Towards Wearables and Medical IoT Integration, Interoperability and real-time Analytics
- Posted by Admin
- Posted in Research and Innovation
The Internet of Things (IoT) is a growing trend in our daily lives, with estimates of 30 billion “things” connected to the internet by 2025. In recent years, the healthcare industry has witnessed a growing interest in creating practical medical care arrangements that help medical clinics and providers find and encourage the patient. In order to provide proper and low latency healthcare, new applications are required to handle the increasing patient’s needs. On the individual and clinical level, traditional healthcare has many issues to overcome. Initially, traditional healthcare services struggle to meet rising population demands for quality care. In traditional healthcare systems, a lack of medical staff leads to delayed diagnosis and treatment. With emerging wireless communication, sensor, and 5G technologies, home care and remote healthcare applications for low, medium, and high-risk patients can be easily implemented. In this context, smart healthcare is used to simplify and improve people’s health care methods.
IoT brings adaptability and comfort in conveying in various conditions for observing and communication. A patient’s body or a hospital or home environment can have embedded sensors and IoT devices that monitor ECG signals, blood pressure, temperature, oxygen saturation, and movement. Smart healthcare uses various data related to an individual’s health status to diagnose diseases. This allows for remote monitoring and continuous follow-up of patients’ medical issues, as well as long-term review of their health records, lowering clinical costs and expanding innovation for patient-driven care rather than medical clinic-driven treatment.
Wearable technology has come a long way since its inception, and now making waves in almost every domain of our digitally connected life. Figure 1 illustrates the growth and adaptation trend of various wearable technologies.
One of the significant milestones in IoT-based healthcare systems is the integration of the wearable technologies for real-time patient monitoring. With wearables, we mean small electronic and mobile devices, or computers that can be worn on the body, or even invasive versions like micro-chips or smart tattoos. Unlike today’s smartphones and tablets, wearables can provide various monitoring and scanning functions, including biofeedback and other sensory physiological functions like biometry.
The benefits of wearables for E-Health innovation are widely recognized. Personal wearable devices help patients and caregivers better evaluate and manage health and wellness in many contexts. However, some of the practices governing data collection and use for E-Health monitoring have stifled innovation. For example, traditional blood pressure cuffs were used to collect data for years. There was no true, technically validated standard for data capture from cuff-less metres. Earlier cuff-less devices required precise calibration. A new standard would help disparate devices capture, calibrate, measure, assess, transfer, and archive blood pressure data.
On the other hand, the wearable market is expected to grow exponentially in the coming years. The market is expected to grow by over 20% annually in the next five years, reaching over 40 billion EUR by 2028. Wearable shipments are expected to increase from 113.2 million in 2017 to 222.3 million in 2021, with a market value of $70 billion. The COVID-19 outbreak also had a huge impact on the evolution of wearable devices driven by various crowd-sensing and contact-tracing platforms.
However, it is very crucial to understand the applicability and viability of smart wearable integration in healthcare setting. At present, most of the wearable devices designed for healthcare applications integrates a range of on-board sensors for patient diagnostics. Some of these sensors include heart-rate, blood pressure, ECG, breathing monitoring, temperature, and motion detection sensors. The combined measurements from these sensors enables a primitive real-time health profile for patient that can be monitored by health-care professionals. Furthermore, the availability of lightweight, and efficient Artificial Intelligence (AI) and Machine Learning (ML) algorithms allow autonomous anomaly detection without the need of supervised assistance from a health-care professional.
Although most of the wearable devices integrate open-source sensors, the measurements, algorithms and applications are proprietary in nature. This limits the extent and standardization of health-care services implementation, which differs from platform to platform. One such example is activity recognition using wearable devices. Activities such as walk, running, and physical workouts are detected by on-board sensors (including GPS, gyro-sensors, and Inertial Measurement Unit IMU etc), the actual implementations varies by vendors, thus producing dissimilarities in observations. Typically, the slight variations in such observations can be neglected, however, in health-care settings data precision and integrity is of utmost importance.
Recently, the world faced the impact of COVID-19 virus with the death toll rising up to roughly 4 Million worldwide. World Health Organization (WHO) reported fever, cough, and variations in breathing among the most common symptoms found in every Covid-19 positive patient. It was also reported that early disease detection may help the patient to receive medical assistance that could help prevent the disease spread, and therefore, providing high survival chances in these situations.
This motivated us to investigate the common COVID-19 symptoms for anomaly detection that could potentially help in early diagnosis and thus prevent life-threatening situations. In this context, we decided to choose commercially available off-the-shelf (COTS) components to design an open-source framework that is capable of identifying anomalous observations, leading to health warnings in general, and respiratory tract related alerts, in particular. In our proposed settings, we utilized low-cost, low-power microcontrollers with range of sensors (including heart-rate, blood pressure, and temperature and IMU sensors) to record real-time personal activities and physiological features.