for fit, comfort, and range-of-motion support. The final design provides controlled extension assistance to counter flexor dominance while enabling functional grasp and release. Iterative refinement improved strap placement, joint mechanics, and structural stability, resulting in a prototype that demonstrates improved positioning, adjustability for growth, and enhanced comfort for rehabilitation use.
Sejal Shah, Jeff Feng
Open Access
Article
Conference Proceedings
A spatial language abilities assessment model for diagnosing Chinese speakers in the prodromal stage of dementia
This study investigates spatial language as a clinical biomarker for prodromal dementia in Chinese-speaking populations. A Virtual Reality (VR) assessment model was developed to evaluate spatial fluency, reference frame naming, and spatial memory in individuals with Mild Cognitive Impairment (MCI) versus healthy controls. The results demonstrate an exceptionally high correlation (r = 0.9431) between spatial language performance and the MoCA scale, validating the model’s diagnostic accuracy. Key findings indicate that while Static Spatial Relationship Expression (SRE-S) is a reliable indicator of overall cognitive stability, Allocentric Spatial Language Memorability (ASLM) exhibits the largest performance gap between groups. It confirms that the inability to linguistically encode environmental maps from a non-self perspective is a primary hallmark of early neurodegeneration. By establishing clear scoring norms, this research provides a non-invasive method for early detection and clinical intervention in dementia care among Chinese-speaking populations.
Cheng-Li Liu
Open Access
Article
Conference Proceedings
Skin Temperature Dynamics during Sleep Onset Latency under Different Ambient Temperatures
The bedroom thermal environment is a key determinant of sleep onset, yet most studies manipulate ambient temperature and evaluate sleep onset latency (SOL) between conditions. Continuous skin-temperature dynamics during the SOL, particularly across bedroom temperatures, remain poorly characterised. This study examined skin-temperature dynamics before objectively defined sleep onset across multiple bedroom temperatures. Twelve healthy adults (6 males, 6 females; 23–42 years; BMI 17.7–32.9 kg/m²) completed four overnight sessions at 22°C, 24°C, 26°C, and 28°C. From 23:00 lights-out to 07:00 lights-on, skin temperatures at nine sites were recorded every 30 seconds, while polysomnography was continuously monitored. Sleep stages were scored according to the American Academy of Sleep Medicine manual, and the first non-wake epoch defined sleep onset. Temperature series were aligned to this point, and distal (DST), proximal (PST), and mean skin temperatures (MST) were derived. The distal–proximal gradient (DPG) was calculated as the difference between distal and proximal temperatures. Linear mixed-effects models with ambient temperature and SOL segment (early: −30 to −15 min; late: −15 to 0 min) as fixed factors showed that ambient temperature strongly affected DST, MST, and DPG in the early period of SOL, but effects were markedly attenuated in the late period. Across all four bedroom temperatures, DST, MST, and DPG converged toward similar levels as sleep onset approached. These findings suggest that the body is not passively constrained by ambient temperature but actively adjusts skin temperature and heat dissipation through distal thermoregulation to reach a relatively stable sleep-conducive state.
Shengnan Liu, Jiewei Li, Jingya Zheng, Siyuan Guo, Weidan Sun, Yuan Zhao, Hengyue Zhang, Fanglai Yao, Fujun Zhang, Wenze Chen, DING Li
Open Access
Article
Conference Proceedings
Bridging the Gap: Toward a Unified Framework for Transparent and Patient-Centered Error Disclosure
Medical error remains a major global patient-safety concern and continues to contribute substantially to morbidity and mortality. Despite sustained international efforts to improve healthcare quality, disclosure of medical errors to patients remains inconsistent and operationally challenging. Although widely accepted as an ethical obligation and a cornerstone of patient and family-centered care, disclosure practices vary considerably across healthcare systems. We have previously described Canadian provincial initiatives promoting open disclosure and advocated for their integration into a no-fault framework. The objective of this study is to conduct a systematic comparative analysis of medical error disclosure policies across Canadian provincial and territorial health authorities and to propose a best-practice medical error disclosure model. Using a structured policy review, we evaluated existing frameworks across five indicators: timeliness and accuracy of communication; the presence of a supportive, non-punitive institutional culture; availability of formal education and training; and coordinated, team-based involvement in disclosure. While most jurisdictions endorse transparent disclosure in principle, substantial variability exists in training requirements, team coordination, and enforcement mechanisms. These inconsistencies contribute to persistent provider uncertainty driven by inadequate preparation, medico-legal concerns, fear of damaging therapeutic relationships, and organizational cultures that do not consistently support transparency. Our findings highlight disclosure as a complex, iterative process requiring alignment of ethical principles, communication strategies, patient safety practices, and institutional support. In the absence of a national framework, we recommend the development of a patient-centered, non-punitive disclosure policy embedded within the standard of care.
