The Internet of Medical Things Enabling Technologies and Emerging Applications
The Internet of Medical Things Enabling Technologies and Emerging Applications
Edited by
Subhendu Kumar Pani, Priyadarsan Patra,
Gianluigi Ferrari, Radoslava Kraleva and Xinheng Wang
This publication is copyright under the Berne Convention and the Universal Copyright
Convention. All rights reserved. Apart from any fair dealing for the purposes of research
or private study, or criticism or review, as permitted under the Copyright, Designs and
Patents Act 1988, this publication may be reproduced, stored or transmitted, in any
form or by any means, only with the prior permission in writing of the publishers, or in
the case of reprographic reproduction in accordance with the terms of licences issued
by the Copyright Licensing Agency. Enquiries concerning reproduction outside those
terms should be sent to the publisher at the undermentioned address:
While the authors and publisher believe that the information and guidance given in this
work are correct, all parties must rely upon their own skill and judgement when making
use of them. Neither the authors nor publisher assumes any liability to anyone for any
loss or damage caused by any error or omission in the work, whether such an error or
omission is the result of negligence or any other cause. Any and all such liability is
disclaimed.
The moral rights of the authors to be identified as authors of this work have been
asserted by them in accordance with the Copyright, Designs and Patents Act 1988.
17 Big data analytics and data mining for healthcare and smart city
applications 331
Sohit Kummar and Bharat Bhushan
17.1 Introduction 331
17.2 Theoretical background of smart cities 333
17.2.1 Smart people 334
17.2.2 Smart economy 334
17.2.3 Smart governance 335
17.2.4 Smart mobility 335
17.2.5 Smart environment 335
17.2.6 Smart living 336
17.3 Computational infrastructures for smart cities big data analytics 336
17.3.1 Cloud computing 338
17.3.2 Fog computing 338
17.3.3 Edge computing 339
17.3.4 Big data based on machine learning 339
17.4 Mining methods for big data 340
17.4.1 Classification 340
17.4.2 Clustering 341
17.4.3 Frequent pattern mining 342
17.4.4 Another mining method 342
17.5 Advances in healthcare sector 342
17.5.1 Big data for healthcare 343
17.5.2 Data mining for healthcare 345
17.6 Conclusion 345
References 346
Index 355
About the editors
The Internet of medical things (IoMT) is rapidly changing the healthcare sector. IoMT
paired with enabling technologies like artificial intelligence, cloud computing, wireless
sensor networks can effectively monitor people’s health continuously. In this chapter,
the author provides a systematic review of the architectures of IoMT, applications, and
trends in IoMT. IoMT applications to counter the COVID-19 pandemic are also dis-
cussed. IoMT and enabling technologies will improve the quality of human lives. This
chapter addresses a few challenges while adopting IoMT in healthcare.
1.1 Introduction
In Internet of Things (IoT), all objects are considered smart objects that are paired
with each other. Sensors and actuators blend with the environment and commu-
nication is shared across the platform. As per Atzori [1] IoT can be realized in three
paradigms: (1) semantic oriented, (2) things oriented and (3) Internet oriented.
As per [2], IoT has been recognized as one of the emerging technologies that
transform the world. IoT has a complex architecture, where various components
interact with each other to find the solutions. IoT comprises cyber-physical systems
to facilitate data-driven decision process.
By integrating various sensors and data analytics, IoT offers various solutions
that can be used in smart grids, smart towns, smart homes, etc. There are three
components in IoT that enables seamless applications [3]. The three components
are (1) hardware that consists of sensors and actuators with embedded hardware, (2)
middleware used for computing and storage and (3) visualization tools.
IoT is enabled with various technologies like (1) wireless sensors networks
(WSN), (2) radio-frequency identification (RFID), (3) communication stack for
WSN, (4) hardware for WSN, (5) middleware and (6) methods for secure data
integration [3].
1
Vardhaman College of Engineering, Hyderabad, India
2 The Internet of medical things
Cloud server
IoT edge devices
IoT sensors
Data transformation
Bulk device IoT gateway User
Stream processing
provisioning framework interface
Central management User
etc., system
IoT
Application Network devise
domain domain domain
Satellite
Dish Gateway
Data
Services
Person
Person
User
Smart healthcare one of the important aspects of human life. The use of emerging
technologies in healthcare will help one to reach the health-related services to all
the stakeholders. IoMT along with other emerging technologies is playing a vital
role in smart healthcare. Various applications of IoMT in healthcare are (1)
4 The Internet of medical things
management of chronic diseases, (2) telehealth, (3) lifestyle assessment, (4) remote
intervention, (5) drug management, (6) medical nursing system, (7) rehabilitation
system and (8) medical Bot.
This subsection gives a brief review of research related to applications of IoMT
in healthcare.
Muzammal et al. [12] presented multisensor data fusion framework. IoMT is
used for cuff-less BP measurement. In [13] authors presented a model based on
backward shortcut connections. An IoMT model is designed for emotion recognition.
Internet of medical things (IoMT) 5
LTE
IoT
Gateway
Communication
Sensors technologies
Data generated from various sensors was fused and the output is fed to seven different
classifiers.
In [19] authors presented a method for emotion recognition. Fused and non-
fused physiological signal data sets were used for the evaluation of the proposed
method. A feed-forward neural network classifier is used to train the data set.
Authors in [20] modeled a recognition system for human action. The proposed
method is tested on UTD-MHAD data set.
A Gaussian filter was used to decompose the images. Baloch et al. [21] pre-
sented a method for IoT healthcare applications. The proposed method consists of
three phases to resolve issues and to maintain the efficiency. Applications of IoMT
in healthcare have been shown in Table 1.4. The illustration of IoMT applications is
shown in Figure 1.5 [22].
Internet of medical things (IoMT) 7
Internet of
medical things
AI-enabled IoMT has been deployed at a Taiwan’s hospital to scan the people
coming to hospital, detect the temperature and identify whether the people wearing
the face mask or not [26]. Table 1.5 shows the various technologies used for
COVID-19. Table 1.5 shows applications of IoMT during COVID-19.
Research that is mainly focused to address the challenges faced by COVID-19 are
1. measures to enforce social distancing using the technology,
2. remote patient monitoring of the patients using IoMT,
3. IoMT-based methods for infection detection,
4. IoMT-enabled thermal screening to avoid the spread of disease.
Advance technologies to tackle the pandemic are shown in Figure 1.6 [27].
Cognitive IoT is an emerging technology that is used for an efficient utilization
of a scarce spectrum [28]. This technology is well suited to track the persons affected
with COVID-19. Cognitive IoMT (CIoMT) is a special case of IoT. IoMT has been
used in major areas to handle COVID-19 pandemic. CIoMT has been used for
1. real-time tracking of patients,
2. monitoring COVID-19-affected patients remotely,
IoT in 5G
Cloud
computing and Advance AI and deep
big data Technologies learning
analytics
Blockchain
and Industry
4.0
Figure 1.6 Advanced technologies that are used to tackle COVID-19 [27]
Internet of medical things (IoMT) 9
Real-time
tracking
Remote
Prevention
monitoring
and
of the
control
patient
CIoMT
Reducing
workload Rapid
of medical diagnosis
industry
Contact
Screening
tracking
and
and
surveillance
cluttering
Smart
healthcare
Wearable Patient
devices tracking
IoMT facilities
for
orthopedic
patients in
COVID-19
Internet
Exigency
-based
warning
services
Remote
health
monitoring
Attacks on IoT classified into two types: (1) data based and (2) protocol based.
