Elsevier

Future Generation Computer Systems

Volume 98, September 2019, Pages 69-77
Future Generation Computer Systems

An IoMT cloud-based real time sleep apnea detection scheme by using the SpO2 estimation supported by heart rate variability

https://doi.org/10.1016/j.future.2018.12.001Get rights and content

Highlights

  • The employee of IoMT to detect the Apnea disease.

  • Cloud based remote control and interpretation.

  • The wearable monitoring sensor for continuous monitoring.

  • The combination of two bio-signal (SpO2 and HRV) to improve the detection accuracy.

  • Efficient alarm for instant disease.

Abstract

Obstructive sleep apnea refers to a highly rampant sleep-related breathing disorder. The gold standard examination for diagnosis is polysomnography. Even though it provides highly accurate results, this multi-parametric test is time consuming and expensive. It also does not align with the new trend in health care, where focus is shifted to wellness and prevention. One possible way to address this problem is home health care, through the use of minimal invasive devices, higher accessibility, and provision of low cost diagnosis. To manage this, an automated and portable sleep apnea detector was formulated and assessed. The device utilizes one SpO2 sensor for estimating the heart rate and the oxygen blood level as well. The basis of the proposed analysis method is the connection between heart rate variability and oxygen saturation with d apnea events. The measured signals were then transferred to a cloud-based system architecture to diagnose and warn the remote patients. This solution was used to process the data and display it on both the mobile phone and personal computer. Testing of the proposed algorithms was done using the St. Vincents University Hospital/University College Dublin sleep apnea database. Apart from this database, the researchers utilized data gathered from 10 apnea patient volunteers. The performance of the proposed scheme algorithm achieved an average accuracy, specificity, and sensitivity of 98.54, 98.95%, and 97.05%, respectively.

Introduction

One can define the sleep apnea syndrome (SAS) as a momentary shutting of the upper parts of the human airway during sleep. This prevents air from incoming to the lungs, which could lead to the complete breathing cessation in adults for more than 10 s [1]. Typically, this is accompanied by lower blood oxygen levels saturation and results to sleep arousal in order to breathe. Furthermore, it has been hypothesized that repetitive unhelpful events during sleep could cause irregular hypoxia, which activates an oxidative stress response and oxygen free radicals.

SAS is considered a major healthcare problem that affects at least 2% of women and 4% of men [2]. Furthermore, about 4% of the general population experience SAS to some extent. There is an estimate that less than 25% of people who suffer from SAS are aware that they have the condition [3]. In the United States, these undiagnosed sufferers result into 70 billion dollars in loss and 11.1 billion in damages, and lead to 980 deaths annually [4].

Generally, evaluating sleep quality and examining the presence of OSA is vital in improving the health of the general population and reducing healthcare costs and mortality. As a matter of fact, this illness leads to asphyxia, hypoxemia, and awakenings. It also has instantaneous values such as high blood pressure and increased heart rate. It also causes long-term symptoms that influence quality of life such as poor concentration, extreme fatigue, a compromised immune system, cardio/cerebrovascular problems, and slower reaction times [5].

Certain sleep disorders are serious enough to affect normal mental, physical, and emotional functioning. People who have complaints about sleepiness or daytime fatigue could be suffering from interrupted sleep. This leads to daytime sleepiness and failure to concentrate, which could result in accidents [6]. Therefore, it directly affects a patient’s quality of life. Recently, sleep disorders have become one of the major focuses of public safety in America, Europe, and Japan.

One can typically classify sleep apnea into three kinds: central sleep apnea (CSA), obstructive sleep apnea (OSA), and mixed sleep apnea (MIX). OSA is considered the more popular form of apnea. It results from a blockage of the airway. Furthermore, it is generally associated with reduced blood oxygen saturation. On the other hand, in CSA, there is no blockage of the airway. However, the brain fails to send a signal to the muscles to breathe as a result of instabilities in the respiratory control center. On the other hand, MIX happens as a result of the transition between brief intervals of CSA and long periods of OSA [7].