Zoher Rafid-Hamed, Bryan Johnston, Anjali Saxena, Jay Kalra
Open Access
Article
Conference Proceedings
Work Related Musculoskeletal Injury Rate and Ergonomics in Endourology: A Global Survey
Work-related musculoskeletal injuries (MSI) are frequent among urologists but often underrecognized, affecting both personal wellbeing and healthcare system efficiency. This study aimed to investigate the prevalence of work-related MSI and the correlation with anthropometric and work characteristics in endourology. Fifty-five endoscopists and nurses participated in a global online survey containing questions about anthropometrics, demographics, work and procedure characteristics. Work-related MSI were reported by 49% of respondents. No significant correlation was found between demographic, anthropometric or procedure characteristics and the occurrence of injuries (p>0.05). The most frequent affected areas were lower and upper back, wrist, and thumb. Prolonged procedural posture and repetitive movements were identified as the main contributing factors. Urologists performing more flexible than rigid or semi-rigid endoscopy reported a higher rate of upper back pain, while those sitting during the procedure reported a lower incidence of upper back MSI (p<0.05). 9% of participants had received ergonomic training. Limitations included the small number of nurses (n=2) and potential response bias. Our findings align with earlier studies, confirming a high work-related MSI prevalence among the urologists. This highlights the importance of implementing preventing measures before, during and after the procedures. Given the professional and economic consequences of work-related MSI, promoting ergonomic awareness is essential. Performing the procedure while sitting, use of two-piece lead aprons, pre-procedural warmups, and using light weight endoscope could improve urologist’s ergonomics and should be targeted in ergonomic training programs.
Veronica Bessone
Open Access
Article
Conference Proceedings
Handgrip Strength and Anthropometric Characteristics of Children and Adolescents in Türkiye
Muscle strength capacity and anthropometric characteristics vary across populations, limiting the direct applicability of global ergonomic design standards to local user groups. This study aimed to establish normative reference values for maximal handgrip strength (HGS) and selected anthropometric measures in Turkish children and adolescents aged 6-17 years. A pilot study involving 120 participants was conducted to determine the required sample size for the main cross-sectional study, which included 459 healthy volunteers (232 girls and 227 boys) residing in İstanbul. HGS was measured for both dominant and non-dominant hands using a calibrated digital Jamar dynamometer in accordance with the American Society of Hand Therapists protocol. Height, body mass, hand length, and hand width were also recorded. Statistical analyses (ANOVA, t-tests, and correlation analysis) showed that HGS increased with age in both genders, with boys demonstrating higher strength values across all age groups, except 12-year age group. Grip strength was also strongly associated with height, body mass, and hand dimensions. These findings provide population-specific reference values that can support the ergonomic design of products and environments intended for children and adolescents in Türkiye.
Özge Coşkun, Mahmut Ekşioğlu
Open Access
Article
Conference Proceedings
Human-Centered Sepsis Management in Clinical Work Systems: A Socio-Technical AI Framework for Patient Safety
Sepsis is a time-critical condition associated with substantial morbidity and mortality, where delays in recognition and treatment markedly worsen outcomes. Although machine learning models show promise for early detection, their clinical impact has been constrained by poor integration into workflows, limited interpretability, and insufficient support for coordinated action. This study introduces a human-centered, systems-based agentic AI architecture for sepsis risk modeling and proactive clinical management. Rather than generating static risk scores, the system continuously interprets evolving patient data, situates risk within the clinical workflow, and supports timely, clinician-supervised interventions. Grounded in systems engineering and guided by the Systems Engineering Initiative for Patient Safety (SEIPS) framework, the architecture embeds predictive intelligence within the broader socio-technical work system, enabling closed-loop monitoring, coordination of safety-critical tasks, and feedback-driven adaptation. By reframing sepsis prediction as an adaptive, workflow-aware safety intervention, this approach advances AI from passive decision support toward an accountable, action-oriented partner in care delivery while preserving clinician oversight.