Data-based attacks include threats related to original data packets, whereas
protocol-based attacks exploit protocol-based structure. Protocol-based attacks are
again classified into two types: (1) communication protocol based and (2) network
protocol based.
Attacks will also be classified as active and passive. Category of various
attacks is shown in Figure 1.9.
IoT attacks can also be classified as shown in Figure 1.10.
Rapid evolution of IoMT also raised security problems. IoMT security
requirements and methods include (1) block chain, (2) access control methods,
(3) IDS methods, (4) authentication schemes and (5) key management methods.
Cyberattacks that happen to IoMT are listed as follows: (1) injection attacks,
(2) denial of service attack, (3) device safety and (4) data leakage attacks.
Attack surface for denial service is mostly on databases and cloud services.
Hardware and middleware are the target for device safety attacks. Injection
attacks basically target the databases. Data leakage attacks will occur to
information and network.
Internet of medical things (IoMT) 11
Attack type
Masquerade Port
DOS attack Message replay Traffic sniffing
attack scanning
Attacks
8. Accuracy and precision are also two critical design factors for IoMT.
Inaccurate results could be harmful to the patients.
Standard protocols, 5G-enabled emerging technologies like edge computing,
explainable AI, can be integrated with IoMT for an effective use of IoMT in
healthcare sector.
1.5 Conclusion
Smart healthcare is a current and well-researched area. IoMT, AI, cloud computing,
and wireless sensor networks are all part of the smart healthcare. The main outcome
of this chapter is to highlight the recent advances and applications of IoMT in the
healthcare sector. The potential use of IoMT in handling the COVID-19 pandemic
is discussed. Security attacks and various mitigation techniques were also addres-
sed. We hope that this survey will benefit the research community in understanding
the role of IoMT in smart healthcare and to adopt this technology well in the
healthcare sector.
References
[1] L. Atzori, A. Iera and G. Morabito, The Internet of Things: A survey.
Comput. Networks 2010; 54(15): 2787–2805.
[2] J. Fenn, Gartner’s hype cycle special report for 2005, 2005. http://www.
gartner. com/resources/130100/130115/gartners_hype_c. pdf.
[3] J. Gubbi, R. Buyya, S. Marusic and M. Palaniswami, Internet of Things
(IoT): A vision, architectural elements, and future directions. Future Gener.
Comput. Syst. 2013; 29(7): 1645–1660.
[4] W. Weng, P. Chen, S. He, X. Sun and H. Peng, Smart electronic textiles.
Angew. Chem. Int. Ed. 2016; 55: 6140–6169.
[5] A.K. Raeespour and A.M. Patel, Design and evaluation of a virtual private
network architecture for collaborating specialist users. Asia Pac. J. Inf.
Technol. Multimed. 2016; 5: 13–50.
[6] R.A. Stackowiak, IoT solutions overview. In Azure Internet of Things
Revealed; Springer: Berkeley, CA, USA, 2019; pp. 29–54.
[7] R. Hassan, F. Qamar, M.K. Hasan, A.H.M. Aman and A.S. Ahmed, Internet
of Things and its applications: A comprehensive survey. Symmetry 2020; 12
(10): 1674. https://doi.org/10.3390/sym12101674.
[8] G.J. Joyia, R.M. Liaqat, A. Farooq and S. Rehman, Internet of Medical
Things (IoMT): Applications, benefits and future challenges in Healthcare
domain. J. Commun. 2017; 12(4): 240–247.
[9] J.T. John and S.J. Ramson, Energy-aware duty cycle scheduling for efficient
data collection in wireless sensor networks. IJARCET 2013; 2.
[10] S. Ramson and D.J. Moni, A case study on different wireless networking tech-
nologies for remote health care. Intell. Decis. Technol. 2016; 10(4): 353–364.
Internet of medical things (IoMT) 13
[11] F. Al-Turjman, M.H. Nawaz and U.D. Ulusar, Intelligence in the Internet of
Medical Things era: A systematic review of current and future trends.
Comput. Commun. 2020; 150: 644–660.
[12] M. Muzammal, R. Talat, A.H. Sodhro and S. Pirbhulal, A multi-sensor data
fusion enabled ensemble approach form Medical data from body sensor
networks. Inf. Fusion 2020; 53: 155–164.
[13] T. Van Steenkiste, D. Deschrijver and T. Dhaene, “Sensor fusion using
backward shortcut connections for sleep apnea detection in multi-modal
data,” 2019, arXiv:1912.06879. [Online]. Available: http://arxiv.org/abs/
1912.06879.
[14] K. Lin, Y. Li, J. Sun, D. Zhou and Q. Zhang, Multi-sensor fusion for body
sensor network in medical human–robot interaction scenario. Inf. Fusion
2020; 57: 15–26.
[15] F. Al-Shargie, Fusion of fNIRS and EEG signals: Mental stress study.
engrXiv 2019; 2019: 1–5. doi: 10.31224/osf.io/kaqew.
[16] I. Cabria and I. Gondra, MRI segmentation fusion for brain tumor detection.
Inf. Fusion 2017; 36: 1–9.
[17] D. Fabiano and S. Canavan, Emotion recognition using fused physiological
signals. In Proc. 8th Int. Conf. Affect. Comput. Intell. Interact. (ACII),
Cambridge, UK, 2019; pp. 42–48.
[18] W. Zhang, J. Yang, H. Su, M. Kumar and Y. Mao, Medical data fusion
algorithm based on Internet of Things. Pers. Ubiquitous Comput. 2018; 22
(5–6): 895–902.
[19] J. Du, W. Li and H. Tan, Intrinsic image decomposition-based grey and
pseudo-color medical image fusion. IEEE Access 2019; 7: 56443–56456.
[20] C. Chen, R. Jafari and N. Kehtarnavaz, A real-time human action recogni-
tion system using depth and inertial sensor fusion. IEEE Sens. J. 2016; 16(3):
773–781.
[21] Z. Baloch, F.K. Shaikh and M.A. Unar, A context-aware data fusion
approach for health-IoT. Int. J. Inf. Technol. 2018; 10(3): 241–245.
[22] I. Ud Din, A. Almogren, M. Guizani and M. Zuair, A decade of Internet of
Things: Analysis in the light of healthcare applications. IEEE Access 2019;
7: 89967–89979. doi: 10.1109/ACCESS.2019.2927082.
[23] D.S.W. Ting, L. Carin, V. Dzau and T.Y. Wong, Digital technology and
COVID-19. Nat. Med. 2020; 26: 7.
[24] Y. Song, J. Jiang, X. Wang, D. Yang and C. Bai, Prospect and application of
Internet of Things technology for prevention of SARIs. Clin. eHealth 2020;
3: 1–4. doi: 10.1016/j.ceh.2020.02.001.
[25] D. Koh. 2020. Temp Pal Smart Thermometer Helps Reduce COVID-
19 Spread in Hospitals. https://www.mobihealthnews.com/news/asia-paci-
fic/temp-pal-smart-thermometer-helps-reduce-covid-19-spread-hospitals.
[26] Microsoft. 2020. Trying to Shield Hospital Staff and Patients from COVID-19
with Help from AI, Cloud, and Intelligent Edge – Asia News Center. https://
news.microsoft.com/apac/2020/03/31/trying-to-shield-hospital-staff-and-
patients-from-covid-19-with-help-from-ai-cloud-and-intelligent-edge.
