Can machine learning models predict maternal and newborn healthcare providers' perception of safety during the COVID-19 pandemic? A cross-sectional study of a global online survey.

BACKGROUND: Maternal and newborn healthcare providers are essential professional groups vulnerable to physical and psychological risks associated with the COVID-19 pandemic. This study uses machine learning algorithms to create a predictive tool for maternal and newborn healthcare providers' perception of being safe in the workplace globally during the pandemic. METHODS: We used data collected between 24 March and 5 July 2020 through a global online survey of maternal and newborn healthcare providers. The questionnaire was available in 12 languages. To predict healthcare providers' perception of safety in the workplace, we used features collected in the questionnaire, in addition to publicly available national economic and COVID-19-related factors. We built, trained and tested five machine learning models: Support Vector Machine (SVM), Random Forest (RF), XGBoost, CatBoost and Artificial Neural Network (ANN) for classification and regression. We extracted from RF models the relative contribution of features in output prediction. RESULTS: Models included data from 941 maternal and newborn healthcare providers from 89 countries. ML models performed well in classification and regression tasks, whereby RF had 82% cross-validated accuracy for classification, and CatBoost with 0.46 cross-validated root mean square error for regression. In both classification and regression, the most important features contributing to output prediction were classified as three themes: (1) information accessibility, clarity and quality; (2) availability of support and means of protection; and (3) COVID-19 epidemiology. CONCLUSION: This study identified salient features contributing to maternal and newborn healthcare providers perception of safety in the workplace. The developed tool can be used by health systems globally to allow real-time learning from data collected during a health system shock. By responding in real-time to the needs of healthcare providers, health systems could prevent potential negative consequences on the quality of care offered to women and newborns.

A speckle-tracking strain-based artificial neural network model to differentiate cardiomyopathy type.

Objectives. In heart failure, invasive angiography is often employed to differentiate ischaemic from non-ischaemic cardiomyopathy. We aim to examine the predictive value of echocardiographic strain features alone and in combination with other features to differentiate ischaemic from non-ischaemic cardiomyopathy, using artificial neural network (ANN) and logistic regression modelling. Design. We retrospectively identified 204 consecutive patients with an ejection fraction <50% and a diagnostic angiogram. Patients were categorized as either ischaemic (n = 146) or non-ischaemic cardiomyopathy (n = 58). For each patient, left ventricular strain parameters were obtained. Additionally, regional wall motion abnormality, 13 electrocardiographic (ECG) features and six demographic features were retrieved for analysis. The entire cohort was randomly divided into a derivation and a validation cohort. Using the parameters retrieved, logistic regression and ANN models were developed in the derivation cohort to differentiate ischaemic from non-ischaemic cardiomyopathy, the models were then tested in the validation cohort. Results. A final strain-based ANN model, full feature ANN model and full feature logistic regression model were developed and validated, F1 scores were 0.82, 0.79 and 0.63, respectively. Conclusions. Both ANN models were more accurate at predicting cardiomyopathy type than the logistic regression model. The strain-based ANN model should be validated in other cohorts. This model or similar models could be used to aid the diagnosis of underlying heart failure aetiology in the form of the online calculator (https://cimti.usj.edu.lb/strain/index.html) or built into echocardiogram software.

Artificial neural network-based model enhances risk stratification and reduces non-invasive cardiac stress imaging compared to Diamond-Forrester and Morise risk assessment models: A prospective study.

BACKGROUND: Coronary artery disease (CAD) accounts for more than half of all cardiovascular events. Stress testing remains the cornerstone for non-invasive assessment of patients with possible or known CAD. Clinical utilization reviews show that most patients presenting for evaluation of stable CAD by stress testing are categorized as low risk prior to the test. Attempts to enhance risk stratification of individuals who are sent for stress testing seem to be more in need today. The present study compares artificial neural networks (ANN)-based prediction models to the other risk models being used in practice (the Diamond-Forrester and the Morise models). METHODS: In our study, we prospectively recruited patients who were 19 years of age or older, and were being evaluated for coronary artery disease with imaging-based stress tests. For ANN, the network architecture employed a systematic method, where the number of neurons is changed incrementally, and bootstrapping was performed to evaluate the accuracy of the models. RESULTS: We prospectively enrolled 486 patients. The mean age of patients undergoing stress test was 55.2 ± 11.2 years, 35% were women, and 12% had a positive stress test for ischemic heart disease. When compared to Diamond-Forrester and Morise risk models, the ANN model for predicting ischemia provided higher discriminatory power (DP)(1.61), had a negative predictive value of 98%, Sensitivity 91% [81%-97%], Specificity 65% [60%-79%], positive predictive value 26%, and a potential 59% reduction of non-invasive imaging. CONCLUSION: The ANN models improved risk stratification when compared to the other risk scores (Diamond-Forrester and Morise) with a 98% negative predictive value and a significant potential reduction in non-invasive imaging tests.

