Global health which denotes equitable access to healthcare, particularly in remote-rural-developing regions, is characterized by unique challenges of affordability, accessibility, and availability for which one of the most promising technological interventions that is emerging is the Internet of Things (IoT) based remote health monitoring. We present an IoT based smart edge system for remote health monitoring, in which wearable vital sensors transmit data into two novel software engines, namely Rapid Active Summarization for effective PROgnosis (RASPRO) and Criticality Measure Index (CMI) alerts, both of which we have implemented in the IoT smart edge. RASPRO transforms voluminous sensor data into clinically meaningful summaries called Personalized Health Motifs (PHMs). The CMI alerts engine computes an aggregate criticality score. Our IoT smart edge employs a risk-stratified protocol consisting of rapid guaranteed push of alerts & PHMs directly to the physicians, and best effort pull of detailed data-on-demand (DD-on-D) through the cloud. We have carried out both clinical validation and performance evaluation of our smart edge system. The clinical validation on 183 patients demonstrated that the IoT smart edge is highly effective in remote monitoring, advance warning and detection of cardiac conditions, as quantified by three measures, precision (0.87), recall (0.83), and F1-score (0.85). Furthermore, performance evaluation showed significant reductions in the bandwidth (98%) and energy (90%), thereby making it suitable for emerging narrow-band IoT networks. In the deployment of our system in the cardiology institute of our University hospital, we observed that our IoT smart edge helped to increase the availability of physicians by 59%. Hence, our IoT smart edge system is a significant step towards addressing the requirements for global health.

A large number of obstructive sleep apnea (OSA) cases are under-diagnosed due unavailability, inconvenience or expense of sleep labs. Hence, an automated detection by applying computational techniques to multivariate signals has already become a well-researched subject. However, the best-known techniques that use various features have not achieved the gold standard of polysomnography (PSG) tests. In this paper, we substantiate the medical conjecture that OSA directly impacts body parameters such as Instantaneous Heart Rate (IHR) and blood oxygen saturation (SpO2). We then use a deep learning technique called LSTM-RNN (long short-term memory recurrent neural networks) to experimentally prove that OSA severity detection can be solely based on either IHR or SpO2 signals, which can be easily, obtained using off-the-shelf non-intrusive wearable single sensors. The results obtained from LSTM-RNN model shows an area under curve (AUC) of 0.98 associated with very high accuracy on a dataset of more than 16,000 apnea non-apnea minutes. These results have encouraged our collaborating doctors to further come up with a diagnostic protocol that is based on LSTM-RNN, SpO2, and IHR, thereby increasing the chances of larger adoption among medical community.

Automated sleep apnea detection and severity identification has largely focused on multivariate sensor data in the past two decades. Clinically too, sleep apnea is identified using a combination of markers including blood oxygen saturation, respiration rate etc. More recently, scientists have begun to investigate the use of instantaneous heart rates for detection and severity measurement of sleep apnea. However, the best-known techniques that use heart rate and its derivatives have been able to achieve less than 85% accuracy in classifying minute-to-minute apnea data. In our research reported in this paper, we apply a deep learning technique called LSTM-RNN (long short-term memory recurrent neural network) for identification of sleep apnea and its severity based only on instantaneous heart rates. We have tested this model on multiple sleep apnea datasets and obtained perfect accuracy. Furthermore, we have also tested its robustness on an arrhythmia dataset (that is highly probable in mimicking sleep apnea heart rate variability) and found that the model is highly accurate in distinguishing between the two.

Objective: To carry out exhaustive data-driven computations for the performance of noninvasive vital signs heart rate (HR), respiratory rate (RR), peripheral oxygen saturation (SpO2), and temperature (Temp), considered both independently and in all possible combinations, for early detection of sepsis. Materials and methods: By extracting features interpretable by clinicians, we applied Gradient Boosted Decision Tree machine learning on a dataset of 2630 patients to build 240 models. Validation was performed on a geographically distinct dataset. Relative to onset, predictions were clocked as per 16 pairs of monitoring intervals and prediction times, and the outcomes were ranked. Results: The combination of HR and Temp was found to be a minimal feature set yielding maximal predictability with area under receiver operating curve 0.94, sensitivity of 0.85, and specificity of 0.90. Whereas HR and RR each directly enhance prediction, the effects of SpO2 and Temp are significant only when combined with HR or RR. In benchmarking relative to standard methods Systemic Inflammatory Response Syndrome (SIRS), National Early Warning Score (NEWS), and quick-Sequential Organ Failure Assessment (qSOFA), Vital-SEP outperformed all 3 of them. Conclusion: It can be concluded that using intensive care unit data even 2 vital signs are adequate to predict sepsis upto 6 h in advance with promising accuracy comparable to standard scoring methods and other sepsis predictive tools reported in literature. Vital-SEP can be used for fast-track prediction especially in limited resource hospital settings where laboratory based hematologic or biochemical assays may be unavailable, inaccurate, or entail clinically inordinate delays. A prospective study is essential to determine the clinical impact of the proposed sepsis prediction model and evaluate other outcomes such as mortality and duration of hospital stay.

Remote health monitoring and delivery through mobile devices and wireless networks offers unique challenges related to performance, reliability, data size, power management, and analytical complexity. We present a multi-layered architecture that matches communication performance to medical importance of data being monitored. The priority of vital data and the context of sensing are used to select the communication medium and the power management policies. Further smartness is introduced into data summarization by employing a severity level quantizer, followed by a consensus abnormality motif discovery and an alert mechanism that prioritizes doctors' consultative time. We also present our successful implementation of the above multi-layered architecture in a system developed to remotely monitor cardiac patients.