14 The Internet of medical things
2.1 Introduction
Driving is one of the major experiences linking together people: every day, while
travelling by car, drivers are exposed to different events and situations, which can
impact on their psychophysiological state. External factors such as road design
1
Internet of Things (IoT) Lab, Department of Engineering and Architecture, University of Parma, Parma,
Italy
2
Multimedia Lab, Department of Engineering and Architecture, University of Parma, Parma, Italy
3
Stress Physiology Lab, Department of Chemistry, Life Sciences and Environmental Sustainability,
University of Parma, Parma, Italy
4
Networks for Humans (Net4U) Lab, Department of Electrical and Electronic Engineering, University of
Cagliari, Cagliari, Italy
*
The work of Marco Martalò was carried out at the following place: in 2020, at the IoT Lab of the
University of Parma; and in 2021 while collaborating with the IoT Lab.
20 The Internet of medical things
(motorways vs. rural roads vs. city roads, etc.), road layout (straight vs. curves,
steep road vs. downhill road, etc.), traffic flow (high vs. low), and weather can
influence and increase driving-related stress. To this end, literature studies exam-
ined the relationship between traffic conditions and stress levels and, as expected,
they found that driving-related stress is greater in high traffic jam areas rather than
in low congestion ones [1–3]. Moreover, it has been observed that these stressful
situations may alter the driver’s psychophysiological state.
In this regard, HRV represents an indirect and noninvasive measurement of
beat-to-beat temporal changes in heart rate (HR), which reflects cardiac autonomic
influences, particularly of vagal origin, at the sinoatrial node of the heart. HRV
analysis has been extensively applied in various research fields, including psy-
chophysiology, cardiology, and psychiatry and has been increasingly recognized as
a biomarker of health and stress. In fact, a healthy subject is characterized by higher
levels of resting HRV that are in turn associated with better flexibility and adapt-
ability to environmental challenges [4]. Moreover, HRV analysis is able to index
psychophysiological states during stressful conditions or mental efforts (such as
during driving activities) [4]. Indeed, lower tonic HRV has been associated with
psychosocial stress and mental workload [5–7], whereas higher HRV reflects the
capability of an individual to successfully adapt to external stimuli and manifesting
itself in a reduction in daily performance [8]. An example of the potential utility of
HRV as a biomarker of physiological driver’s state is shown in [9], where driving-
related stress is associated with a reduction in HRV indexes of vagal tone.
From a more technological perspective, the definition and analysis of mon-
itoring systems inside vehicles that are able to simultaneously collect physiolo-
gical indexes and useful data to determine the status of the driver–vehicle system
is a topic of strong interest. The final aim for both academic players and industrial
manufacturer is to improve safety and security while driving for both drivers and
passengers [10]. Furthermore, recent technological progress has introduced new
possibilities with respect to traditional manual driving, especially with regard to
mechanisms aiming at vehicle—e.g., cars, trucks, etc.—driving assistance, such
as onboard advanced driver-assistance systems (ADAS) [11]. In detail, examples
of ADAS mechanisms equipping modern vehicles are antilock braking system
[12], adaptive cruise control [13], electronic stability control (ESC) [14], lane
departure warning system [15], forward collision warnings [16], traffic sign
recognition [17], automotive night vision [18], alcohol ignition interlock devices
[19], collision avoidance systems [20], and driver drowsiness detection [21].
Then, as a common objective, all these ADAS aim to compensate for human
errors, in order to reduce road fatalities—activating alarms or, when necessary,
taking control of the vehicle itself.
Furthermore, focusing on a more general perspective, in the last years the
ubiquitous diffusion of the IoT paradigm [22] has influenced and changed our
lifestyle, thanks to the ability to interconnect heterogeneous devices (such as sen-
sors, actuators, or more in general, smart objects [23,24]), combining different
technologies and communication protocols to build services for end users [25].
Thanks to this heterogeneity, IoT applications are innumerable—nowadays often
Noninvasive psychophysiological driver monitoring 21
cameras, useful integration could involve the use of normal and general-purpose ima-
ging sensors, such as in-vehicle fixed cameras, or the front camera of a smartphone
[38,39]. In detail, these additional video sensing elements can integrate the experi-
mental data obtained by thermal cameras, thanks to their well-known portability, thus
representing an optimal way for gathering information [40].
Data processing
V2X BLE
ECU
HUB
Wi-Fi
CAN
Figure 2.3 FLIR One Pro LT thermal camera positioning inside the vehicle
end, it is necessary to carefully and accurately define the position of the smartphone
connected with the thermal camera; this is as much true as it represents a trade-off
between the quality of the framing—further image-processing and analysis tasks
require a frontal viewpoint—and the degree of obtrusiveness for the driver’s per-
spective toward the windscreen.
(ii) low frequency (LF), and (iii) high frequency (HF) bands. The VLF compo-
nent reflects R–R interval variations that are due to long-term regulation
mechanisms (e.g., thermoregulation and hormonal mechanisms). The LF band
reflects a mix between sympathetic and vagal influences. The HF band reflects
vagal tone and is linked to respiratory-related changes in cardiac autonomic
modulation [45].
City 2
City 1
Highway
Highway
Rest
Baseline
Figure 2.5 Urban and highway roads traveled in the city of Parma, Italy, during
the experimental driver monitoring
95
90
HR (bpm)
85
80
ine
y1
y2
st
wa
wa
Re
sel
Cit
Cit
gh
gh
Ba
Hi
Hi
Condition/Location
38
36
34
RMSSD (ms)
32
30
28
26
24
ine
y1
y2
st
wa
wa
Re
sel
Cit
Cit
gh
gh
Ba
Hi
Hi
Condition/Location
Figure 2.7 Experimental RMSSD values obtained during the driving route
700
600
HF (ms2)
500
400
300
ine
y1
y2
st
wa
wa
Re
sel
Cit
Cit
gh
gh
Ba
Hi
Hi
Condition/Location
In Figure 2.8, the values of HF (dim: [ms2]) are shown. As for RMSSD, the
driving phase was characterized by a reduction of HF values in both city and
highway scenarios, thus suggesting a decrease in parasympathetic modulation.
Finally, the average RR values (dim: counts per minute [cpm]) are shown in
Figure 2.9. The R–R interval was steady during the overall recording session.
Noninvasive psychophysiological driver monitoring 29
18.5
18
17.5
RR (cpm)
17
16.5
16
15.5
15
ine
y1
y2
st
wa
wa
Re
sel
Cit
Cit
gh
gh
Ba
Hi
Hi
Condition/Location
Figure 2.9 Experimental R–R values obtained during the driving route
These preliminary results suggest that the driving experience can modulate the
psychophysiological state of the driver, as hinted by the reduction in HRV para-
meters associated with the increase of HR. Therefore, further driving recording
sessions with a larger sample size are needed to confirm these results and indicate
the extent to which HRV parameters may vary according to the psychophysiolo-
gical state of the driver in different driving scenarios. Moreover, a future integra-
tion of thermal imaging data would provide a clearer picture of the physiological
correlates of drivers’ mental effort.
Acknowledgments
The work of the authors is supported by the European Union’s Horizon 2020
research and innovation program ECSEL Joint Undertaking (JU) under grant
30 The Internet of medical things
References
[1] Hennessy DA and Wiesenthal DL. The Relationship Between Traffic
Congestion, Driver Stress and Direct Versus Indirect Coping Behaviours.
Ergonomics. 1997; 40(3): 348–361. doi:10.1080/001401397188198.
[2] Hill JD and Boyle LN. Driver Stress as Influenced by Driving Maneuvers and
Roadway Conditions. Transportation Research Part F: Traffic Psychology and
Behaviour. 2007; 10(3): 177–186. doi:10.1016/j.trf.2006.09.002.
