Search

Abstract Accurate and comprehensive nursing documentation is essential to ensure quality patient care. To streamline this process, we present SONAR, a publicly available dataset of nursing activities recorded using inertial sensors in a nursing home. The dataset includes 14 sensor streams, such as acceleration and angular velocity, and 23 activities recorded by 14 caregivers using five sensors for 61.7 hours. The caregivers wore the sensors as they performed their daily tasks, allowing for continuous monitoring of their activities. We additionally provide machine learning models that recognize the nursing activities given the sensor data. In particular, we present benchmarks for three deep learning model architectures and evaluate their performance using different metrics and sensor locations. Our dataset, which can be used for research on sensor-based human activity recognition in real-world settings, has the potential to improve nursing care by providing valuable insights that can identify areas for improvement, facilitate accurate documentation, and tailor care to specific patient conditions.

Virtual reality (VR) technology is often referred to as the 'ultimate empathy machine' due to its capability to immerse users in alternate perspectives and environments beyond their immediate physical reality. In this study, participants will be immersed in 3-dimensional 360° VR videos where actors express different emotions (sadness, happiness, anger, and anxiousness). The primary objective is to investigate the potential relationship between participants' empathy levels and the changes in their physiological attributes. The empathy levels will be self-reported with questionnaires, and physiological attributes will be measured using different sensors. The main outcome of the study will be a machine learning model to predict a person's empathy level based on their physiological responses while watching VR videos. Despite the existence of established methodologies and metrics in research and clinical domains, our aim is to contribute to addressing the gap of a universally accepted "gold standard" for assessing empathy. Additionally, we expect to deepen our understanding of the relationship between different emotions and psychological attributes, gender differences in empathy, and the impact of narrative context on empathic responses.

Sensor-based human activity recognition is becoming ever more prevalent. The increasing importance of distinguishing human movements, particularly in healthcare, coincides with the advent of increasingly compact sensors. A complex sequence of individual steps currently characterizes the activity recognition pipeline. It involves separate data collection, preparation, and processing steps, resulting in a heterogeneous and fragmented process. To address these challenges, we present a comprehensive framework, HARE, which seamlessly integrates all necessary steps. HARE offers synchronized data collection and labeling, integrated pose estimation for data anonymization, a multimodal classification approach, and a novel method for determining optimal sensor placement to enhance classification results. Additionally, our framework incorporates real-time activity recognition with on-device model adaptation capabilities. To validate the effectiveness of our framework, we conducted extensive evaluations using diverse datasets, including our own collected dataset focusing on nursing activities. Our results show that HARE’s multimodal and on-device trained model outperforms conventional single-modal and offline variants. Furthermore, our vision-based approach for optimal sensor placement yields comparable results to the trained model. Our work advances the field of sensor-based human activity recognition by introducing a comprehensive framework that streamlines data collection and classification while offering a novel method for determining optimal sensor placement.

One key task in the early fight against the COVID-19 pandemic was to plan non-pharmaceutical interventions to reduce the spread of the infection while limiting the burden on the society and economy. With more data on the pandemic being generated, it became possible to model both the infection trends and intervention costs, transforming the creation of an intervention plan into a computational optimization problem. This paper proposes a framework developed to help policy-makers plan the best combination of non-pharmaceutical interventions and to change them over time. We developed a hybrid machine-learning epidemiological model to forecast the infection trends, aggregated the socio-economic costs from the literature and expert knowledge, and used a multi-objective optimization algorithm to find and evaluate various intervention plans. The framework is modular and easily adjustable to a real-world situation, it is trained and tested on data collected from almost all countries in the world, and its proposed intervention plans generally outperform those used in real life in terms of both the number of infections and intervention costs.