In terms of SAS diagnosis, one crucial and standard method is nocturnal polysomnography (PSG). Included in this diagnosis is the monitoring of the breath airflow [8], breathing events [9], snore [10], oxygen saturation (SpO2) [11], electrooculography (EOG) [12], electroencephalography (EEG) [13], and electrocardiography (ECG) [14]. Nonetheless, this examination contains several disadvantages [15]. First, it is fairly complicated because it is made up of the recording of electroencephalogram, oronasal airflow, electrooculogram, chest wall as well as abdominal wall movement, ElectroCardioGram (ECG), and oxygen saturation measurements. Second, another drawback associated with this is that when a PSG test is conducted, a patient needs to stay in the same location during the night due to the large amounts of wires, tubes, leads, and so on. The third issue with the present OSA tests is that they limit the patients to a hospital setting for at least one night. They can also generate stresses that may affect the actual OSA form and the outcomes of this observing activity. Lastly, PSG is expensive. For instance, the coast-to-coast average price of PSG medicare rates in Massachusetts is around 1018 $ [16].

Thus, there is a need to have new simplified techniques and methods for screening and diagnosis. However, because of the complexity of handling with SAS symptoms, as proven by the monitoring of multi-parameter offered by the PSG. Moreover, it is ideal for state of the art approaches to be able to summarize all the collected information with the goal of obtaining the most relevant data and lowering the complexity of the model before the process of classification begins. Therefore, SAS screening and diagnosis often includes three major methodology stages: feature extraction, pattern classification, and selection of features. First, the main goal is to accomplish a smaller set of features, usually extracted through algorithms designed from the observed data. Second, one should prioritize these features and offer an adequate selection. This is expressive since majority of classification and diagnosis algorithms fail to achieve high accuracy when managing a large amount of weakly relevant and/or redundant characteristics. Third, one should wisely choose a classification method that aims to provide reliable, reasonable, and consistent decisions.

SpO2 refers to the percentage (%) of haemoglobin within the blood which is saturated with oxygen and noted by a pulse oximeter. Various commonly utilized SpO2 features include the accumulative time (TSA) that is go below a specific saturation level [17], the oxygen desaturation index (ODI, the amount of oxyhemoglobin desaturation that goes below a specific threshold) [18], and the saturation variability index (Delta index) [19]. In terms of the spectral-domain characteristics, the studies in [19] presented a periodogram of the SpO2 signal and chose four associated with the 30–70 s period for the purpose of detection. However, all the previously methods depend on the whole overnight SpO2 records, which result into delayed offline diagnosis and analysis.

Numerous studies have been performed for OSAS screening as they tried to lower PSG complexity and cost. Various methods have been suggested; with oximetry-based screening considered to be one of the most commonly suggested method for both the paediatric and adult populations. Promising technologies such as big data and cloud computing and the Internet of things (IoT) have important potential for applying home-based solutions to support and monitor the sleep apnea. The Internet of Medical Things, or the (IoMT) paradigm, which is the IoT and health customized version of IoT, has become a key enabler for empowering elderly people regarding wellness, independent aging, and disease management [20]. IoT also allows for the omnipresent interconnection of various smart objects with sensing, networking, acting, and processing capabilities via the Internet by offering the ability to impart information [21]. This interconnection has the potential to produce vast quantities of data that need scalable computing infrastructure that can be used to conduct efficient real-time analysis and processing. In this context, cloud computing may be helpful since it can offer scalable and on-demand storage that allow processing services and big data analysis.

In this study, a new sleep apnea scheme was proposed. This method adopts the IoMT techniques for real time detection which enable the true alarm and give a prediction alarm. Besides, the combination of using the HRV signal and oxygen level in the blood contributes to improve the quality of detection and reduce the false alarm cases.

Section snippets

Scheme architecture

The proposed scheme is illustrated in Fig. 1. It is split into three key components: the SpO2 sensor, the gateway, and the cloud. The first part contains the wearable SpO2 device which provides the required signal to extract both HRV and SpO2 as well. These features then are subjected to pre-processing analysis to isolate unwanted noise and improve the collected data quality. The second part is the system gateway where it is an important section that represents the pathway for data transfers.