[3] Neighbors C, Vietor NA, and Knee CR. A Motivational Model of Driving
Anger and Aggression. Personality and Social Psychology Bulletin. 2002;
28(3): 324–335. doi:10.1177/0146167202286004.
[4] Thayer JF, Hansen AL, Saus-Rose E, et al. Heart Rate Variability, Prefrontal
Neural Function, and Cognitive Performance: The Neurovisceral Integration
Perspective on Self-regulation, Adaptation, and Health. Annals of Behavioral
Medicine. 2009; 37(2): 141–153. doi:10.1007/s12160-009-9101-z.
[5] Lischke A, Jacksteit R, Mau-Moeller A, et al. Heart Rate Variability is Associated
With Psychosocial Stress in Distinct Social Domains. Journal of Psychosomatic
Research. 2018; 106: 56–61. doi:10.1016/j.jpsychores.2018.01.005.
[6] Castaldo R, Melillo P, Bracale U, et al. Acute Mental Stress Assessment via
Short Term HRV Analysis in Healthy Adults: A Systematic Review With
Meta-Analysis. Biomedical Signal Processing and Control. 2015; 18: 370–
377. doi:10.1016/j.bspc.2015.02.012.
[7] Mulder T, De Waard D, and Brookhuis KA. Estimating Mental Effort Using
Heart Rate and Heart Rate Variability. In: Stanton NA, Hedge A, Brookhuis
K, Salas E, Hendrick HW, editors. Handbook of Human Factors and
Ergonomics Methods. 1st ed. Boca Raton, FL: CRC Press; 2004. p. 227–236.
doi:10.1201/9780203489925-30.
[8] Melillo P, Bracale M, and Pecchia L. Nonlinear Heart Rate Variability
Features for Real-Life Stress Detection. Case study: Students Under Stress
Due to University Examination. BioMedical Engineering OnLine. 2011; 10
(1): 96. doi:10.1186/1475-925x-10-96.
[9] Riener A, Ferscha A, and Aly M. Heart on the Road: HRV Analysis for
Monitoring a Driver’s Affective State. In: Proceedings of the 1st International
Conference on Automotive User Interfaces and Interactive Vehicular Applications.
Noninvasive psychophysiological driver monitoring 31
AutomotiveUI ’09. New York, NY, USA: Association for Computing Machinery;
2009. p. 99–106. doi:10.1145/1620509.1620529.
[10] Davoli L, Martalò M, Cilfone A, et al. On Driver Behavior Recognition for
Increased Safety: A Roadmap. Safety. 2020; 6(4): 1–33. doi:10.3390/
safety6040055.
[11] Ziebinski A, Cupek R, Grzechca D, et al. Review of Advanced Driver
Assistance Systems (ADAS). AIP Conference Proceedings. 2017; 1906(1):
120002. doi:10.1063/1.5012394.
[12] Satoh M and Shiraishi S. Performance of Antilock Brakes with Simplified
Control Technique. In: SAE International Congress and Exposition; 28
February–4 March. Detroit, MI: SAE International; 1983. doi:10.4271/
830484.
[13] Vollrath M, Schleicher S, and Gelau C. The Influence of Cruise Control and
Adaptive Cruise Control on Driving Behaviour—A Driving Simulator Study.
Accident Analysis & Prevention. 2011; 43(3): 1134–1139. doi:10.1016/j.aap.
2010.12.023.
[14] Hall-Geisler K. How Electronic Stability Control Works 2021 [cited 11 May
2021]. Available from: https://auto.howstuffworks.com/car-driving-safety/
safety-regulatory-devices/electronic-stability-control.htm.
[15] Kortli Y, Marzougui M, and Atri M. Efficient Implementation of a Real-Time
Lane Departure Warning System. In: 2016 International Image Processing,
Applications and Systems (IPAS); 2016. p. 1–6. doi:10.1109/IPAS.2016.788
0072.
[16] Wang C, Sun Q, Li Z, et al. A Forward Collision Warning System Based on
Self-Learning Algorithm of Driver Characteristics. Journal of Intelligent &
Fuzzy Systems. 2020; 38(2): 1519–1530. doi:10.3233/JIFS-179515.
[17] Luo H, Yang Y, Tong B, et al. Traffic Sign Recognition Using a Multi-Task
Convolutional Neural Network. IEEE Transactions on Intelligent
Transportation Systems. 2018; 19(4): 1100–1111. doi:10.1109/TITS.2017.
2714691.
[18] Martinelli NS and Seoane R. Automotive Night Vision System. In:
Thermosense XXI. vol. 3700. International Society for Optics and Photonics.
SPIE; 1999. p. 343–346. doi:10.1117/12.342304.
[19] Centers for Disease Control and Prevention (CDC). Increasing Alcohol
Ignition Interlock Use; 2016 [cited 28 April 2021]. Available from: https://
www.cdc.gov/transportationsafety/impaired_driving/ignition_interlock_states.
html.
[20] Fuller J. How Pre-Collision Systems Work; 2009 [cited 28 April 2021].
Available from: https://auto.howstuffworks.com/car-driving-safety/safety-
regulatory-devices/pre-collision-systems.htm.
[21] Jacobé de Naurois C, Bourdin C, Stratulat A, et al. Detection and Prediction
of Driver Drowsiness Using Artificial Neural Network Models. Accident
Analysis & Prevention. 2019; 126: 95–104. 10th International Conference on
Managing Fatigue: Managing Fatigue to Improve Safety, Wellness, and
Effectiveness. doi:10.1016/j.aap.2017.11.038.
32 The Internet of medical things
[22] Atzori L, Iera A, and Morabito G. The Internet of Things: A survey. Computer
Networks. 2010; 54(15): 2787–2805. doi:10.1016/j.comnet.2010.05.010.
[23] Davoli L, Belli L, Cilfone A, et al. Integration of Wi-Fi Mobile Nodes in a
Web of Things Testbed. ICT Express. 2016; 2(3): 96–99. Special Issue on ICT
Convergence in the Internet of Things (IoT). doi:10.1016/j.icte.2016.07.001.
[24] Atzori L, Iera A, and Morabito G. From “Smart Objects” to “Social Objects”:
The Next Evolutionary Step of the Internet of Things. IEEE Communications
Magazine. 2014; 52(1): 97–105. doi:10.1109/MCOM.2014.6710070.
[25] Belli L, Cilfone A, Davoli L, et al. IoT-Enabled Smart Sustainable Cities:
Challenges and Approaches. Smart Cities. 2020; 3: 1039–1071. doi:10.3390/
smartcities3030052.
[26] Belli L, Cirani S, Davoli L, et al. An Open-Source Cloud Architecture for Big
Stream IoT Applications. In: Podnar Žarko I, Pripužić K, and Serrano M,
editors. Interoperability and Open-Source Solutions for the Internet of Things:
International Workshop, FP7 OpenIoT Project, Held in Conjunction with
SoftCOM 2014, Split, Croatia, September 18, 2014, Invited Papers. Springer
International Publishing; 2015. p. 73–88. doi:10.1007/978-3-319-16546-2_7.
[27] Welch KC, Harnett C, and Lee YC. A Review on Measuring Affect with
Practical Sensors to Monitor Driver Behavior. Safety. 2019; 5(4): 72.
doi:10.3390/safety5040072.
[28] Li M and Narayanan S. Robust ECG Biometrics by Fusing Temporal and
Cepstral Information. In: 2010 20th International Conference on Pattern
Recognition; 2010. p. 1326–1329. doi:10.1109/ICPR.2010.330.