ObjectivesThis study aimed to investigate the associations between day-to-day work-related stress exposures (i.e., job demands and lack of job control), job strain, and next-day work engagement among office workers in academic settings. Additionally, we assessed the influence of psychological detachment and relaxation on next-day work engagement and tested for interaction effects of these recovery variables on the relationship between work-related stressors and next-day work engagement.MethodsOffice workers from two academic settings in Belgium and Slovenia were recruited. This study is based on an Ecological Momentary Assessment (EMA) with a 15-working day data collection period using our self-developed STRAW smartphone application. Participants were asked repeatedly about their work-related stressors, work engagement, and recovery experiences. Fixed-effect model testing using random intercepts was applied to investigate within- and between-participant levels.ResultsOur sample consisted of 55 participants and 2710 item measurements were analysed. A significant positive association was found between job control and next-day work engagement (β = 0.28, p < 0.001). Further, a significant negative association was found between job strain and next-day work engagement (β = -0.32, p = 0.05). Furthermore, relaxation was negatively associated with work engagement (β = -0.08, p = 0.03).ConclusionsThis study confirmed previous results, such as higher job control being associated with higher work engagement and higher job strain predicting lower work engagement. An interesting result was the association of higher relaxation after the working day with a lower next-day work engagement. Further research investigating fluctuations in work-related stressors, work engagement, and recovery experiences is required.

Understanding the growth pattern is important in view of child and adolescent development. Due to different tempo of growth and timing of adolescent growth spurt, individuals reach their adult height at different ages. Accurate models to assess the growth involve intrusive radiological methods whereas the predictive models based solely on height data are typically limited to percentiles and therefore rather inaccurate, especially during the onset of puberty. There is a need for more accurate non-invasive methods for height prediction that are easily applicable in the fields of sports and physical education, as well as in endocrinology. We developed a novel method, called Growth Curve Comparison (GCC), for height prediction, based on a large cohort of > 16,000 Slovenian schoolchildren followed yearly from ages 8 to 18. We compared the GCC method to the percentile method, linear regressor, decision tree regressor, and extreme gradient boosting. The GCC method outperformed the predictions of other methods over the entire age span both in boys and girls. The method was incorporated into a publicly available web application. We anticipate our method to be applicable also to other models predicting developmental outcomes of children and adolescents, such as for comparison of any developmental curves of anthropometric as well as fitness data. It can serve as a useful tool for assessment, planning, implementation, and monitoring of somatic and motor development of children and youth.

Abstract Background While chronic workplace stress is known to be associated with health-related outcomes like mental and cardiovascular diseases, research about day-to-day occupational stress is limited. This systematic review includes studies assessing stress exposures as work environment risk factors and stress outcomes, measured via self-perceived questionnaires and physiological stress detection. These measures needed to be assessed repeatedly or continuously via Ecological Momentary Assessment (EMA) or similar methods carried out in real-world work environments, to be included in this review. The objective was to identify work environment risk factors causing day-to-day stress. Methods The search strategies were applied in seven databases resulting in 11833 records after deduplication, of which 41 studies were included in a qualitative synthesis. Associations were evaluated by correlational analyses. Results The most commonly measured work environment risk factor was work intensity, while stress was most often framed as an affective response. Measures from these two dimensions were also most frequently correlated with each other and most of their correlation coefficients were statistically significant, making work intensity a major risk factor for day-to-day workplace stress. Conclusions This review reveals a diversity in methodological approaches in data collection and data analysis. More studies combining self-perceived stress exposures and outcomes with physiological measures are warranted.

Abstract This study tested the effectiveness of HeartMan—a mobile personal health system offering decisional support for management of congestive heart failure (CHF)—on health-related quality of life (HRQoL), self-management, exercise capacity, illness perception, mental and sexual health. A randomized controlled proof-of-concept trial (1:2 ratio of control:intervention) was set up with ambulatory CHF patients in stable condition in Belgium and Italy. Data were collected by means of a 6-min walking test and a number of standardized questionnaire instruments. A total of 56 (34 intervention and 22 control group) participants completed the study (77% male; mean age 63 years, sd 10.5). All depression and anxiety dimensions decreased in the intervention group (p

Fondazione Bruno Kessler is developing a mobile app prototype for empowering citizens to improve their health conditions through different lifestyle interventions that will be incorporated into a mobile application for lifestyle promotion of the Province of Trento in the context of the Trentino Salute 4.0 Competence Center. The envisioned interventions are based on promoting behaviour change in various domains such as physical activity, mental health and nutrition. In particular, the nutrition component is a self-monitoring module that collects dietary habits to analyse them and recommend healthier eating behaviours. Dietary assessment is completed using a Food Frequency Questionnaire on the Mediterranean diet that is presented to the user as a grid of images. The questionnaire returns feedback on 11 aspects of nutrition. Although the questionnaire used in the application only consists of 24 questions, it still could be a bit overwhelming and a bit crowded when shown on the screen. In this paper, we tried to find a machine-learning-based solution to reduce the number of questions in the questionnaire. We proposed a method that uses the user’s previous answers as additional information to find the goals that need more attention. We compared this method with a case where the subset of questions is randomly selected and with a case where the subset is chosen using feature selection. We also explored how large the subset should be to obtain good predictions. All the experiments are conducted as a multi-target regression problem, which means several goals are predicted simultaneously. The proposed method adjusts well to the user in question and has the slightest error when predicting the goals.

From 2018 to 2021, the Sussex-Huawei Locomotion-Transportation Recognition Challenge presented different scenarios in which participants were tasked with recognizing eight different modes of locomotion and transportation using sensor data from smartphones. In 2019, the main challenge was using sensor data from one location to recognize activities with sensors in another location, while in the following year, the main challenge was using the sensor data of one person to recognize the activities of other persons. We use these two challenge scenarios as a framework in which to analyze the effectiveness of different components of a machine-learning pipeline for activity recognition. We show that: (i) selecting an appropriate (location-specific) portion of the available data for training can improve the F1 score by up to 10 percentage points (p. p.) compared to a more naive approach, (ii) separate models for human locomotion and for transportation in vehicles can yield an increase of roughly 1 p. p., (iii) using semi-supervised learning can, again, yield an increase of roughly 1 p. p., and (iv) temporal smoothing of predictions with Hidden Markov models, when applicable, can bring an improvement of almost 10 p. p. Our experiments also indicate that the usefulness of advanced feature selection techniques and clustering to create person-specific models is inconclusive and should be explored separately in each use-case.

Starting renal replacement therapy (RRT) for patients with chronic kidney disease (CKD) at an optimal time, either with hemodialysis or kidney transplantation, is crucial for patient's well-being and for successful management of the condition. In this paper, we explore the possibilities of creating forecasting models to predict the onset of RRT 3, 6, and 12 months from the time of the patient's first diagnosis with CKD, using only the comorbidities data from National Health Insurance from Taiwan. The goal of this study was to see whether a limited amount of data (including comorbidities but not considering laboratory values which are expensive to obtain in low- and medium-income countries) can provide a good basis for such predictive models. On the other hand, in developed countries, such models could allow policy-makers better planning and allocation of resources for treatment. Using data from 8,492 patients, we obtained the area under the receiver operating characteristic curve (AUC) of 0.773 for predicting RRT within 12 months from the time of CKD diagnosis. The results also show that there is no additional advantage in focusing only on patients with diabetes in terms of prediction performance. Although these results are not as such suitable for adoption into clinical practice, the study provides a strong basis and a variety of approaches for future studies of forecasting models in healthcare.

Introduction: The preferences of people with profound intellectual and multiple disabilities (PIMD) often remain unfulfilled since it stays challenging to decode their idiosyncratic behavior resulting in a negative impact on their quality of life (QoL). Physiological data (i.e. heart rate (variability) and motion data) might be the missing piece for identifying emotions of people with PIMD, which positively affects their QoL. Method: Machine learning (ML) processes and statistical analyses are integrated to discern and predict the potential relationship between physiological data and emotional states (i.e. numerical emotional states, descriptive emotional states and emotional arousal) in everyday interactions and activities of two participants with PIMD. Results: Emotional profiles were created enabling a differentiation of the individual emotional behavior. Using ML classifiers and statistical analyses, the results regarding the phases partially confirm previous research, and the findings for the descriptive emotional states were good and even better for the emotional arousal. Conclusion: The results show the potential of the emotional profiles especially for practitioners and the possibility to get a better insight into the emotional experience of people with PIMD including physiological data. This seems to be the missing piece to better recognize emotions of people with PIMD with a positive impact on their QoL.

Chronic kidney disease (CKD) represents a heavy burden on the healthcare system because of the increasing number of patients, high risk of progression to end-stage renal disease, and poor prognosis of morbidity and mortality. The aim of this study is to develop a machine-learning model that uses the comorbidity and medication data obtained from Taiwan’s National Health Insurance Research Database to forecast the occurrence of CKD within the next 6 or 12 months before its onset, and hence its prevalence in the population. A total of 18,000 people with CKD and 72,000 people without CKD diagnosis were selected using propensity score matching. Their demographic, medication and comorbidity data from their respective two-year observation period were used to build a predictive model. Among the approaches investigated, the Convolutional Neural Networks (CNN) model performed best with a test set AUROC of 0.957 and 0.954 for the 6-month and 12-month predictions, respectively. The most prominent predictors in the tree-based models were identified, including diabetes mellitus, age, gout, and medications such as sulfonamides and angiotensins. The model proposed in this study could be a useful tool for policymakers in predicting the trends of CKD in the population. The models can allow close monitoring of people at risk, early detection of CKD, better allocation of resources, and patient-centric management.

Understanding people’s eating habits plays a crucial role in interventions promoting a healthy lifestyle. This requires objective measurement of the time at which a meal takes place, the duration of the meal, and what the individual eats. Smartwatches and similar wrist-worn devices are an emerging technology that offers the possibility of practical and real-time eating monitoring in an unobtrusive, accessible, and affordable way. To this end, we present a novel approach for the detection of eating segments with a wrist-worn device and fusion of deep and classical machine learning. It integrates a novel data selection method to create the training dataset, and a method that incorporates knowledge from raw and virtual sensor modalities for training with highly imbalanced datasets. The proposed method was evaluated using data from 12 subjects recorded in the wild, without any restriction about the type of meals that could be consumed, the cutlery used for the meal, or the location where the meal took place. The recordings consist of data from accelerometer and gyroscope sensors. The experiments show that our method for detection of eating segments achieves precision of 0.85, recall of 0.81, and F1-score of 0.82 in a person-independent manner. The results obtained in this study indicate that reliable eating detection using in the wild recorded data is possible with the use of wearable sensors on the wrist.

Contact-free sensors offer important advantages compared to traditional wearables. Radio-frequency sensors (e.g., radars) offer the means to monitor cardiorespiratory activity of people without compromising their privacy, however, only limited information can be obtained via movement, traditionally related to heart or breathing rate. We investigated whether five complex hemodynamics scenarios (resting, apnea simulation, Valsalva maneuver, tilt up and tilt down on a tilt table) can be classified directly from publicly available contact and radar input signals in an end-to-end deep learning approach. A series of robust k-fold cross-validation evaluation experiments were conducted in which neural network architectures and hyperparameters were optimized, and different data input modalities (contact, radar and fusion) and data types (time and frequency domain) were investigated. We achieved reasonably high accuracies of 88% for contact, 83% for radar and 88% for fusion of modalities. These results are valuable in showing large potential of radar sensing even for more complex scenarios going beyond just heart and breathing rate. Such contact-free sensing can be valuable for fast privacy-preserving hospital screenings and for cases where traditional werables are impossible to use.