Firda Rahmadani, Mecit Can Emre Simsekler, Siddiq Anwar
Open Access
Article
Conference Proceedings
RE-TEST THEM: A Human Factors Pilot Study to De-Risk the Adoption of Breath-Based Therapeutic Monitoring in Oesophagogastric Cancer
Oesophagogastric (OG) cancer remains a cancer of unmet need, with poor survival driven by delayed diagnosis, aggressive tumour biology, and inherent treatment resistance. Analysis of exhaled breath volatile organic compounds (VOCs) offers a promising, non-invasive approach to characterising tumour biology and identifying potential therapeutic vulnerabilities. Within the RE-TEST THEM (bREath TESTing for THErapeutic Monitoring) project, a breath test has been developed to identify patients who may benefit from experimental therapies such as PARP inhibitors. However, its clinical value depends on its integration into clinical decision-making rather than analytical performance alone.This exploratory pilot study applied a human factors approach to examine the usability, acceptability, and decision-making implications of the breath test. Semi-structured interviews were conducted with eleven patients, alongside a vignette-based decision-making exercise with four consultant oncologists using four realistic scenarios covering curative and palliative pathways.Findings demonstrated a clear division in perceived utility between clinical settings. In the curative pathway, clinicians were highly cautious, requiring strong evidence and clear thresholds before allowing test results to influence potentially curative care. For palliative patients, the test was viewed as more acceptable for supporting nuanced decisions about treatment continuation or cessation. Importantly, from a practical perspective, breath testing was seen as easily feasible with limited patient burden, however, the key challenge of integrating results into complex clinical decision-making repeatedly highlighted. These findings demonstrate the value of early human factors evaluation in de-risking experimental diagnostics and guiding their responsible adoption in cancer care.
Massimo Micocci, Henry Robb, Philip Leung, Bibek Das, Peter Buckle, George Hanna
Open Access
Article
Conference Proceedings
Assessing Innovative Healthcare Models: An Integrated Measurement Framework of Clinical, Economic, Social, and Environmental Performance
Population ageing is among the most consequential demographic shifts of the twenty-first century, with the share of adults aged 65 and over rising rapidly across Europe and particularly in Italy. This transition is accompanied by increasing multimorbidity, chronic disease burden, and long-term care needs that strain traditional hospital-centred systems, which are ill-equipped to address the complex, multidimensional needs of older populations. Consequently, policymakers are advancing innovative socio-healthcare models that prioritize community-based services, integration of health and social care, prevention, and person-centred approaches. These reforms promote multidisciplinary pathways, home-based assistance, digital health solutions, and new organizational structures aimed at improving sustainability and quality of care.Despite their growing adoption, evaluating the effectiveness of these models remains challenging. Existing performance assessment frameworks focus largely on clinical outputs and financial metrics, overlooking broader outcomes such as quality of life, accessibility, social participation, and system resilience. Addressing this gap requires multidimensional evaluation tools capable of capturing the full value generated by integrated socio-healthcare systems. This study responds by systematically identifying indicators proposed in the scientific and grey literature to assess innovative care models for ageing populations. Through a structured review of evaluation frameworks across healthcare, social services, and integrated care contexts, the research highlights fragmentation in terminology and methodologies that limits comparability and decision-making. The resulting indicator system incorporates measures of accessibility, workforce capacity, environmental impact, social inclusion, well-being, clinical effectiveness, and cost efficiency. This comprehensive framework supports balanced performance appraisal, continuous monitoring, and evidence-informed policy decisions to enhance care delivery and outcomes for ageing societies
Manila Caragiuli, Agnese Brunzini, Rebecca Posa, Michele Germani
Open Access
Article
Conference Proceedings
A Chest-Worn Quad-Modal Cardiac Monitoring Device Combining ECG, PCG, SCG, and GCG with Cross-Modal Motion Artifact Suppression
Continuous monitoring of both electrical and mechanical cardiac activity is essential for early detection and management of cardiovascular diseases in real-life environments. This paper presents the design and preliminary evaluation of a chest-worn, Holter-like device that enables 24-hour quad-modal cardiac monitoring by synchronously acquiring electrocardiogram (ECG), phonocardiogram (PCG), seismocardiogram (SCG), and gyrocardiogram (GCG) signals. The main unit is attached to the chest and integrates a heart sound sensor, a 6-axis inertial measurement unit (IMU), data acquisition and storage circuits, and a battery into a single compact housing, while four limb leads (RA, RL, LA, LL) are extended from the device to record ECG. All cardiac signals are sampled at 10 kHz under a shared hardware clock, ensuring absolute temporal synchronization across modalities.Building on the IMU, SCG (chest wall micro-acceleration) and GCG (chest wall micro-rotation) are treated not only as auxiliary motion references, but also as cardio-mechanical signals that are jointly analyzed with ECG. A cross-modal motion artifact suppression framework is proposed, in which ECG, SCG, and GCG mutually constrain each other: motion-dominated components are identified via their inconsistent morphology across modalities, while cardiac components exhibit stable beat-synchronous patterns. The denoised ECG then serves as a temporal reference to perform ECG-guided heart sound segmentation on the PCG, enabling robust extraction of the first to fourth heart sounds (S1–S4). A custom desktop software platform supports synchronized visualization, beat-level quality assessment, and batch analysis of 24-hour recordings.Preliminary tests on healthy subjects during daily activities (resting, walking, posture changes) show that the proposed quad-modal system effectively reduces motion-induced artifacts, improves the morphological consistency of ECG, SCG, and GCG, and achieves reliable multi–heart sound segmentation under ambulatory conditions.The chest-worn, integrated design and cross-modal processing pipeline demonstrate strong potential as a user-friendly and low-cost solution for continuous, multi-dimensional cardiovascular monitoring in clinical and home settings.
Yingwei Li, Qianxiang Zhou, Zhongqi Liu, Mengmeng Jin
Open Access
Article
Conference Proceedings
Digital Biomarkers for the Assessment of Motor Symptoms in Parkinson’s Disease: From Daily Life to Intervention Evaluation
The objective and continuous assessment of motor symptoms in Parkinson’s disease (PD) remains limited by the episodic nature of clinical evaluations. This work presents a dual monitoring methodology developed within the BioCliTe project, integrating both standardized MDS‑UPDRS Part III exercises and an ecologically valid daily‑life task. Data were collected using smartwatches that acquired accelerometer and gyroscope signals at 50 Hz in supervised and free‑living settings, guided by a mobile application that enabled automatic labelling. The recordings were processed through a reproducible pipeline including filtering, segmentation, windowing, and feature extraction in both time and frequency domains. Explainable machine‑learning models, such as decision‑tree ensembles, logistic regression, and SVM, were trained using interpretability methods (LIME, SHAP) to define digital biomarkers of tremor, bradykinesia, and gait. These biomarkers demonstrated strong capability to differentiate PD patients from healthy controls and to reflect motor severity in unsupervised environments. Results confirm the feasibility of diagnosing and monitoring PD symptoms outside clinical facilities through wearable‑based biomechanical analysis. Notably, the free-living task yielded a low‑cost and reproducible bradykinesia biomarker with robust performance in clinical and remote conditions. The defined digital biomarkers establish the basis for the EVINTERS project, aimed at evaluating therapeutic effects on symptom progression. This approach supports more personalized, continuous, and patient‑centered management beyond point‑in‑time assessments.
Ignacio Pavón Garcia, Carlos Polvorinos-Fernández, Irene Sánchez-Hernández, Guillermo De Arcas, Luis Sigcha
Open Access
Article
Conference Proceedings
Leveraging AI Tools for Emotion-Safe Dental Imaging: Enhancing Patient Communication While Preserving Anatomical Accuracy in Surgical Visualizations
This paper presents a co-design methodology applied to the development of Mixed Reality (MR) interfaces within the Motivate XR project, targeting industrial training and operational support scenarios. The proposed approach emphasizes a user-centered design process, achieved through the continuous involvement of end users alongside developers, with the objective of improving usability while addressing ergonomic and operational constraints. The design process was structured around a systematic, ergonomics-driven methodology, tailored to the specific use cases of multiple industrial pilots. Insights derived from individual case studies were combined to define a design solution adaptable across heterogeneous real-world environments. A series of co-design workshops were coordinated to iteratively present, evaluate, and refine interface concepts based on direct user feedback, ensuring alignment with operational goals. To validate the proposed designs, 360-degree photos and videos of real industrial environments were used to simulate MR interactions within the actual pilot contexts. Interface layouts and visual elements, such as contrast, spatial arrangement, and content readability, were optimized through the creation of MR mockups, supporting both functional effectiveness and perceptual clarity. The methodology placed strong emphasis on real-time user feedback, enabling rapid iteration and continuous refinement of design decisions.The co-design activities were supported by a combination of UX tools and MR technologies, including 2D prototyping platforms and immersive VR environments, facilitating collaborative evaluation and validation. Results demonstrate that co-design represents an effective strategy for the development of scalable, user-centered XR interfaces in industrial contexts, contributing to improved usability and stronger alignment between technical solutions and end-user needs.