[29] Abo-Zahhad M, Ahmed SM, and Abbas SN. Biometric Authentication
Based on PCG and ECG Signals: Present Status and Future Directions.
Signal, Image and Video Processing. 2014; 8(4): 739–751. doi:10.1007/
s11760-013-0593-4.
[30] Cassani R, Falk TH, Horai A, et al. Evaluating the Measurement of Driver
Heart and Breathing Rates from a Sensor-Equipped Steering Wheel Using
Spectrotemporal Signal Processing. In: 2019 IEEE Intelligent Transportation
Systems Conference (ITSC); 2019. p. 2843–2847. doi:10.1109/ITSC.2019.
8916959.
[31] Lanatà A, Valenza G, Greco A, et al. How the Autonomic Nervous System
and Driving Style Change With Incremental Stressing Conditions During
Simulated Driving. IEEE Transactions on Intelligent Transportation
Systems. 2015; 16(3): 1505–1517. doi:10.1109/TITS.2014.2365681.
[32] Eilebrecht B, Wolter S, Lem J, et al. The Relevance of HRV Parameters for
Driver Workload Detection in Real World Driving. In: 2012 Computing in
Cardiology. Krakow, Poland; 2012. p. 409–412.
[33] Samarin N and Sannella D. A Key to Your Heart: Biometric Authentication
Based on ECG Signals. CoRR. 2019; abs/1906.09181. Available from:
http://arxiv.org/abs/1906.09181.
[34] Coughlin JF, Reimer B, and Mehler B. Monitoring, Managing, and
Motivating Driver Safety and Well-Being. IEEE Pervasive Computing.
2011; 10(3): 14–21. doi:10.1109/MPRV.2011.54.
Noninvasive psychophysiological driver monitoring 33
The Internet of Things (IoT) enables physical items and devices to see, hear, and
think by exchanging data. The IoT works in the other direction, converting ordinary
items into intelligent ones. Embedded devices, communication protocols, sensor
networks, Internet protocols, and applications are examples of IoT-specific tech-
nology. Certain IoT-based healthcare solutions, such as mobile health and telecare,
as well as preventative, diagnostic, therapeutic, and monitoring systems, funda-
mentally modify everything. Wireless body area networks (WBANs) and radio-
frequency identification (RFID) are unquestionably important components of the
IoT. This chapter is covered, in addition to research on the usage of IoT in the field
of biomedical systems to be applied within the framework.
3.1 Introduction
The fast progress of IoT technology in recent years has permitted the connecting of a
huge number of smart items and sensors, as well as the formation of seamless data
interchange between them. As a result, data analysis and storage systems such as
cloud computing and fog computing are required. Healthcare is one of the IoT
application areas that has aroused the interest of industry, academia, and government.
The advancement of IoT and biomedical healthcare is enhancing patient
safety, staff happiness, and operational efficiency in the medical industry.
Biomedical engineering is a relatively new area of study that integrates funda-
mental principles from physical and applied science to life and medicine. Engineers
and experts in biomedical engineering work at the interface of engineering,
healthcare and biological sciences. Recent advancements in the semiconductor
industry, such as the down-scaling of electrical devices, as well as advancements in
computer sciences, such as the advent of data science and cloud computing, have
sparked an explosion of novel therapeutic applications. The effect was slow, but it
helped the whole field of medicine. Nowadays, healthcare tailored to individual
pathological requirements, the retention of long-term data (ECG, SPO2, etc.), and
1
School of Computer Science and Applied Mathematics, University of the Witwatersrand, Johannesburg,
South Africa
36 The Internet of medical things
Internet
Physiological signal
Heart
ECG
ABP Feedback
Arteries
PPG Classification and analysis
Outpatient care
EHR systems
Electronic
medical consultation
Smart emergency
Network-based
community- EHR
Internet of
Things
Healthcare PAN
Elderly monitoring
Smart medicine
Heart Rate
sensor
Motion
sensor
Patient management
Nursing-care
server
Nurses with
mobile gateway
Rehabilitation
Smart biosensor
Fixed environmental sensor
biodata, with the caregiver engaging with them through a portable smart smartphone.
The patient gives biometric data, namely, electrocardiography, electroencephalography,
electromyography, and blood pressure. Following that, the caregiver may combine
ambient data and the patient’s biodata to make a more rapid assessment of the patient’s
requirements. As a consequence, patients will get more precise and timely therapeutic
intervention. Numerous studies have been conducted on physical injury healing for
enhancement of biomedical applications using IoT. In [4], Fan et al. describe the process
of creating a framework that culminates in a recovery strategy personalized to an indi-
vidual’s symptoms. This is accomplished by linking the patient’s diagnosis to a history of
past patients’ symptoms, illnesses, and therapies. The technology allows doctors to
manually input symptoms and approve prescription medication; in 87.9% of cases. The
physician followed the system’s instructions to the letter and made no changes to the
recommended treatment plan.
The WBAN is made up of implanted sensors that monitor vital indicators as well as
specific organs inside the body. Sensors capture data from various parts of the body
and transmit it to the sink. Recent research has concentrated on the dependability,
security, and efficiency of sensor communications in order to convey data from
sensors to sinks. Reduced network life span is a critical criterion in WBANs, and
the majority of recent publications have focused on developing cooperative routing
algorithms that achieve low-energy consumption and increased network longevity.
The packet reception ratio, on the other hand, is another statistic that has been used
into the routing algorithms of WBANs in order to determine whether or not packets
were successfully received in the sink [5].
40 The Internet of medical things
Reference [6] assesses the usefulness of existing IoT technologies as a tool for
monitoring Parkinson’s disease patients. Their research implies that wearable
monitors capable of detecting gait patterns, tremors, and general activity thresholds
might be used in conjunction with vision-based sensors (i.e., webcams) in the home
to detect Parkinson’s disease beginning. Additionally, the authors believe that
machine learning might result in more successful treatment options in the long run.
According to [7], WBAN is on the way to improve biomedical application based on
IoT. It consists of a network of small intelligent devices that can communicate
wirelessly and analyze data. The garment incorporates sensing knots and may be
worn as clothes. There are numerous things that are unique or may be worn over
others. The growth of technology has resulted in a reduction in the size of elec-
tronic devices capable of wireless communication, while increasing their proces-
sing and data storage capabilities. WBANs have become a crucial component of
medical monitoring systems due to their ability to be used at any time and in
any place.
Figure 3.4 displays the fundamental WBAN structure as well as information
transmission strategies collected from individuals. This structure’s first layer
facilitates wireless communication, allowing sensor nodes and nodes to connect
with the network structure on the higher layer [9]. A Wi-Fi gateway is provided
(PDA, smart mobile phone, etc.) for communication. The gateway is often a
computer with an Internet connection that is attached to an access point and wire-
lessly captures data sent over it. The data gathered is kept on this computer.
Currently, a medical monitoring mechanism is in place, enabling all data to be
analyzed. It is preserved in such a manner that authorities may access it remotely.
EEG
Vision sensor
Temperature sensor
ECG
Breathing sensor
Blood pressure
Glucose level
Network
Motion sensor
Since the primary goal of WBAN is to improve people’s lives, several appli-
cations are accessible. For the most part, these applications may be classified as
either medical or nonmedical. Monitoring physiological aspects of the human body
in order to identify anomalies and warn relevant people in real-time is one example
of a medical application. Nonmedical uses include entertainment, emotion detec-
tion, secure identification, and nonmedical emergencies that are accomplished by
collecting environmental data and alerting people to hazards like as fire [10]. While
WBAN systems are often used to monitor a single person, collaborative WBAN
systems may be used to monitor a group of people.