Results and discussion

Different combinations between HRV and SpO2 features are tested to validate the main concept and gain the best and most suitable features. Table 1 illustrate the possible combinations which named “Case”. Besides, the effect of training samples is considered in this study where it can play an important factor. The training number is effective in both accuracy and system resources which should be compromised. For this reason the training — testing setting is set to 70%–30% in this study. Fig. 5

Conclusions

We put forward an automatic diagnostic system, based only on oximetry analysis, which has numerous possible configurations. Also, assessment of the classification algorithm has been done. Multiple possible cases were considered to test the developed architecture, which enabled a multitude of applications. In longer time periods, it was easy to identify the features oscillations. A good discriminant capacity was demonstrated with the 1-min variance and procurement of relevant information to

Li Haoyu, he is a master of Electrical Engineering student at Fujian university of Technology. His research interests include power transformer fault analysis.

References (35)

  • ChenL. et al.

    An automatic screening approach for obstructive sleep apnea diagnosis based on single-lead electrocardiogram

    IEEE Trans. Autom. Sci. Eng.

    (2015)
  • ChokrovertyS. et al.

    Sleep, breathing, and neurologic disorders

  • ForceP.A.P.T.T. et al.

    Clinical guidelines for the manual titration of positive airway pressure in patients with obstructive sleep apnea

    J. Clin. Sleep Med.

    (2008)
  • GüderF. et al.

    Paper-based electrical respiration sensor

    Angew. Chem. Int. Ed.

    (2016)
  • DafnaE. et al.

    Automatic detection of whole night snoring events using non-contact microphone

    PLoS One

    (2013)
  • BabbarA. et al.

    Use of non-invasive mechanical ventilation to prevent hypoxia during bronchoscopy in patients with sleep apnea or at high risk of sleep apnea

  • KapurV.K. et al.

    Clinical practice guideline for diagnostic testing for adult obstructive sleep apnea: an American Academy of Sleep Medicine clinical practice guideline

    J. Clin. Sleep Med.

    (2017)
  • Cited by (71)

    • An overview of the Internet of medical things (IoMT): Applications, benefits, and challenges

      2023, Security and Privacy Issues in Internet of Medical Things
    • Toward sleep apnea detection with lightweight multi-scaled fusion network

      2022, Knowledge-Based Systems
      Citation Excerpt :

      A wearable device is worn or attached to the surface of the skin. Thus, wearable devices also can be used for SA detection, based on collected ECG signals [25], EEG signals [42], accelerator-based signals [43], and photoplethysmogram (PPG) signals [44]. EEG signals, with many electrodes placed along the scalp, are quite burdensome for users to monitor their sleep conditions at home.

    • Advances on networked ehealth information access and sharing: Status, challenges and prospects

      2022, Computer Networks
      Citation Excerpt :

      The clinical-care IoMT devices are highly personalized and designed to address specific chronic diseases of the individual patient. Examples include sleep apnea detectors [4] and body sensor network (BSN) based health monitors [5]. The massive adoption of IoT in the healthcare industry in creating a smart health environment for monitoring the medical conditions of chronic patients has been making a positive impact on the development of the economy and social welfare [6].

    View all citing articles on Scopus

    Li Haoyu, he is a master of Electrical Engineering student at Fujian university of Technology. His research interests include power transformer fault analysis.

    Li Jianxing, Master of Control Theory and Control Engineering,Graduated from the Central South University . He is currently a professor of School of Information Science and Engineering, Fujian University of Technology. His research interest is mainly in the area of Process Control and PLC Application Technology.

    Arunkumar N Is associated with SASTRA University and is active in research. He has published several papers in engineering streams in various top journals. He has been guiding several students from various countries in their research works.

    Ahmed Faeq Hussein Received his B.Sc. degree in Electrical Engineering from Al-Mustansiriyah University, Iraq in 1998, the M.Sc. degree in Computer Engineering from University of Technology, Iraq in 2004 and the PhD. Degree in Computer and Embedded System Engineering from Universiti Putra Malaysia in 2018. He was senior engineer at Medical Department, Ministry of Health, Iraq until 2009, and lecturer at Bio-Medical Engineering Department, Al-Nahrain University since 2009. His research include Bio-medical signal processing, Low Energy Bluetooth Communication and Cloud based application.

    Mustafa Musa Jaber is lecturer at Dijlah University College, He did his bachelor in field of software engineering from Alrafiadain university college and then he completed his master and PhD in field of IT from Malaysia, his interest is human interaction factors, telemedicine, and data warehouse applications.

    View full text