Robotics in Nursing: A Scoping Review.

PURPOSE: Driven by the shortage in qualified nurses and the high percentage of aging populations, the past decade has witnessed a significant growth in the use of robots in nursing, especially in countries like Japan. This article is a scoping review of the different tracks in which robots are used in nursing. Whereas assistive robots are used for physical care, including service and monitoring tasks, social assistive robots focus on the cognitive and emotional well-being of patients in need of companionship. METHODS: A total of six electronic databases were used in the search for journal papers and conference proceedings. The key words used in searching the databases were nursing OR nurses, AND robots OR robotics. Topics covering surgical robotics, nursing education robotics, and clinical procedures were excluded. FINDINGS: A total of 1,758 articles were retrieved, from which 69 articles were included in the final review. The analysis of the chosen papers led to the categorization of robots into two main categories: assistive robots and social assistive robots. CONCLUSIONS: After a detailed review of the state of the art in assistive robots and social assistive robots, an insight into the future of robotics in this field is provided. The recommendations include the need to intensify research on human robot interaction, greater focus on monitoring robots, and analysis of the psychological barriers that need to be surmounted to achieve more tolerance and higher acceptance of robots. CLINICAL RELEVANCE: For researchers and developers to provide suitable technological solutions, a full understanding of robotics in nursing is needed. An overview of the most recent applications and their proper categorization is key to finding areas for contribution.

CENTERA: a centralized trust-based efficient routing protocol with authentication for wireless sensor networks.

In this paper, we present CENTERA, a CENtralized Trust-based Efficient Routing protocol with an appropriate authentication scheme for wireless sensor networks (WSN). CENTERA utilizes the more powerful base station (BS) to gather minimal neighbor trust information from nodes and calculate the best routes after isolating different types of "bad" nodes. By periodically accumulating these simple local observations and approximating the nodes' battery lives, the BS draws a global view of the network, calculates three quality metrics-maliciousness, cooperation, and compatibility-and evaluates the Data Trust and Forwarding Trust values of each node. Based on these metrics, the BS isolates "bad", "misbehaving" or malicious nodes for a certain period, and put some nodes on probation. CENTERA increases the node's bad/probation level with repeated "bad" behavior, and decreases it otherwise. Then it uses a very efficient method to distribute the routing information to "good" nodes. Based on its target environment, and if required, CENTERA uses an authentication scheme suitable for severely constrained nodes, ranging from the symmetric RC5 for safe environments under close administration, to pairing-based cryptography (PBC) for hostile environments with a strong attacker model. We simulate CENTERA using TOSSIM and verify its correctness and show some energy calculations.

Improved accuracy of anticoagulant dose prediction using a pharmacogenetic and artificial neural network-based method.

BACKGROUND: The unpredictability of acenocoumarol dose needed to achieve target blood thinning level remains a challenge. We aimed to apply and compare a pharmacogenetic least-squares model (LSM) and artificial neural network (ANN) models for predictions of acenocoumarol dosing. METHODS: LSM and ANN models were used to analyze previously collected data on 174 participants (mean age: 67.45 SD 13.49 years) on acenocoumarol maintenance therapy. The models were based on demographics, lifestyle habits, concomitant diseases, medication intake, target INR, and genotyping results for CYP2C9 and VKORC1. LSM versus ANN performance comparisons were done by two methods: by randomly splitting the data as 50 % derivation and 50 % validation cohort followed by a bootstrap of 200 iterations, and by a 10-fold leave-one-out cross-validation technique. RESULTS: The ANN-based pharmacogenetic model provided higher accuracy and larger R value than all other LSM-based models. The accuracy percentage improvement ranged between 5 % and 24 % for the derivation cohort and between 12 % and 25 % for the validation cohort. The increase in R value ranged between 6 % and 31 % for the derivation cohort and between 2 % and 31 % for the validation cohort. ANN increased the percentage of accurately dosed subjects (mean absolute error ≤1 mg/week) by 14.1 %, reduced the percentage of mis-dosed subjects (mean absolute error 2-3 mg/week) by 7.04 %, and reduced the percentage of grossly mis-dosed subjects (mean absolute error ≥4 mg/week) by 24 %. CONCLUSIONS: ANN-based pharmacogenetic guidance of acenocoumarol dosing reduces the error in dosing to achieve target INR. These results need to be ascertained in a prospective study.