Context recognition using wearable devices is a mature research area, but one of the biggest issues it faces is the high energy consumption of the device that is sensing and processing the data. In this work we propose three different methods for optimizing its energy use. We also show how to combine all three methods to further increase the energy savings. The methods work by adapting system settings (sensors used, sampling frequency, duty cycling, etc.) to both the detected context and directly to the sensor data. This is done by mathematically modeling the influence of different system settings and using multiobjective optimization to find the best ones. The proposed methodology is tested on four different context-recognition tasks where we show that it can generate accurate energy-efficient solutions—in one case reducing energy consumption by 95% in exchange for only four percentage points of accuracy. We also show that the method is general, requires next to no expert knowledge about the domain being optimized, and that it outperforms two approaches from the related work.

Food frequency questionnaires (FFQs) are the most commonly selected tools in nutrition monitoring, as they are inexpensive, easily implemented and provide useful information regarding dietary intake. They are usually carefully drafted by experts from nutritional and/or medical fields and can be validated by using other dietary monitoring techniques. FFQs can get very extensive, which could indicate that some of the questions are less significant than others and could be omitted without losing too much information. In this paper, machine learning is used to explore how reducing the number of questions affects the predicted nutrient values and diet quality score. The paper addresses the problem of removing redundant questions and finding the best subset of questions in the Extended Short Form Food Frequency Questionnaire (ESFFFQ), developed as part of the H2020 project WellCo. Eight common machine-learning algorithms were compared on different subsets of questions by using the PROMETHEE method, which compares methods and subsets via multiple performance measures. According to the results, for some of the targets, specifically sugar intake, fiber intake and protein intake, a smaller subset of questions are sufficient to predict diet quality scores. Additionally, for smaller subsets of questions, machine-learning algorithms generally perform better than statistical methods for predicting intake and diet quality scores. The proposed method could therefore be useful for finding the most informative subsets of questions in other FFQs as well. This could help experts develop FFQs that provide the necessary information and are not overbearing for those answering.

Diabetes is both heavily affected by the patients’ lifestyle, and it affects their lifestyle. Most diabetic patients can manage the disease without technological assistance, so we should not burden them with technology unnecessarily, but lifestylemonitoring technology can still be beneficial both for patients and their physicians. Because of that we developed an approach to lifestyle monitoring that uses the smartphone, which most patients already have. The approach consists of three steps. First, a number of features are extracted from the data acquired by smartphone sensors, such as the user’s location from GPS coordinates and visible wi-fi access points, and the physical activity from accelerometer data. Second, several classifiers trained by machine learning are used to recognise the user’s activity, such as work, exercise or eating. And third, these activities are refined by symbolic reasoning encoded in Event Calculus. The approach was trained and tested on five people who recorded their activities for two weeks each. Its classification accuracy was 0.88.

This study introduces two datasets for multimodal research on cognitive load inference and personality traits. Different to other datasets in Affective Computing, which disregard participants’ personality traits or focus only on emotions, stress, or cognitive load from one specific task, the participants in our experiments performed seven different tasks in total. In the first dataset, 23 participants played a varying difficulty (easy, medium, and hard) game on a smartphone. In the second dataset, 23 participants performed six psychological tasks on a PC, again with varying difficulty. In both experiments, the participants filled personality trait questionnaires and marked their perceived cognitive load using NASA-TLX after each task. Additionally, the participants’ physiological response was recorded using a wrist device measuring heart rate, beat-to-beat intervals, galvanic skin response, skin temperature, and three-axis acceleration. The datasets allow multimodal study of physiological responses of individuals in relation to their personality and cognitive load. Various analyses of relationships between personality traits, subjective cognitive load (i.e., NASA-TLX), and objective cognitive load (i.e., task difficulty) are presented. Additionally, baseline machine learning models for recognizing task difficulty are presented, including a multitask learning (MTL) neural network that outperforms single-task neural network by simultaneously learning from the two datasets. The datasets are publicly available to advance the field of cognitive load inference using commercially available devices.

