Smart Health最新文献

A robust health care system is crucial to reducing patient stress and contributing to economic growth. The future of the health care industry depends upon a reliable and efficient system to deal with the increasing number of patients. However, in today's healthcare system, patients face negative experiences as a result of long wait times. Now the pressing question is how to develop an effective healthcare system? To address this issue, this study uses Systems Modeling Language (SysML) coupled with a simulation approach to assess the performance of the healthcare system, identify the problem, and offer recommended alternatives. To elaborate, a systemic magic-grid methodology will be used to model and analyze the blood laboratory by using four pillars (structural, behavioral, requirement, and parametric) of SysML. To represent these pillars, a set of SysML diagrams will be used to visualize the layered system architecture, interactions, and activity between its different components. Furthermore, Discrete Event Simulation (DES) is utilized through Flexsim simulation software for the analysis of the parametric aspect of the system of interest. A blood laboratory within an outpatient clinic located at southern US State is considered a testing bed. The detailed architecture of the system of interest is studied, and required data are collected for modeling and simulation. The simulation results indicate that the combination of 50% Type I routes and 50% Type II routes resulted in the shortest wait times in the system of 22 min, the shortest wait times in the phlebotomist queue of 2 min, and the highest system throughput of 11369 patients per nine months. This article will provide a reference point for practitioners who want to apply the SysML approach to address health sector-related issues. More importantly, with this comprehensive approach, stakeholders of the blood laboratory system can utilize the hospital infrastructure in a more effective and optimized manner.

Ventricular Arrhythmia (VA) is a leading cause of sudden cardiac death (SCD), which kills an average of 180,000 to 350,000 people annually, accounting for 15%–20% of all deaths. Furthermore, fewer than 6% of those who experience sudden cardiac arrest outside the hospital survive, compared to 24% of those who experience SCD inside a hospital. To aid in earlier detection and improve outcomes for out-of-hospital cardiac events, an automated passive detection system for these events could be used. Such automated detection would allow users to raise their self-awareness of potential cardiac risks in life-threatening situations. Diagnosis and detection of heart dysfunctions at early stages can help to prevent complications of a patient’s condition.

In this work, we propose VANet and design framework for ECG-related application, a small-scale deep learning-based real-time inference solution for VA detection. VANet achieves milliseconds scale inference speed on various platforms, including desktop CPUs, mobile devices, micro-controllers, and devices with constrained computation resources. It only requires a minimum of 13 kb of storage space and 34 kb of available run-time, making it small enough to be integrated into portable devices such as smartwatches and other Internet of Things (IoT) medical monitoring devices. VANet can trigger an alarm whenever it is necessary to alert someone with cardiac dysfunction.

VANet leverages optimization techniques, such as residual connections, and architecture designs, such as transformers and RNNs, to maximize neural network performance and minimize computational and storage costs. Our architecture achieved a 96.89% accuracy using multiple different ECG collection devices.

Electroencephalogram (EEG)-based brain–computer interface (BCI) system is a promising tool for personalized rehabilitation post-stroke. Previous research has demonstrated the fundamental elements of these systems, including efficient classification, validated source localization, and visualization of EEG from stroke survivors. However, little attention has been given to developing a holistic framework to employ these elements in a human-in-the-loop personalized rehabilitation system. Without a holistic examination and development, we undervalue the context of personalized rehabilitation, ultimately hindering the agreement and acceptability in clinical practices and patients’ preferences. Therefore, this study proposed a holistic computational pipeline to employ classification, source localization, and visualization of EEG for personalized rehabilitation of stroke survivors. To this end, we designed an experiment focusing on upper limb movement in which a participant voluntarily performed the left hand’s shoulder, elbow, and wrist movements several times while simultaneously brain data were recorded with an EEG cap. Based on the recorded EEG data, we first developed feature engineering, importance analysis, and machine learning approaches with considerations of real-time implementations. EEG source localization was performed using the sLORETA method in Brainstorm to illustrate consistency for enabling agreement upon clinical conclusion about the brain areas that have been repeatedly activated when performing each of these movements. The experimental results demonstrated that decision trees of optimal selected EEG-channel features could achieve the best classification performance (95.63% Accuracy, 0.89 AUC). Furthermore, the EEG channels chosen by the decision trees showed consistency with the source localization.