Tackling the COVID-19 infodemic among Syrian refugees in Lebanon: Development and evaluation of the "Wikaytek" tool.

Objective: The COVID-19 infodemic has been a global public health challenge, especially affecting vulnerable populations such as Syrian refugees with limited internet access and functional, health, digital, and media literacies. To address this problem, we developed Wikaytek, a software to diffuse reliable COVID-19 information using WhatsApp, the preferred communication channel among Syrian refugees. In this paper, we describe the systematic development of the tool. Methods: We undertook a pilot study guided by the Humanitarian Engineering Initiative (HEI)'s user-centered design framework, comprising five stages: (a) user research, including needs assessment and desk review of interventions with target users; (b) concept design based on platform and source selection, message format, concept testing, and architecture design; (c) prototyping and implementation, encompassing software development and system operation; (d) user testing (alpha and beta); and (e) evaluation through software analytics and user interviews. We reported a qualitative process evaluation. Results: Wikaytek scrapes validated and reliable COVID-19-related information from reputable sources on Twitter, automatically translates it into Arabic, attaches relevant media (images/video), and generates an audio format using Google text-to-speech. Then, messages are broadcast to WhatsApp. Our evaluation shows that users appreciate receiving "push" information from reliable sources they can trust and prefer the audio format over text. Conclusions: Wikaytek is a useful and well-received software for diffusing credible information on COVID-19 among Syrian refugees with limited literacy, as it complements the texts with audio messages. The tool can be adapted to diffuse messages about other public health issues among vulnerable communities, extending its scope and reach in humanitarian settings.

Data Management and Privacy Policy of COVID-19 Contact-Tracing Apps: Systematic Review and Content Analysis.

BACKGROUND: COVID-19 digital contact-tracing apps were created to assist public health authorities in curbing the pandemic. These apps require users' permission to access specific functions on their mobile phones, such as geolocation, Bluetooth or Wi-Fi connections, or personal data, to work correctly. As these functions have privacy repercussions, it is essential to establish how contact-tracing apps respect users' privacy. OBJECTIVE: This study aimed to systematically map existing contact-tracing apps and evaluate the permissions required and their privacy policies. Specifically, we evaluated the type of permissions, the privacy policies' readability, and the information included in them. METHODS: We used custom Google searches and existing lists of contact-tracing apps to identify potentially eligible apps between May 2020 and November 2021. We included contact-tracing or exposure notification apps with a Google Play webpage from which we extracted app characteristics (eg, sponsor, number of installs, and ratings). We used Exodus Privacy to systematically extract the number of permissions and classify them as dangerous or normal. We computed a Permission Accumulated Risk Score representing the threat level to the user's privacy. We assessed the privacy policies' readability and evaluated their content using a 13-item checklist, which generated a Privacy Transparency Index. We explored the relationships between app characteristics, Permission Accumulated Risk Score, and Privacy Transparency Index using correlations, chi-square tests, or ANOVAs. RESULTS: We identified 180 contact-tracing apps across 152 countries, states, or territories. We included 85.6% (154/180) of apps with a working Google Play page, most of which (132/154, 85.7%) had a privacy policy document. Most apps were developed by governments (116/154, 75.3%) and totaled 264.5 million installs. The average rating on Google Play was 3.5 (SD 0.7). Across the 154 apps, we identified 94 unique permissions, 18% (17/94) of which were dangerous, and 30 trackers. The average Permission Accumulated Risk Score was 22.7 (SD 17.7; range 4-74, median 16) and the average Privacy Transparency Index was 55.8 (SD 21.7; range 5-95, median 55). Overall, the privacy documents were difficult to read (median grade level 12, range 7-23); 67% (88/132) of these mentioned that the apps collected personal identifiers. The Permission Accumulated Risk Score was negatively associated with the average App Store ratings (r=-0.20; P=.03; 120/154, 77.9%) and Privacy Transparency Index (r=-0.25; P<.001; 132/154, 85.7%), suggesting that the higher the risk to one's data, the lower the apps' ratings and transparency index. CONCLUSIONS: Many contact-tracing apps were developed covering most of the planet but with a relatively low number of installs. Privacy-preserving apps scored high in transparency and App Store ratings, suggesting that some users appreciate these apps. Nevertheless, privacy policy documents were difficult to read for an average audience. Therefore, we recommend following privacy-preserving and transparency principles to improve contact-tracing uptake while making privacy documents more readable for a wider public.