This paper addresses patient-reported outcomes (PROs) and telemonitoring in congestive heart failure (CHF), both increasingly important topics. The interest in CHF trials is shifting from hard end-points such as hospitalization and mortality, to softer end-points such health-related quality of life. However, the relation of these softer end-points to objective parameters is not well studied. Telemonitoring is suitable for collecting both patient-reported outcomes and objective parameters. Most telemonitoring studies, however, do not take full advantage of the available sensor technology and intelligent data analysis. The Chiron clinical observational study was performed among 24 CHF patients (17 men and 7 women, age 62.9 ± 9.4 years, 15 NYHA class II and 9 class III, 10 of ishaemic, aetiology, 6 dilated, 2 valvular, and 6 of multiple aetiologies or cardiomyopathy) in Italy and UK. A large number of physiological and ambient parameters were collected by wearable and other devices, together with PROs describing how well the patients felt, over 1,086 days of observation. The resulting data were mined for relations between the objective parameters and the PROs. The objective parameters (humidity, ambient temperature, blood pressure, SpO2, and sweeting intensity) could predict the PROs with accuracies up to 86% and AUC up to 0.83, making this the first report providing evidence for ambient and physiological parameters to be objectively related to PROs in CHF patients. We also analyzed the relations in the predictive models, gaining some insights into what affects the feeling of health, which was also generally not attempted in previous investigations. The paper strongly points to the possibility of using PROs as primary end-points in future trials.

Blood pressure (BP) is a direct indicator of hypertension, a dangerous and potentially deadly condition. Regular monitoring of BP is thus important, but many people have aversion towards cuff-based devices, and their limitation is that they can only be used at rest. Using just a photoplethysmogram (PPG) to estimate BP is a potential solution investigated in our study. We analyzed the MIMIC III database for high-quality PPG and arterial BP waveforms, resulting in over 700 h of signals after preprocessing, belonging to 510 subjects. We then used the PPG alongside its first and second derivative as inputs into a novel spectro-temporal deep neural network with residual connections. We have shown in a leave-one-subject-out experiment that the network is able to model the dependency between PPG and BP, achieving mean absolute errors of 9.43 for systolic and 6.88 for diastolic BP. Additionally we have shown that personalization of models is important and substantially improves the results, while deriving a good general predictive model is difficult. We have made crucial parts of our study, especially the list of used subjects and our neural network code, publicly available, in an effort to provide a solid baseline and simplify potential comparison between future studies on an explicit MIMIC III subset.

The recognition of the user’s context with wearable sensing systems is a common problem in ubiquitous computing. However, the typically small battery of such systems often makes continuous recognition impractical. The strain on the battery can be reduced if the sensor setting is adapted to each context. We propose a method that efficiently finds near-optimal sensor settings for each context. It uses Markov chains to simulate the behavior of the system in different configurations and the multi-objective genetic algorithm to find a set of good non-dominated configurations. The method was evaluated on three real-life datasets and found good trade-offs between the system’s energy expenditure and the system’s accuracy. One of the solutions, for example, consumed five-times less energy than the default one, while sacrificing only two percentage points of accuracy.

Epitope-antibody-reactivities (EAR) of intravenous immunoglobulins (IVIGs) determined for 75,534 peptides by microarray analysis demonstrate that roughly 9% of peptides derived from 870 different human protein sequences react with antibodies present in IVIG. Computational prediction of linear B cell epitopes was conducted using machine learning with an ensemble of classifiers in combination with position weight matrix (PWM) analysis. Machine learning slightly outperformed PWM with area under the curve (AUC) of 0.884 vs. 0.849. Two different types of epitope-antibody recognition-modes (Type I EAR and Type II EAR) were found. Peptides of Type I EAR are high in tyrosine, tryptophan and phenylalanine, and low in asparagine, glutamine and glutamic acid residues, whereas for peptides of Type II EAR it is the other way around. Representative crystal structures present in the Protein Data Bank (PDB) of Type I EAR are PDB 1TZI and PDB 2DD8, while PDB 2FD6 and 2J4W are typical for Type II EAR. Type I EAR peptides share predicted propensities for being presented by MHC class I and class II complexes. The latter interaction possibly favors T cell-dependent antibody responses including IgG class switching. Peptides of Type II EAR are predicted not to be preferentially presented by MHC complexes, thus implying the involvement of T cell-independent IgG class switch mechanisms. The high extent of IgG immunoglobulin reactivity with human peptides implies that circulating IgG molecules are prone to bind to human protein/peptide structures under non-pathological, non-inflammatory conditions. A webserver for predicting EAR of peptide sequences is available at www.sysmed-immun.eu/EAR.

Although wearable accelerometers can successfully recognize activities and detect falls, their adoption in real life is low because users do not want to wear additional devices. A possible solution is an accelerometer inside a wrist device/smartwatch. However, wrist placement might perform poorly in terms of accuracy due to frequent random movements of the hand. In this paper we perform a thorough, large-scale evaluation of methods for activity recognition and fall detection on four datasets. On the first two we showed that the left wrist performs better compared to the dominant right one, and also better compared to the elbow and the chest, but worse compared to the ankle, knee and belt. On the third (Opportunity) dataset, our method outperformed the related work, indicating that our feature-preprocessing creates better input data. And finally, on a real-life unlabeled dataset the recognized activities captured the subject’s daily rhythm and activities. Our fall-detection method detected all of the fast falls and minimized the false positives, achieving 85% accuracy on the first dataset. Because the other datasets did not contain fall events, only false positives were evaluated, resulting in 9 for the second, 1 for the third and 15 for the real-life dataset (57 days data).

Background: The number of elderly people in need of help with the activities of daily living in the EU is rapidly increasing, while the number of young workers is decreasing. Elderly care will, therefore, also have to be provided by intelligent computer systems. Methods: A prototype elderly-care system, developed at the Jožef Stefan Institute, mostly as part of the Confidence project, is presented. The prototype detects falls and behavior changes in the elderly. It learns from experience and is based on intelligent interpretation of movement patterns. Three sets of tests were performed to evaluate its properties on various subjects when engaged in normal activities, falling and imitations of several health problems under medical supervision. The key novelty was in locationbased sensors and advanced intelligent methods. Results: The prototype using the Ubisense sensor system, which detects the locations of tags worn on the body, correctly recognized 96 % of falls, significantly outperforming simple accelerometer- based systems. In addition, it recognized up to 99 % of abnormal behavior. Conclusions: Experimental results showed that an intelligent system coupled with advanced location sensors can achieve the level of performance needed in real life. The system offers significantly better performance than commercially available solutions, and once the price of sensors decreases, its widespread application seems likely.

Abstract Tools for predicting COVID-19 outcomes enable personalized healthcare, potentially easing the disease burden. This collaborative study by 15 institutions across Europe aimed to develop a machine learning model for predicting the risk of in-hospital mortality post-SARS-CoV-2 infection. Blood samples and clinical data from 1286 COVID-19 patients collected from 2020 to 2023 across four cohorts in Europe and Canada were analyzed, with 2906 long non-coding RNAs profiled using targeted sequencing. From a discovery cohort combining three European cohorts and 804 patients, age and the long non-coding RNA LEF1-AS1 were identified as predictive features, yielding an AUC of 0.83 (95% CI 0.82–0.84) and a balanced accuracy of 0.78 (95% CI 0.77–0.79) with a feedforward neural network classifier. Validation in an independent Canadian cohort of 482 patients showed consistent performance. Cox regression analysis indicated that higher levels of LEF1-AS1 correlated with reduced mortality risk (age-adjusted hazard ratio 0.54, 95% CI 0.40–0.74). Quantitative PCR validated LEF1-AS1’s adaptability to be measured in hospital settings. Here, we demonstrate a promising predictive model for enhancing COVID-19 patient management.