Sleep apnea is a common and severe sleep-related respiratory disorder. Since the symptoms of sleep apnea are often ambiguous, it is difficult for a physician to decide whether to prescribe a clinical diagnosis, i.e., polysomnography (PSG), which results in a large percentage of undiagnosed and very late diagnosed cases. To reduce the time to diagnosis we investigate whether sleep monitoring data collected with a low-cost strain gauge respiration belt (called Flow) and a smartphone can be used to estimate with machine learning (ML) the severity of a patient’s sleep apnea. The Flow belt and the Type III sleep monitor Nox T3 were used together by 29 patients for unattended sleep monitoring at home, resulting each in 235 hours of sleep data. Through experimental analysis, we found that Convolutional Neural Networks are best suited to analyze the Flow data, because they are most robust against the frequently occurring baseline issues and exhibit the best performance with an accuracy of 0.7609, sensitivity of 0.7833, and specificity of 0.7217. These results can be achieved even if the classifier is trained only on high-quality data from the Nox T3. Thus, there are good chances that future ML experiments with data from other low-cost respiration belts can benefit from existing open PSG datasets without new extensive data collection. On a low-end smartphone, the classifier needs approximately one second to analyze the sleep data from one night. The results demonstrate the potential of low-cost strain gauge belts, smartphones, and ML to enable large parts of the population to perform sleep apnea pre-screening at home.

Emerging evidence has suggested that prenatal resting energy expenditure (REE) may be an important determinant of gestational weight gain. Advancements in technology such as the real-time, mobile indirect calorimetry device (Breezing™) have offered the novel opportunity to continuously assess prenatal REE while also potentially capturing fluctuations in REE. The purpose of this study was to examine feasibility and user acceptability of Breezing™ to assess weekly REE from 8 to 36 weeks gestation in pregnant women with overweight or obesity participating in the Healthy Mom Zone intervention study. Participants (N = 27) completed REE assessments once per week from 8 to 36 gestation using Breezing™. Feasibility of the device was calculated as compliance (# of weeks used/total # of weeks). User acceptability was measured by asking women to report on the device's enjoyability and perceived barriers. Median compliance was 68%. However, when weeks women experienced technical difficulties (11 of 702 total events) and the device was unavailable were removed (13 of 702 total events), median compliance increased to 71%. Over half (56%) of the women reported that the device was enjoyable or they had neutral feelings about it whereas the remaining 44% reported that it was not enjoyable. The most common barrier reported (44%) was the experience of technical issues. Study compliance data suggest the feasibility of using Breezing™ to assess prenatal REE is promising. However, acceptability data suggest future interventionists should develop transparent and informative protocols to address any barriers prior to implementing the device to increase use.

Nowadays, interpretable machine learning models are one of the most critical topics in the medical domain. The lack of interpretation leads to blind and unreliable models for clinicians, despite the fact that the aim is to support diagnosis through these tools. This problem has been increasing since the creation of large models such as those based on deep learning, which, despite providing good performance in prediction and classification tasks, are not transparent to human understanding. One of the increasingly prevalent clinical problems related to acoustic and linguistic disorders is Alzheimer’s disease (AD), where one important challenge is to provide speech markers that help in supporting, understanding, and facilitating the diagnosis and monitoring of the disease. It motivates this study which proposes a methodology focused on analyzing acoustic features in AD and at the same time providing interpretation from the results. The proposed approach consists of using decision tree-based methods together with neural networks (ForestNet) for analyzing the classification results. Only features that can give interpretation were considered. Unweighted average recalls of up to 79% were achieved for discriminating AD patients. Then, we looked at the relevant features that provided most of the information for assessing AD, which were those related to rhythm, voiced rates, duration, and phone rates. This confirms that this kind of approach can be suitable for the discrimination of AD while maintaining a good performance.

Wireless Body Area Networks (WBANs) provide wireless remote patient monitoring services where doctors get patients' health records without physically visiting them. In WBANs, biosensors are placed on the patient's body that sense and transmit physiological data to the paired medical personnel. Such medical setups are appropriate for COVID-19 patient monitoring, where the patient remains isolated for an extended period. Sometimes, human body parts impede the signals transmitted by biosensors to the coordinator and this type of occlusion lasts for a longer duration during sleeping human postures. In such circumstances, an intermediate biosensor forwards the signals of the occluded biosensor node. The forwarding of messages results in quick depletion of energy resources at the intermediate biosensor, affecting the overall WBAN services. To resolve this, first, we propose an adaptive Relay-Node Centric (RNC) relay-based communication protocol for WBANs, which reduces energy used in relaying and improves the stability period of the network. Second, we design a novel simulation model using an existing real-life experimental dataset to simulate a WBAN placed on the sleeping patient's body. We derive a Discrete Markov Chain (DTMC) model from real-life data and use human biomechanisms to simulate biosensors' connectivity status in four human sleeping positions. Lastly, we evaluate the performance of RNC against the existing cost-function-based and Analytical Hierarchical Process (AHP) based relay selection protocols. Results obtained on the real-life dataset and designed simulation model show that RNC outperforms the existing methods in terms of network stability period and packet success ratio.