Augmented Reality Meets Artificial Intelligence in Robotics: A Systematic Review.

Recently, advancements in computational machinery have facilitated the integration of artificial intelligence (AI) to almost every field and industry. This fast-paced development in AI and sensing technologies have stirred an evolution in the realm of robotics. Concurrently, augmented reality (AR) applications are providing solutions to a myriad of robotics applications, such as demystifying robot motion intent and supporting intuitive control and feedback. In this paper, research papers combining the potentials of AI and AR in robotics over the last decade are presented and systematically reviewed. Four sources for data collection were utilized: Google Scholar, Scopus database, the International Conference on Robotics and Automation 2020 proceedings, and the references and citations of all identified papers. A total of 29 papers were analyzed from two perspectives: a theme-based perspective showcasing the relation between AR and AI, and an application-based analysis highlighting how the robotics application was affected. These two sections are further categorized based on the type of robotics platform and the type of robotics application, respectively. We analyze the work done and highlight some of the prevailing limitations hindering the field. Results also explain how AR and AI can be combined to solve the model-mismatch paradigm by creating a closed feedback loop between the user and the robot. This forms a solid base for increasing the efficiency of the robotic application and enhancing the user's situational awareness, safety, and acceptance of AI robots. Our findings affirm the promising future for robust integration of AR and AI in numerous robotic applications.

Directing and Orienting ICT Healthcare Solutions to Address the Needs of the Aging Population.

BACKGROUND: With an aging population, it is essential to maintain good health and autonomy for as long as possible. Instead of hospitalisation or institutionalisation, older people with chronic conditions can be assisted in their own home with numerous "smart" devices that support them in their activities of daily living, manage their medical conditions, and prevent fall incidents. Information and Communication Technology (ICT) solutions facilitate the monitoring and management of older people's health to improve quality of life and physical activity with a decline in caregivers' burden. METHOD: The aim of this paper was to conduct a systematic literature review to analyse the state of the art of ICT solutions for older people with chronic conditions, and the impact of these solutions on their quality of life from a biomedical perspective. RESULTS: By analysing the literature on the available ICT proposals, it is shown that different approaches have been deployed by noticing that the more cross-interventions are merged then the better the results are, but there is still no evidence of the effects of ICT solutions on older people's health outcomes. Furthermore, there are still unresolved ethical and legal issues. CONCLUSION: While there has been much research and development in healthcare ICT solutions for the aging population, ICT solutions still need significant development in order to be user-oriented, affordable, and to manage chronic conditions in the aging wider population.

A comparison of artificial intelligence-based algorithms for the identification of patients with depressed right ventricular function from 2-dimentional echocardiography parameters and clinical features.

BACKGROUND: Recognizing low right ventricular (RV) function from 2-dimentiontial echocardiography (2D-ECHO) is challenging when parameters are contradictory. We aim to develop a model to predict low RV function integrating the various 2D-ECHO parameters in reference to cardiac magnetic resonance (CMR)-the gold standard. METHODS: We retrospectively identified patients who underwent a 2D-ECHO and a CMR within 3 months of each other at our institution (American University of Beirut Medical Center). We extracted three parameters (TAPSE, S' and FACRV) that are classically used to assess RV function. We have assessed the ability of 2D-ECHO derived parameters and clinical features to predict RV function measured by the gold standard CMR. We compared outcomes from four machine learning algorithms, widely used in the biomedical community to solve classification problems. RESULTS: One hundred fifty-five patients were identified and included in our study. Average age was 43±17.1 years old and 52/156 (33.3%) were females. According to CMR, 21 patients were identified to have RV dysfunction, with an RVEF of 34.7%±6.4%, as opposed to 54.7%±6.7% in the normal RV population (P<0.0001). The Random Forest model was able to detect low RV function with an AUC =0.80, while general linear regression performed poorly in our population with an AUC of 0.62. CONCLUSIONS: In this study, we trained and validated an ML-based algorithm that could detect low RV function from clinical and 2D-ECHO parameters. The algorithm has two advantages: first, it performed better than general linear regression, and second, it integrated the various 2D-ECHO parameters.

Bio-inspired upper limb soft exoskeleton to reduce stroke-induced complications.

Stroke has become the leading cause of disability and the second-leading cause of mortality worldwide. Dyskinesia complications are the major reason of these high death and disability rates. As a tool for rapid motion function recovery in stroke patients, exoskeleton robots can reduce complications and thereby decrease stroke mortality rates. However, existing exoskeleton robots interfere with the wearer's natural motion and damage joints and muscles due to poor human-machine coupling. In this paper, a novel ergonomic soft bionic exoskeleton robot with 7 degrees of freedom was proposed to address these problems based on the principles of functional anatomy and sports biomechanics. First, the human motion system was analysed according to the functional anatomy, and the muscles were modelled as tension lines. Second, a soft bionic robot was established based on the musculoskeletal tension line model. Third, a robot control method mimicking human muscle control principles was proposed and optimized on a humanoid platform manufactured using 3D printing. After the control method was optimized, the motion trajectory similarities between humans and the platform exceeded 87%. Fourth, the force-assisted effect was tested based on electromyogram signals, and the results showed that muscle signals decreased by 58.17% after robot assistance. Finally, motion-assistance experiments were performed with stroke patients. The joint movement level increased by 174% with assistance, which allowed patients to engage in activities of daily living. With this robot, stroke patients could recover their motion functions, preventing complications and decreasing fatality and disability rates.

Examining the need & potential for biomedical engineering to strengthen health care delivery for displaced populations & victims of conflict.

Conflict and the subsequent displacement of populations creates unique challenges in the delivery of quality health care to the affected population. Equitable access to quality care demands a multi-pronged strategy with a growing need, and role, for technological innovation to address these challenges. While there have been significant contributions towards alleviating the burden of conflict via data informatics and analytics, communication technology, and geographic information systems, little has been done within biomedical engineering. This article elaborates on the causes for gaps in biomedical innovation for refugee populations affected by conflict, tackles preconceived notions, takes stock of recent developments in promising technologies to address these challenges, and identifies tangible action items to create a stronger and sustainable pipeline for biomedical technological innovation to improve the health and well-being of an increasing group of vulnerable people around the world.

Haptic and Tactile Adjectives Are Consistently Mapped onto Color Space.

Cross-modal associations refer to non-arbitrary associations of features across sensory modalities. Such associations have been observed between many different sensory features. One association that has rarely been studied so far is between touch and color. In this study, participants were asked to match tactile and haptic adjectives to color samples shown individually on a screen. They could select one to 11 tactile and haptic terms, presented in 11 pairs of opposed adjectives. The results showed a regular pattern in the way tactile and haptic terms were matched to color. Our results further revealed that the colors to which tactile and haptic terms were matched did not fall within the boundaries of color lexical categories, suggesting that the associations were not based on lexicon--despite the frequent occurrence of linguistic expressions such as 'soft pink', not all colors called 'pink' were matched to 'soft'. In contrast with one recent study, the distribution of tactile and haptic terms across the Munsell array suggests that along with brightness and chroma, hue was also relevant to participants' responses. Specifically in the case of hue, several opposed adjectives were relatively well matched to opposed colors, along the orthogonal Yellow/Blue and Red/Green axes, which are suggested to structure the space of hue experience. Possible accounts of these results are considered.

Artificial neural network modeling enhances risk stratification and can reduce downstream testing for patients with suspected acute coronary syndromes, negative cardiac biomarkers, and normal ECGs.

Despite uncertain yield, guidelines endorse routine stress myocardial perfusion imaging (MPI) for patients with suspected acute coronary syndromes, unremarkable serial electrocardiograms, and negative troponin measurements. In these patients, outcome prediction and risk stratification models could spare unnecessary testing. This study therefore investigated the use of artificial neural networks (ANN) to improve risk stratification and prediction of MPI and angiographic results. We retrospectively identified 5354 consecutive patients referred from the emergency department for rest-stress MPI after serial negative troponins and normal ECGs. Patients were risk stratified according to thrombolysis in myocardial infarction (TIMI) scores, ischemia was defined as >5 % reversible perfusion defect, and obstructive coronary artery disease was defined as >50 % angiographic obstruction. For ANN, the network architecture employed a systematic method where the number of neurons is changed incrementally, and bootstrapping was performed to evaluate the accuracy of the models. Compared to TIMI scores, ANN models provided improved discriminatory power. With regards to MPI, an ANN model could reduce testing by 59 % and maintain a 96 % negative predictive value (NPV) for ruling out ischemia. Application of an ANN model could also avoid 73 % of invasive coronary angiograms while maintaining a 98 % NPV for detecting obstructive CAD. An online calculator for clinical use was created using these models. The ANN models improved risk stratification when compared to the TIMI score. Our calculator could also reduce downstream testing while maintaining an excellent NPV, though further study is needed before the calculator can be used clinically.

The impact of digital technology on health of populations affected by humanitarian crises: Recent innovations and current gaps.

Digital technology is increasingly used in humanitarian action and promises to improve the health and social well-being of populations affected by both acute and protracted crises. We set out to (1) review the current landscape of digital technologies used by humanitarian actors and affected populations, (2) examine their impact on health and well-being of affected populations, and (3) consider the opportunities for and challenges faced by users of these technologies. Through a systematic search of academic databases and reports, we identified 50 digital technologies used by humanitarian actors, and/or populations affected by crises. We organized them according to the stage of the humanitarian cycle that they were used in, and the health outcomes or determinants of health they affected. Digital technologies were found to facilitate communication, coordination, and collection and analysis of data, enabling timely responses in humanitarian contexts. A lack of evaluation of these technologies, a paternalistic approach to their development, and issues of privacy and equity constituted major challenges. We highlight the need to create a space for dialogue between technology designers and populations affected by humanitarian crises.

Cross-Modal Associations between Color and Haptics.

The objective of the present study was to explore cross-modal associations between color and tactile sensation while using haptically rendered virtual stimuli with substance properties of roughness/smoothness, hardness/softness, heaviness/lightness, elasticity/inelasticity, and adhesiveness/nonadhesiveness. The stimuli with the indicated properties were rendered with the aid of SensAble PHANTOM OMNI® haptic device. The experimental setup required the participants to use exploratory procedures typical to real object interaction, and select a color from the HSV color space that matched the experienced sensation. The findings of our investigation reveal systematic mapping between color characteristics and intensity of the haptic stimuli. Qualitatively different haptic sensations, however, produced relatively similar patterns of cross-modal associations.

Artificial neural network modeling using clinical and knowledge independent variables predicts salt intake reduction behavior.

BACKGROUND: High dietary salt intake is directly linked to hypertension and cardiovascular diseases (CVDs). Predicting behaviors regarding salt intake habits is vital to guide interventions and increase their effectiveness. We aim to compare the accuracy of an artificial neural network (ANN) based tool that predicts behavior from key knowledge questions along with clinical data in a high cardiovascular risk cohort relative to the least square models (LSM) method. METHODS: We collected knowledge, attitude and behavior data on 115 patients. A behavior score was calculated to classify patients' behavior towards reducing salt intake. Accuracy comparison between ANN and regression analysis was calculated using the bootstrap technique with 200 iterations. RESULTS: Starting from a 69-item questionnaire, a reduced model was developed and included eight knowledge items found to result in the highest accuracy of 62% CI (58-67%). The best prediction accuracy in the full and reduced models was attained by ANN at 66% and 62%, respectively, compared to full and reduced LSM at 40% and 34%, respectively. The average relative increase in accuracy over all in the full and reduced models is 82% and 102%, respectively. CONCLUSIONS: Using ANN modeling, we can predict salt reduction behaviors with 66% accuracy. The statistical model has been implemented in an online calculator and can be used in clinics to estimate the patient's behavior. This will help implementation in future research to further prove clinical utility of this tool to guide therapeutic salt reduction interventions in high cardiovascular risk individuals.