Asset tracking
system Drug
Patient container
Doctor Patient
identification
system RFID
tag
Drug
administration Medical
Asset supplies
system
Nurse
Access control
system
RFID readers
Patient Terminal
Internet/
wireless network
The author of [11] suggested a simple method for diabetics to assess their
blood glucose levels. Patients must manually measure blood glucose levels at
certain intervals during this technique. It next goes through two types of blood
sugar irregularities. The first includes increased blood glucose levels, while the
second includes a failure to assess blood glucose levels. Furthermore, the system
determines how to tell the patient, his or her parents and family friends, or emer-
gency healthcare professionals such as physicians, dependent on the severity of the
irregularity. While this methodology is feasible and has been shown to be feasible,
it might be improved by automating blood glucose monitoring.
In authors’ suggestions [12], an off-the-shelf component and a custom antenna
were combined to create a device for predicting cardiac disorders. A micro-
controller analyses data from an ECG sensor to identify the heart rhythm.
This data is then wirelessly transferred to the user’s device through Bluetooth,
where it is processed and presented in a user application. The authors suggest that
creating algorithms to forecast heart attacks might help the approach. Additional
progress might be achieved by incorporating respiration rate monitoring, which has
been shown to aid in the detection of cardiac arrests [13].
IoT-based biomedical healthcare approach 43
results in decreased time spent acquiring all necessary lab tests. Additionally, they
are advantageous for a number of reasons, including relieving demand on an over-
burdened healthcare system.
emerging side effects, as well as addressing and alleviating concerns about it.
Better treatments with smart technology may reduce patients’ inconveniencing.
fact that smartphone and computer technology have been in use for a long time, the
increasing availability of apps and the IoT-based biomedical applications, wireless
technology, and the virtual economy has all contributed to the rapid IoT, causing
the overall technology ecosystem of society to expand [16].
Other physical instruments (sensors, actuators, and so on) have been com-
bined with IoT devices to gather and communicate data through Bluetooth, Wi-
Fi, and IEEE 802.11 [17]. Heart-related applications incorporate clinical data
about the patient using integrated or wearable sensors such as thermistors, a
palpator (a pressure sensor), a lead electrograph physiologist’s (an ECG), and
electro-of pathologist’s the brain’s electric potential (an EEG). Environmental
parameters that may be recorded include temperature, humidity, date, period,
and time of day. These medical information may be used to make intriguing and
precise conclusions about a patient’s clinical status. Because IoT sensors gen-
erate/capture a variety of data through the Internet, enormous amounts of data
are provided by a variety of sources (sensors, mobile phones, e-mail, software,
and applications). The previous research’s findings are made accessible to phy-
sicians, caregivers, and anyone else allowed to use the devices. The growth of
cloud/server-based stile diagnosis and the transmission of these facts through
healthcare services is an efficient use of information, and cures are delivered as
soon as possible when required.
Because the app’s users, the healthcare institution, and the contact module all
function together, all parties have access to the app’s data. The bulk of IoT com-
ponents are user-interface-based (IoT in the framework largely functions as a
dashboard for medical practitioners, enabling patient monitoring, data visualiza-
tion, and apprehension capabilities). The research revealed that IoT in healthcare
has improved on previous year’s results by delving further into these issues [14].
Healthcare monitoring, regulation, and privacy are three potential IoT applications
that should be examined first. These technical advances demonstrate the IoT
potential for success and profit in the healthcare business. The fundamental pro-
blem, however, is to maintain service quality matrices that encourage knowledge
exchange, stability, cost efficiency, and resilience while also maintaining all users’
data privacy [18].
Table 3.1 summarizes the contributions of several academics to IoT in
healthcare. In 2019, Dang et al. developed a cloud-based method for processing IoT
data in healthcare [24]. They selected to address a number of security problems and
difficulties related to the use of IoT and cloud computing in healthcare. In 2019
[27] and 2018 [23,28], a significant number of peer-reviewed journal articles were
released, both of which searched for prior-to-2017 publications. Nazir et al. per-
formed a systematic review of research published between 2011 and 2019 that
addressed the security and privacy risks associated with mobile-based healthcare
IoT [27].
There are presently no research papers that are solely devoted to medical IoT
applications; nonetheless, Table 3.1 provides relevant research papers on healthcare
IoT. The table summarizes the outcomes of ten independent researches on IoT in
healthcare (e.g., systemic analyses and other types of evaluations). The initial
48 The Internet of medical things
3.10 Conclusion
One of the aims of e-health is to deliver healthcare services to patients in their
homes, especially via the use of IoT-based biomedical applications. In general, IoT
applications are intended to save money and stimulate patients at home, resulting in
increased patient engagement. Everyone will benefit from improved health pro-
motion and a more fulfilled lifestyle as a consequence of this.
According to the study’s findings, the IoT-based biomedical applications in
medicine are still in its infancy. The utilization-limiting application seems to have
put a significant strain on the healthcare system in a variety of subfields.
Furthermore, since the number of medical IoT studies and research fields increased
in 2018, the number of studies on this topic will grow in the future, resulting in the
engagement of new research areas.
References
[1] C.-F. Chiang, F.-M. Hsu, and K.-H. Yeh. Robust IOT-based nursing-care
support system with smart bio-objects. Biomedical Engineering Online, 17
(2): 1–15, 2018.
[2] Y.J. Fan, Y.H. Yin, L.D. Xu, Y. Zeng, and F. Wu. IoT-based smart rehabi-
litation system. IEEE Transactions on Industrial Informatics, 10(2): 1568–
1577, 2014.
[3] R. Jafari and M. EffatParvar. Cooperative Routing Protocols in Wireless
Body Area Networks (WBAN) : A Survey, 2017.
[4] C.F. Pasluosta, H. Gassner, J. Winkler, J. Klucken, and B.M. Eskofier. An
emerging era in the management of Parkinson’s disease: wearable technol-
ogies and the internet of things. IEEE Journal of Biomedical and Health
Informatics, 19(6): 1873–1881, 2015.
[5] I. Kırbas. Online Kablosuz İnkübatör İzleme ve Kontrol Sistemi Tasarımı ve
Uygu-laması. PhD thesis, Doktora Tezi, Sakarya Üniversitesi, Sakarya-
Türkiye, 2013.
52 The Internet of medical things
[6] S. Saleem, S. Ullah, and K.S. Kwak. A study of IEEE 802.15.4 security fra-
mework for wireless body area networks. Sensors, 11(2): 1383–1395, 2011.
[7] A. Pise, H. Vadapalli, and I. Sanders. Facial emotion recognition using
temporal relational network: an application to e-learning. Multimedia Tools
and Applications; Springer: Netherland, 2020; pp. 1–21.
[8] S.-H. Chang, R.-D. Chiang, S.-J. Wu, and W.-T. Chang. A context-aware,
interactive m-health system for diabetics. IT Professional, 18(3): 14–22,
2016.
[9] G. Wolgast, C. Ehrenborg, A. Israelsson, J. Helander, E. Johansson, and H.
Manefjord. Wireless body area network for heart attack detection [education
corner]. IEEE Antennas and Propagation Magazine, 58(5): 84–92, 2016.
[10] M.A. Cretikos, R. Bellomo, K. Hillman, J. Chen, S. Finfer, and A. Flabouris.
Respiratory rate: the neglected vital sign. Medical Journal of Australia, 188
(11): 657–659, 2008.
[11] R.C.A. Alves, L.B. Gabriel, B.T. de Oliveira, C.B. Margi, and F.C.L. dos
Santos. Assisting physical (hydro) therapy with wireless sensors networks.
IEEE Internet of Things Journal, 2(2): 113–120, 2015.
[12] A. Pise, H. Chabukswar, and D. Jadhav. Hard authentication using dot array
carrier structure. International Journal of Scientific & Engineering
Research, 4(5): 670–676, 2014.
[13] V. Jagadeeswari, V. Subramaniyaswamy, R. Logesh, and V. Vijayakumar. A
study on medical internet of things and big data in personalized healthcare
system. Health Information Science and Systems, 6(1): 1–20, 2018.
[14] H. Peng, Y. Tian, J. Kurths, L. Li, Y. Yang, and D. Wang. Secure and
energy-efficient data transmission system based on chaotic compressive
sensing in body-to-body networks. IEEE Transactions on Biomedical
Circuits and Systems, 11(3): 558–573, 2017.
[15] A. Gatouillat, Y. Badr, B. Massot, and E. Sejdić. Internet of medical things:
a review of recent contributions dealing with cyber-physical systems in
medicine. IEEE Internet of Things Journal, 5(5): 3810–3822, 2018.
[16] L.M. Dang, M. Piran, D. Han, et al. A survey on internet of things and cloud
computing for healthcare. Electronics, 8(7): 768, 2019.
[17] I. de Morais Barroca Filho and G.S. de Aquino Junior. IoT-based healthcare
applications: a review. In International Conference on Computational
Science and Its Applications, pages 47–62. Springer, 2017.
[18] A. Awad Mutlag, M.K. Abd Ghani, N. Arunkumar, et al. Enabling tech-
nologies for fog computing in healthcare IoT systems. Future Generation
Computer Systems, 90: 62–78, 2019.
[19] S. Nazir, Y. Ali, N. Ullah, and I. Garcı́a-Magariño. Internet of things for
healthcare using effects of mobile computing: a systematic literature review.
Wireless Communications and Mobile Computing, 2019, 2019.
[20] B. Farahani, F. Firouzi, V. Chang, M. Badaroglu, N. Constant, and K.
Mankodiya. Towards fog-driven IoT eHealth: promises and challenges of
IoT in medicine and healthcare. Future Generation Computer Systems, 78:
659–676, 2018.
IoT-based biomedical healthcare approach 53
For the COVID-19 pandemic, international health facilities have been issued in
financial section with difficulty. According to the economic status, several hospi-
tals and healthcare facilities loses a month in sales. Furthermore, providing an
adequate healthcare solution to COVID-19 could cost $52 billion (around $8.60 per
person) for low- and middle-income countries (LMICs). Could burden have a sig-
nificant effect on health treatment, surgeries, and outcomes through the use of
COVID-19. This year the World Bank predicts that the global economy would
contract by about 8%, with developing countries bearing the brunt of the burden.
Lack of planning was a key in dealing with healthcare facility issues everywhere.
On a national scale, items such as gowns, gloves, facemasks, syringes, disin-
fectants, and toilet paper ran out. Healthcare worldwide has felt threatened by
COVID-19’s findings and has responded by formulating new programs to deal with
pandemics. In this article, we will talk about the financial implications of COVID-
19 that include clinics, surgery, and medical procedures on both the US and foreign
healthcare systems, in the United States and abroad. In the case of natural or
human-made disaster, the United States and the rest of the world would make profit
from being prepared for a single blueprint to react.
By December 29, 2019, the novel coronavirus (nCoV) sparked a worldwide
pandemic. Such groups—especially the older adults and those with weakened
immune systems, such as low blood pressure, cardiovascular disease (CVD), or
asthma, or chronic kidney disease—are at an increased risk of the infection.
Systemic socioeconomic status (SES) is linked to a higher prevalence of diabetes,
hypertension, coronary artery disease, and obesity in both demographic groups.
Those already established conditions increase the likelihood of catching the cor-
onavirus (2019-nSRS-2), also known as severe acute respiratory syndrome cor-
onavirus (SARS-CoV-2). Emphasis should be put on the root of the issue at
1
Department of Information Technology, VFSTRU, Guntur, India
2
Krupajal Engineering College, Odisha, India
56 The Internet of medical things
vulnerable periods. The study of the social determinants of health (SDOH), which
involves research on the various facets of a population’s health in the context of
various aspects of their society, will assist policymakers in managing health crises
in ways that treat everybody fairly. During the COVID-19 pandemic, the SDOH
directly affected patient recovery and health outcomes.
Determine how these three variables influence national healthcare, social,
and economic well-being. Many forms of quantitative and nonnumerical data
have been employed. We have found small well-equipped clinics, poor rule
awareness, unemployment, and shortage of research equipment to be influential
in this COVID-19 infection. To combat the spread of diseases, enforcement
policies that are tough yet respect people’s rights could be more effective. The
delivery of adequate medical treatment is necessary for setting up appropriate
medical facilities.
4.1 Introduction
Once there were 7.5 billion inhabitants on Earth, 1 year ago, no one could have
expected COVID-19. World Health Organization (WHO) was the first to identify a
COVID-19 on December 30, 2019 and put the disease on a global emergency list
on January 30, 2020. But also scientists say that the virus was at work for some
time before this. Respiratory problems may occur with single-stranded RNA viru-
ses [1] in the air. It is estimated by the CDC project an infection, or virulence (R0),
of 2.5 for this new Corona virus strain. It found one hundredfold greater lethality of
the mutant variants than of the normal influenza virus [2,3]. When the pandemic
was already new, the global production chain was disrupted.*
About 51% of Americans recorded a loss of income and an increase in
unemployment of 14.7% due to the outbreak [4]. It has been difficult for young
people to make their mortgage and cover the bills this year due to missed income
and jobs. These consequences are more likely to occur in young adults in
America, who have not saved enough for an economic recession. Predictions
show that one in five tenants was likely to face homelessness this year. COVID-19 has
foreign implications as well. As soon as the Chinese government imposed a
mandatory quarantine, several countries lost a large portion of their critical trade
products. It is increasingly clear that many countries, particularly those in the
developing world, heavily rely on China. Unfortunately, several valuable equip-
ment pieces, such as respirators, gloves, were lost in the explosion. In turn, the
lack of appropriate personal protective equipment (PPE) promoted the global
pandemic. The World Bank expects global growth to drop by almost 8% in places
that have seen the most significant losses in development and 2 trillion dollars
this year [5]. COVID-19 has had an influence on all US healthcare services and
foreign organizations as well.
*
Projected to be the leading cause of USS disease costs in 2020. It is estimated that it would cost the US
economy $3.3 trillion. It is forecast to cost the US$3 trillion in the next 2 years [3].
Impact of world pandemic “COVID-19” 57
draw on numerous revenue sources, which was not as bad for them. For certain
physicians who practice in the Department of Veterans Affairs (VA) specialty and
procedural practices, the income impact of cancelled treatments was not as worri-
some since they do not get money for their services the same way fee-for-based
organizations do in the private sector does. Often, as a VA system works as a whole,
it will change the supply chain to serve best regions devastated by the pandemic. The
Department of Veterans Affairs provided over 16,500 intensive care units and offered
3,000 ventilators and 4,000 medical assistances for deployment in the form of
deployable disaster services. The Veterans’ Administration switched several of their
official trips to telehealth care. It was done by the VA through the usage of telehealth,
employing the first Chief Telehealth Officer in 1999. Before the arrival of the pan-
demic, the VA predicted the rapid increase in virtual treatment. Prompted by the first
COVID-19 cases in the United States, they started to invest in the technology,
including providing patients with mobile computers, pulse oximeters, and cell phones
to track their temperature remotely, such as smartphones.
During the plague, there was a campaign to help the elderly of our community.
One small nursing home outbreak in Washington State helped start the pandemic. A
nurse at the nursing home developed an infection called COVID-19 on February 28.
Both employees and residents in the facility reported signs of respiratory disease on
that day. By March 18, a total of 167 COVID-19-infected people, 50 COVID-19
travelers, and a lone agent had been found. Accordingly, the hospitalization rate was
high for 54% of the inhabitants and 33% of the workers, resulting in a fatality rate of
33% [12]. More than 40% of all pandemic patients and staff in long-term care
facilities have succumbed to COVID-19. We found that eight out of every ten cancer
of the COVID-19 population was in people aged 65 and over. Thus, the Coronavirus
Assistance, Relief, and Economic Security Act allotted $5 billion to long-term care
centers and older people’s homes for the elderly. Thanks to the improved funds,
additional infection prevention procedures, as well as personnel and program being
put in place for those unable to leave their house, the nursing home was able to have
extended visitation features to its residents [13].
4.2.3.2 India
It was a step in the right direction for the Indians, who, after ages of isolation and
British colonial domination, had finally regained their freedom and identity, had
the opportunity to start constructing a modern democratic nation on the Indian
model. Trissur is the first in India to be hit by COVID-19, on January 30, 2020. The
government initially employed policies to limit the spread, including quarantine of
persons who traveled from high-risk locations, tracing those who had contracted the
virus, and preventing movement of those with a large caseload of cases. When the
number of cases grew, local dissemination became apparent, attention turned to
dealing with the virus, and control steps began to dominate the Ebola. The moves
India has introduced are similar to those in China, including closing public venues
and rerouting foreign air transport. Furthermore, these regulations inflicted a
financial burden on the local economies and had long-term detrimental effects in
the agriculture, industrial, and utility sectors [15,16]. The disease spread among the
countries with which India had trade relations and within them, which put a con-
siderable dent in India’s exports. Furthermore, the general pandemic and resulting
mass incarceration have significantly increased anxiety and depression in India.
4.2.3.3 Brazil
COVID-19 did not get to Brazil until February of 2020. Still, it soon became a
celebrity in this field as a hotbed of creativity contrast to some national responses.
State/Federal governments initiated measures. President Bolsonaro attracted wide
backlash for his apparent inattention to the virus. Since he spoke out so passionately
about FU lockdowns, he usually touted the overhyped hydroxychloroquine cure and
had a generally low attitude toward the pandemic. An estimated 13 million people are
cut off from access to safe water in Brazil’s informal slums, poverty-stricken shanty
towns. Due to the global spread of the pandemic, the economic consequences in Brazil
have been substantial. On March 9–13, the equity exchange in São Paulo fell by over
15% to its worst weekly percentage decline since the Great Recession of 2008.
Additionally, the GDP declined by 11.4% in the second quarter of 2020 relative to the
same period in the previous year. As of September of this year, with an incidence of
4 million and a mortality rate of about 6, behind just the United States and Brazil in
the world, the United States in incidence and rate of, the United States, Brazil has over
a million reported cases and over 125,000 people die per year from them [17].
4.2.3.4 Singapore
Singapore’s reaction to the pandemic has been extolled as one of the most pro-
mising ones in the world. With the first cases arising in December 2019, the nation
was one of the first to register COVID-19 and initially was second only to China in
overall cases. The SARS-CoV epidemic of 2002 revealed several vulnerabilities in
Singapore’s capacity to cope with pandemics and was the catalyst for the nation to
patch up those deficiencies. The country developed the National Centre for
Infectious Diseases (NCID) and 900 quick responses to public health preparedness
clinics. Daily drills were held to simulate the arrival of a pandemic. As a con-
sequence, Singapore was well prepared to conduct mass screening of possibly
62 The Internet of medical things
infectious persons. They have identified themselves as one of the most prolific in
research and touch tracing. The National University of Singapore (NUS) Yong Loo
Lin School of Medicine created an engaging series of comics entitled “COVID-19
Chronicles” that provides essential information about the virus in a format that is
easily understood by most demographics. Messages such as these have been widely
disseminated across the world and have received universal acclaim.
Additionally, Singapore used several strategies similar to other countries that
suffered outbreaks, such as reducing the scale of meetings, promoting physical
distancing, restricting movement, and screening and quarantine anyone entering the
nation. Strict fines were levied for breaching these rules. Interestingly, Singapore
did not initially resort to issuing a systematic shutdown or shutting schools until a
flare-up of cases leads to a 21-day order on April 28, 2020. This was some months
after the virus was first observed in the world [18]. Despite the country’s progress
in managing the contagious burden of the epidemic, owing to the interconnected-
ness of the global economy, Singapore have not sparred the economic effect of the
pandemic as the country recorded 13.2% contraction in GDP in the second quarter
of 2020 as opposed to the previous year [19].
Unpreparedness was a big part of the problem of healthcare services. PPE [22] was
also lacking for healthcare staff. Different healthcare staff used PPE to protect their
health during medical services. The Pakistani analysis discovered that just 34.5%
had helmets, and 0.5% had face masks or goggles. In Jordan, a survey claims that
just 18.5% of physicians had access to PPE. Even the US healthcare sector was
troubled by the accessibility of prescription supplies. Approximately 15% of the
doctors indicated that they did not have N95 respirators. Approximately 50% of the
doctors reported that they did not have facemasks, and approximately 50% reported
that they did not have complete suits or gowns. Furthermore, about 7% of physi-
cians recorded caring for patients who had contracted COVID-19, and approxi-
mately 80% of PPE had been reused. Healthcare services across the world
requested nonmedical PPE donations and resourceful residents modified products
to meet this need. People around the globe faced similar issues with the availability
of intensive care units and ventilators. We did not have the resources to test locally.
Disturbances in global supply chains caused more shortages of food and
consumer goods.
The pandemic has diminished efforts for the diagnosis and treatment of non-
COVID-19 disease as a side effect. Noncommunicable disease (NCD) preventive and
recovery interventions have been compromised since the COVID-19 pandemic
started. As the virus advanced, healthcare staff, who had usually handled non-
catasthma cardiovascular conditions were pulled from treating patients to assist with
the COVID-19 response. In compliance with public health recommendations, certain
procedures and appointments were also delayed. Shrinking of public transportation
impacted patient’s travel times to appointments. Sometimes, cancer and diabetic
patients were unable to get the treatment and medication they needed. Ironically,
these results were seen more in low-income nations, where countries still have a large
percentage of their national income tied up in battling the pandemic. The results of a
survey in Bangladesh, Kenya, Pakistan, and Nigeria have found decreased access to
antenatal and vector control systems, as well as tuberculosis and HIV programs,
before the pandemic. Moreover, there was a notable decline in family income at the
same period as healthcare expense rose [23]. COVID-19 also prompted healthcare
providers to develop contingency plans that guarantee that necessary treatment is not
postponed or denied due to a global viral or another epidemic. Alternative solutions
such as telemedicine have gained a significant foothold in healthcare settings
worldwide and have aided in mitigating the impact of the epidemic on people.
these two specialty practices that asked “excessive financial challenges due to the
trauma pandemic.” Ironically, being self-employed was shown to be a strong
indicator of financial hardship, one of which was referred to as “deplete” in the
study’s multiple regression model. