Search Results (145)

The modified Rankin Scale (mRS) is a widely used outcome measure for assessing disability in stroke care; however, its administration is often affected by subjectivity and variability, leading to poor inter-rater reliability and inconsistent scoring. Originally designed for hospital discharge evaluations, the mRS has evolved into an outcome tool for disability assessment and clinical decision-making. Inconsistencies persist due to a lack of standardization and cognitive biases during its use. This paper presents design principles for creating a standardized clinical assessment tool (CAT) for the mRS, grounded in human–computer interaction (HCI) and cognitive engineering principles. Design principles were informed in part by an anonymous online survey conducted with clinicians across Canada to gain insights into current administration practices, opinions, and challenges of the mRS. The proposed design principles aim to reduce cognitive load, improve inter-rater reliability, and streamline the administration process of the mRS. By focusing on usability and standardization, the design principles seek to enhance scoring consistency and improve the overall reliability of clinical outcomes in stroke care and research. Developing a standardized CAT for the mRS represents a significant step toward improving the accuracy and consistency of stroke disability assessments. Future work will focus on real-world validation with healthcare stakeholders and exploring self-completed mRS assessments to further refine the tool. Full article

National and international aged care frameworks recommend family-integrated care to enhance care quality and outcomes, supported by evidence demonstrating improvements in patient and clinician experiences. Yet uncertainty remains about how to integrate family carers effectively in diverse healthcare models and settings for neurodegenerative and mental health conditions. A systematic integrative review was conducted to answer two research questions: how do the studies describe the integration of family carers in health services design and delivery for older patients with neurodegenerative and mental health conditions? And what is the evidence for family-integrated care models impacting the health and wellbeing of these older patients? Structured and iterative searches of five databases (CINAHL, Medline (Ovid), Web of Science, PsycINFO, and ProQuest) and the Google Scholar search engine identified 2271 records. A Covidence screening process resulted in 14 studies for review, comprising randomised controlled trials, mixed methods studies, qualitative studies, and quasi-experimental designs. The following four themes emerged from the evidence synthesis: (1) family participation in service delivery, (2) health and wellbeing outcomes, (3) satisfaction with care, and (4) service dynamics in enabling family-integrated care successfully. This review highlights that while family-integrated care models contribute to positive health and wellbeing outcomes for older patients with neurodegenerative and mental health conditions, challenges remain for implementation due to the extent and variability in integration strategies, a lack of rigorous evaluation, and an absence of standardised frameworks. Full article

Human Digital Twins are an emerging type of Digital Twin used in healthcare to provide personalized support. Following this trend, we intend to elevate our virtual fitness coach, a coaching platform using wearable data on physical activity, to the level of a personalized Human Digital Twin. Preliminary investigations revealed a significant difference in performance, as measured by prediction accuracy and F1-score, between the optimal choice of machine learning algorithms for generalized and personalized processing of the available data. Based on these findings, this survey aims to establish the state of the art in the selection and application of machine learning algorithms in Human Digital Twin applications in healthcare. The survey reveals that, unlike general machine learning applications, there is a limited body of literature on optimization and the application of meta-learning in personalized Human Digital Twin solutions. As a conclusion, we provide direction for further research, formulated in the following research question: how can the optimization of human data feature engineering and personalized model selection be achieved in Human Digital Twins and can techniques such as meta-learning be of use in this context? Full article

Background: Chronic diseases significantly burden healthcare systems due to the need for long-term treatment. Early diagnosis is critical for effective management and minimizing risk. The current traditional diagnostic approaches face various challenges regarding efficiency and cost. Digitized healthcare demonstrates several opportunities for reducing human errors, increasing clinical outcomes, tracing data, etc. Artificial Intelligence (AI) has emerged as a transformative tool in healthcare. Subsequently, the evolution of Generative AI represents a new wave. Large Language Models (LLMs), such as ChatGPT, are promising tools for enhancing diagnostic processes, but their potential in this domain remains underexplored. Methods: This study represents the first systematic evaluation of ChatGPT’s performance in chronic disease prediction, specifically targeting heart disease and diabetes. This study compares the effectiveness of zero-shot, few-shot, and CoT reasoning with feature selection techniques and prompt formulations in disease prediction tasks. The two latest versions of GPT4 (GPT-4o and GPT-4o-mini) are tested. Then, the results are evaluated against the best models from the literature. Results: The results indicate that GPT-4o significantly beat GPT-4o-mini in all scenarios regarding accuracy, precision, and F1-score. Moreover, a 5-shot learning strategy demonstrates superior performance to zero-shot, few-shot (3-shot and 10-shot), and various CoT reasoning strategies. The 5-shot learning strategy with GPT-4o achieved an accuracy of 77.07% in diabetes prediction using the Pima Indian Diabetes Dataset, 75.85% using the Frankfurt Hospital Diabetes Dataset, and 83.65% in heart disease prediction. Subsequently, refining prompt formulations resulted in notable improvements, particularly for the heart dataset (5% performance increase using GPT-4o), emphasizing the importance of prompt engineering. Conclusions: Even though ChatGPT does not outperform traditional machine learning and deep learning models, the findings highlight its potential as a complementary tool in disease prediction. Additionally, this work provides value by setting a clear performance baseline for future work on these tasks Full article

Resistive gas sensors have attracted significant attention due to their simple architecture, low cost, and ease of integration, with widespread applications in environmental monitoring, industrial safety, and healthcare diagnostics. This review provides a comprehensive overview of recent advances in resistive gas sensors, focusing on their fundamental working mechanisms, sensing material design, device architecture optimization, and intelligent system integration. These sensors primarily operate based on changes in electrical resistance induced by interactions between gas molecules and sensing materials, including physical adsorption, charge transfer, and surface redox reactions. In terms of materials, metal oxide semiconductors, conductive polymers, carbon-based nanomaterials, and their composites have demonstrated enhanced sensitivity and selectivity through strategies such as doping, surface functionalization, and heterojunction engineering, while also enabling reduced operating temperatures. Device-level innovations—such as microheater integration, self-heated nanowires, and multi-sensor arrays—have further improved response speed and energy efficiency. Moreover, the incorporation of artificial intelligence (AI) and Internet of Things (IoT) technologies has significantly advanced signal processing, pattern recognition, and long-term operational stability. Machine learning (ML) algorithms have enabled intelligent design of novel sensing materials, optimized multi-gas identification, and enhanced data reliability in complex environments. These synergistic developments are driving resistive gas sensors toward low-power, highly integrated, and multifunctional platforms, particularly in emerging applications such as wearable electronics, breath diagnostics, and smart city infrastructure. This review concludes with a perspective on future research directions, emphasizing the importance of improving material stability, interference resistance, standardized fabrication, and intelligent system integration for large-scale practical deployment. Full article

This paper provides a systematic and in-depth examination of large language models (LLMs) in the healthcare domain, addressing their significant potential to transform medical practice through advanced natural language processing capabilities. Current implementations demonstrate LLMs’ promising applications across clinical decision support, medical education, diagnostics, and patient care, while highlighting critical challenges in privacy, ethical deployment, and factual accuracy that require resolution for responsible integration into healthcare systems. This paper provides a comprehensive understanding of the background of healthcare LLMs, the evolution and architectural foundation, and the multimodal capabilities. Key methodological aspects—such as domain-specific data acquisition, large-scale pre-training, supervised fine-tuning, prompt engineering, and in-context learning—are explored in the context of healthcare use cases. The paper highlights the trends and categorizes prominent application areas in medicine. Additionally, it critically examines the prevailing technical and social challenges of healthcare LLMs, including issues of model bias, interpretability, ethics, governance, fairness, equity, data privacy, and regulatory compliance. The survey concludes with an outlook on emerging research directions and strategic recommendations for the development and deployment of healthcare LLMs. Full article

The analysis of electrocardiogram (ECG) signals is profoundly affected by the presence of electromyographic (EMG) noise, which can lead to substantial misinterpretations in healthcare applications. To address this challenge, we present ECGDnet, an innovative architecture based on Transformer technology, specifically engineered to denoise multi-channel ECG signals. By leveraging multi-head self-attention mechanisms, positional embeddings, and an advanced sequence-to-sequence processing architecture, ECGDnet effectively captures both local and global temporal dependencies inherent in cardiac signals. Experimental validation on real-world datasets demonstrates ECGDnet’s remarkable efficacy in noise suppression, achieving a Signal-to-Noise Ratio (SNR) of 19.83, a Normalized Mean Squared Error (NMSE) of 0.9842, a Reconstruction Error (RE) of 0.0158, and a Pearson Correlation Coefficient (PCC) of 0.9924. These results represent significant improvements from traditional deep learning approaches while maintaining complex signal morphology and effectively mitigating noise interference. Full article

Clinical documentation, particularly the hospital discharge report (HDR), is essential for ensuring continuity of care, yet its preparation is time-consuming and places a considerable clinical and administrative burden on healthcare professionals. Recent advancements in Generative Artificial Intelligence (GenAI) and the use of prompt engineering in large language models (LLMs) offer opportunities to automate parts of this process, improving efficiency and documentation quality while reducing administrative workload. This study aims to design a digital system based on LLMs capable of automatically generating HDRs using information from clinical course notes and emergency care reports. The system was developed through iterative cycles, integrating various instruction flows and evaluating five different LLMs combined with prompt engineering strategies and agent-based architectures. Throughout the development, more than 60 discharge reports were generated and assessed, leading to continuous system refinement. In the production phase, 40 pneumology discharge reports were produced, receiving positive feedback from physicians, with an average score of 2.9 out of 4, indicating the system’s usefulness, with only minor edits needed in most cases. The ongoing expansion of the system to additional services and its integration within a hospital electronic system highlights the potential of LLMs, when combined with effective prompt engineering and agent-based architectures, to generate high-quality medical content and provide meaningful support to healthcare professionals. Hospital discharge reports (HDRs) are pivotal for continuity of care but consume substantial clinician time. Generative AI systems based on large language models (LLMs) could streamline this process, provided they deliver accurate, multilingual, and workflow-compatible outputs. We pursued a three-stage, design-science approach. Proof-of-concept: five state-of-the-art LLMs were benchmarked with multi-agent prompting to produce sample HDRs and define the optimal agent structure. Prototype: 60 HDRs spanning six specialties were generated and compared with clinician originals using ROUGE with average scores compatible with specialized news summarizing models in Spanish and Catalan (lower scores). A qualitative audit of 27 HDR pairs showed recurrent divergences in medication dose (56%) and social context (52%). Pilot deployment: The AI-HDR service was embedded in the hospital’s electronic health record. In the pilot, 47 HDRs were autogenerated in real-world settings and reviewed by attending physicians. Missing information and factual errors were flagged in 53% and 47% of drafts, respectively, while written assessments diminished the importance of these errors. An LLM-driven, agent-orchestrated pipeline can safely draft real-world HDRs, cutting administrative overhead while achieving clinician-acceptable quality, not without errors that require human supervision. Future work should refine specialty-specific prompts to curb omissions, add temporal consistency checks to prevent outdated data propagation, and validate time savings and clinical impact in multi-center trials. Full article

Durable plastics are a sustainability challenge for healthcare products. Orthopedic products are regulated with strict specifications for human tissue interactions. Healthcare engineers and managers select plastic to meet the full range of material properties. Plastic is plentiful, low cost, and reliable, with established supply chains. Used plastic products can be discarded using existing waste management systems with low externality costs for orthopedic businesses. However, plastic is produced from fossil petroleum, raising issues for sustainability commitments of healthcare product companies. Barriers to the transition away from single-use plastic toward circular systems and bio-based healthcare products have been studied, but the transition is a goal that has yet to be realized. This research article reports on a transition engineering design sprint with a medium-sized orthopedic company specializing in orthoses for children and teenagers. The design sprint process engages company experts with systems perspectives on the role of unsustainable plastic in orthopedic healthcare and illuminates opportunities for capturing value in business transition. Two system transition project concepts were co-developed. The first concept is a plastics value map that aims to converge the satisfaction of essential needs with the usefulness of plastics under the limitations of a biophysically constrained future economy. The second concept is an orthopedics library data system concept that would allow reusing of fit-for-purpose used products and to inform the refurbishment of used products. In addition to an explanation of the design of the two concepts, the article presents reflections of co-design stakeholders on the usefulness and usability of the concepts. The article provides a real-world application of the co-design processes in transition engineering and the reflection by the company on the value of the results. The results indicate that the co-designed concepts could enable the company to address its sustainability aspirations and potentially resolve the dissonance of sustainability and business viability. Full article

In the past decade, extended reality (XR) has been introduced into healthcare due to several potential benefits, such as scalability and cost savings. As there is no comprehensive study covering all the factors influencing the selection of an XR system in the healthcare and medical domain, a Decision Support System is proposed in this paper to identify and rank factors impacting the performance of XR in this domain from an engineering design perspective. The proposed system is built upon the Supply Chain Operations Reference (SCOR) model supported by a literature survey and experts’ knowledge to extract and identify important factors. Subsequently, the factors are categorized into distinct categories, and their relative importance is specified by Analytic Network Process (ANP) models under a fuzzy environment. Two fuzzy approaches for the ANP models are compared, and the results are analyzed using statistical testing. The computational results show that the ranking agreement between the two fuzzy approaches is strong and corresponds to the fact that both approaches yield the same ranking of primary factors, highlighting the significance of reliability as the topmost factor, followed by responsiveness, cost, and agility. It is shown that while the top three important sub-factors are identical between the two approaches, their relative order is slightly varied. Safety is considered to be the most critical aspect within the reliability category in both approaches, but there are discrepancies in the rankings of accuracy and user control and freedom. Both approaches also consider warranty and depreciation costs as the least significant criteria. Full article

The rapid growth of wearable technology has opened new possibilities for smart health-monitoring systems. Among various sensing methods, bio-impedance sensing has stood out as a powerful, non-invasive, and energy-efficient way to track physiological changes and gather important health information. This review looks at the basic principles behind bio-impedance sensing, how it is being built into wearable devices, and its use in healthcare and everyday wellness tracking. We examine recent progress in sensor design, signal processing, and machine learning, and show how these developments are making real-time health monitoring more effective. While bio-impedance systems offer many advantages, they also face challenges, particularly when it comes to making devices smaller, reducing power use, and improving the accuracy of collected data. One key issue is that analyzing bio-impedance signals often relies on complex digital signal processing, which can be both computationally heavy and energy-hungry. To address this, researchers are exploring the use of neuromorphic processors—hardware inspired by the way the human brain works. These processors use spiking neural networks (SNNs) and event-driven designs to process signals more efficiently, allowing bio-impedance sensors to pick up subtle physiological changes while using far less power. This not only extends battery life but also brings us closer to practical, long-lasting health-monitoring solutions. In this paper, we aim to connect recent engineering advances with real-world applications, highlighting how bio-impedance sensing could shape the next generation of intelligent wearable devices. Full article

Background/Objectives: Gram staining is an essential diagnostic technique used for the rapid identification of bacterial and fungal infections, playing a pivotal role in clinical decision-making, especially in point-of-care (POC) settings. Manual staining, while effective, is labor-intensive and prone to variability, relying heavily on the skill of laboratory personnel. Current automated Gram-staining systems are primarily designed for high-throughput laboratory environments, limiting their feasibility in decentralized healthcare settings such as emergency departments and rural clinics. This study aims to introduce and evaluate the Point-of-Care Gram Stainer (PoCGS^®), a compact, automated device engineered for single-slide processing, addressing challenges related to portability, standardization, and efficiency in POC applications. Methods: The PoCGS^® device was developed to emulate expert manual staining techniques through features such as methanol fixation and programmable reagent application. A comparative evaluation was performed using 40 urine samples, which included both clinical and artificial specimens. These samples were processed using PoCGS^®, manual staining by skilled experts, and manual staining by unskilled personnel. The outcomes were assessed based on microbial identification concordance, the staining uniformity, presence of artifacts, and agreement with the culture results. Statistical analyses, including agreement rates and quality scoring, were conducted to compare the performance of PoCGS^® against manual staining methods. Results: PoCGS^® achieved a 100% concordance rate with expert manual staining in terms of microbial identification, confirming its diagnostic accuracy. However, staining quality parameters such as the uniformity and presence of artifacts showed statistically significant differences when compared to skilled and unskilled personnel. Despite these limitations, PoCGS^® demonstrated a comparable performance regarding artifact reduction and agreement with the culture results, indicating its potential utility in POC environments. Challenges such as fixed processing times and limited adaptability to varying specimen characteristics were identified as areas for further improvement. Conclusions: The study findings suggest that PoCGS^® is a reliable and valuable tool for microbial identification in POC settings, with a performance comparable to skilled manual staining. Its compact design, automation, and ease of use make it particularly beneficial for resource-limited environments. Although improvements in staining uniformity and background clarity are required, PoCGS^® has the potential to standardize Gram staining protocols and improve diagnostic turnaround times. Future developments will focus on optimizing staining parameters and expanding its application to other clinical sample types, ensuring robustness and broader usability in diverse healthcare settings. Full article

Large language models (LLMs) have revolutionized natural language processing across diverse domains, yet they also raise critical fairness and ethical concerns, particularly regarding gender bias. In this study, we conduct a systematic, mathematically grounded investigation of gender bias in four leading LLMs—GPT-4o, Gemini 1.5 Pro, Sonnet 3.5, and LLaMA 3.1:8b—by evaluating the gender distributions produced when generating “perfect personas” for a wide range of occupational roles spanning healthcare, engineering, and professional services. Leveraging standardized prompts, controlled experimental settings, and repeated trials, our methodology quantifies bias against an ideal uniform distribution using rigorous statistical measures and information-theoretic metrics. Our results reveal marked discrepancies: GPT-4o exhibits pronounced occupational gender segregation, disproportionately linking healthcare roles to female identities while assigning male labels to engineering and physically demanding positions. In contrast, Gemini 1.5 Pro, Sonnet 3.5, and LLaMA 3.1:8b predominantly favor female assignments, albeit with less job-specific precision. These findings demonstrate how architectural decisions, training data composition, and token embedding strategies critically influence gender representation. The study underscores the urgent need for inclusive datasets, advanced bias-mitigation techniques, and continuous model audits to develop AI systems that are not only free from stereotype perpetuation but actively promote equitable and representative information processing. Full article

The term “big data analytics (BDA)” defines the computational techniques to study complex datasets that are too large for common data processing software, encompassing techniques such as data mining (DM), machine learning (ML), and predictive analytics (PA) to find patterns, correlations, and insights in massive datasets. Cardiovascular diseases (CVDs) are attributed to a combination of various risk factors, including sedentary lifestyle, obesity, diabetes, dyslipidaemia, and hypertension. We searched PubMed and published research using the Google and Cochrane search engines to evaluate existing models of BDA that have been used for CVD prediction models. We critically analyse the pitfalls and advantages of various BDA models using artificial intelligence (AI), machine learning (ML), and artificial neural networks (ANN). BDA with the integration of wide-ranging data sources, such as genomic, proteomic, and lifestyle data, could help understand the complex biological mechanisms behind CVD, including risk stratification in risk-exposed individuals. Predictive modelling is proposed to help in the development of personalized medicines, particularly in pharmacogenomics; understanding genetic variation might help to guide drug selection and dosing, with the consequent improvement in patient outcomes. To summarize, incorporating BDA into cardiovascular research and treatment represents a paradigm shift in our approach to CVD prevention, diagnosis, and management. By leveraging the power of big data, researchers and clinicians can gain deeper insights into disease mechanisms, improve patient care, and ultimately reduce the burden of cardiovascular disease on individuals and healthcare systems. Full article

Conductive hydrogels, particularly those incorporating poly(3,4-ethylenedioxythiophene):polystyrene sulfonate (PEDOT:PSS), have revolutionized wearable health monitoring by merging tissue-like softness with robust electronic functionality. This review systematically explores design strategies for PEDOT:PSS-based hydrogels, focusing on advanced gelation methods, including polymer crosslinking, ionic interactions, and light-induced polymerization, to engineer hierarchical networks that balance conductivity and mechanical adaptability. Cutting-edge fabrication techniques such as electrochemical patterning, additive manufacturing, and laser-assisted processing further enable precise microstructural control, enhancing interfacial compatibility with biological systems. The applications of these hydrogels in wearable sensors are highlighted through their capabilities in real-time mechanical deformation tracking, dynamic tissue microenvironment analysis, and high-resolution electrophysiological signal acquisition. Environmental stability and long-term durability are critical for ensuring reliable operation under physiological conditions and mitigating performance degradation caused by fatigue, oxidation, or biofouling. By addressing critical challenges in environmental stability and long-term durability, PEDOT:PSS hydrogels demonstrate transformative potential for personalized healthcare, where their unique combination of softness, biocompatibility, and tunable electro-mechanical properties enables seamless integration with human tissues for continuous, patient-specific physiological monitoring. These systems offer scalable solutions for multi-modal diagnostics, empowering tailored therapeutic interventions and chronic disease management. The review concludes with insights into future directions, emphasizing the integration of intelligent responsiveness and energy autonomy to advance next-generation bioelectronic interfaces. Full article