Sarah De Cristofaro, Salvatore Parrulli, Bernardo Parrulli, Luca Rizzi
Open Access
Article
Conference Proceedings
Machine learning-based identification of non-responders to a 12-month digital self-management in knee osteoarthritis
Knee osteoarthritis (KOA) places a significant burden on individuals and healthcare systems worldwide. Although self-management programs (SMPs) offer accessible support for KOA management, individual responses vary. Therefore, early identification of those unlikely to show substantial improvement with self-management is important to adapt treatment plans promptly, introduce more effective interventions when needed, and ultimately improve patient outcomes by avoiding prolonged use of strategies that do not produce responses. This study aims to develop a machine learning-based approach to identify potential non-responders to a digital SMP using characteristics at program entry. Data were obtained from a previously conducted 12-month app-based SMP for KOA. Responders were defined as individuals whose arthritis self-efficacy (ASE) and health-related quality of life (HRQoL) outcomes both improved. Non-responders included those who showed improvement in only one outcome, no improvement in either, or deterioration in one or both outcomes. After excluding participants who did not complete the SMP or who underwent knee-related surgery and/or hospitalization during the study period, 57 participants were included in the analysis. Body mass index, presence of non-musculoskeletal comorbidities, ASE score, HRQoL index, and hip range of motion were the input features for model development. Gradient boosting decision tree achieved the best performance, with an AUC of 0.822 and balanced sensitivity and specificity in identifying non-responders. These findings present the feasibility of using machine learning to early identify individuals with limited expected benefit from digital self-management. Such identification may facilitate more efficient, tailored, and proactive strategies for managing KOA. Future research should prioritize external validation in larger and more diverse cohorts.
Yingyi Li, Xiaoyi Wang, Calvin Or
Open Access
Article
Conference Proceedings
Human Factors Influencing Trust in Healthcare Providers as Primary COVID-19 Information Sources Among Cancer Survivors: A Health Belief Model Analysis
Due to Covid-19 rapid escalation at the global level, a growing body of misinformation sources became available to patients. However, patterns and determinants of consulting unreliable sources are not well understood. Using the Health Belief Model, this study investigates the impactors of trust in healthcare providers as the main source of Covid-19 information-seeking patterns. Methods This retrospective study used restricted data from the 2021 Health Information National Trends Survey (HINTS SEER), which collected information from January 11, 2021, to August 20, 2021. We used the partial least squares structural equation modeling (PLS-SEM) method for data analysis. Missing data were handled using a multiple imputations method. Results A total of 1234 cancer survivors were included in the study. The goodness of fit of the structural model indicated an acceptable and satisfactory fit. SEM analysis showed that the "perceived severity" and the "cues to action" did not affect the behavior of the cancer survivors. By contrast, perceived self-efficacy (β=0.088, P<0.001), benefits (β=0.009, P<0.001), barriers (β=-0.064, P=0.001), and susceptibility (β=-0.089, P<0.001) were predictors of the behavior. Conclusions The findings of our study provide important insights into the factors that affect cancer survivors' trust in doctors as the main source of Covid-19 Information. Our results suggest that patient-centered authentic, reliable, and accurate communication centered around the cancer survivors' needs should be adopted to ensure patients continue to trust their providers. The study also suggests that it remains important to support patients' self-efficacy to know how to handle critical situations and their trust in their abilities.
Safa Elkefi, Dario Trapani, Congyu Wu
Open Access
Article
Conference Proceedings
Expert Evaluation and Guidance for a Home-Based Rehabilitation System in Postpartum Low Back Pain
Postpartum low back pain (LBP) is a prevalent health issue. Most patients with postpartum LBP recover within three months after delivery. However, a considerable number of women (50-80%) still report developing LBP after this, and some cases may progress to chronic back pain, lasting for months or even years. Postpartum LBP arises from multiple risk factors spanning physiological, psychological, and social aspects. Traditional rehabilitation protocols often provide generalized guidance, failing to address the individual needs of postpartum women, which vary depending on their physical condition, lifestyle, and recovery trajectory. Therefore, it is necessary to develop more convenient and personalized treatment services for patients. This study aims to obtain professional guidance on developing a personalized home-based rehabilitation service system for postpartum LBP. This study conducted semi-structured interviews with six multidisciplinary experts, including one kinesiology specialist, one rehabilitation physician, two physiotherapists, and two obstetricians. The interviews explored the experts' views on the feasibility, potential benefits, and risks of home-based exercise rehabilitation for patients with postpartum LBP. Analysis of the interview records revealed a consensus among the experts on the value of personalized home-based interventions for treating postpartum LBP.
Xiaoli Li, Wenjing Yang
Open Access
Article
Conference Proceedings
Physics-Informed Neural Networks for Ultrasound-Based Varicose Vein Screening
Early screening of varicose veins using non-invasive and low-cost sensing techniques remains a practical challenge in community and home-based healthcare settings. Conventional ultrasound imaging systems are accurate but require bulky hardware, skilled operators, and clinical environments, which limit accessibility for large-scale screening. This study proposes a physics-informed deep learning framework for one-dimensional ultrasound A-scan analysis to enable lightweight and interpretable varicose vein screening. A convolutional neural network with an encoder–decoder reconstruction branch and a classification head was developed. The reconstruction branch incorporates physics-motivated constraints, including second-order temporal smoothness and signal energy consistency, to regularize the learned waveform representation. These constraints aim to preserve physically meaningful propagation characteristics while suppressing noise and overfitting. A stratified train–test split with a guaranteed test size of 50 samples was employed to improve statistical reliability under small-sample conditions. Experimental results on 75 labeled A-scan segments demonstrated stable convergence and high screening performance. The final model achieved approximately 98% test accuracy with zero false negatives, indicating strong sensitivity for detecting abnormal vascular conditions. The physics-informed reconstruction produced smoother yet structurally consistent waveforms compared with raw signals, suggesting improved interpretability of learned representations. The findings indicate that physics-guided learning can enhance robustness and clinical relevance in small-sample ultrasound screening tasks. While further validation with participant-level separation and larger cohorts is required, the proposed framework provides a feasible direction for portable, AI-assisted vascular screening in non-clinical environments.
Fan Yang, Qian Mao
Open Access
Article
Conference Proceedings
Advances in Intelligent Rehabilitation Systems for Chronic Nonspecific Low Back Pain
Low back pain (LBP) is one of the most prevalent musculoskeletal disorders worldwide, and over 85% of chronic cases are classified as chronic nonspecific low back pain (CNSLBP). Traditional rehabilitation approaches, such as medication, physical therapy, and self-managed exercise, often face challenges like poor adherence and limited feedback. With advances in wearable sensors, virtual reality (VR), artificial intelligence (AI), and tele-rehabilitation, intelligent rehabilitation systems are emerging as innovative home-based solutions. Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) framework, this review analyzes studies on CNSLBP rehabilitation systems published between 2016 and 2025. It identifies the critical components, smart technologies, and design characteristics of the CNSLBP rehabilitation systems, while highlighting their effectiveness in improving user engagement. Current limitations include fragmented system integration, inadequate user experience design, and weak behavioral intervention mechanisms. Future development should emphasize standardized digital therapeutics, multimodal personalization, and user-centered interdisciplinary collaboration to enhance the clinical efficacy and quality of life for patients.
Pengda Lu, Wenjing Yang
Open Access
Article
Conference Proceedings
The Application of Machine Learning in Postpartum Low Back Pain
Postpartum low back pain (LBP) is a common health issue that significantly impacts women’s quality of life; however, traditional rehabilitation models often provide generalized guidance, making it difficult to address the individual differences in patients’ physiological, psychological, and lifestyle factors with precision. With the development of smart healthcare, machine learning (ML) provides the technological foundation for personalized interventions. This study employs a scoping review methodology to systematically analyze the current state of ML applications in postpartum low back pain and related fields between 2015 and 2025, aiming to provide a theoretical framework for the design of future personalized postpartum rehabilitation service systems.The study found that although ML has made significant progress in precise diagnosis and personalized recommendations in adjacent fields such as non-specific low back pain, its application to postpartum LBP remains in its infancy. This paper identifies key design challenges, including multi-dimensional data integration, model interpretability, and the motivation for participation among specific populations. From a design science perspective, the study proposes that future efforts should focus on integrating multimodal data—such as physiological indicators and psychosocial factors—to construct a personalized rehabilitation service system capable of real-time monitoring and dynamic adjustment, thereby meeting the complex needs of postpartum women. This research not only distils a transferable interdisciplinary methodology but also provides directional guidance for the design of smarter postpartum health services.
Xiaoli Li, Wenjing Yang
Open Access
Article
Conference Proceedings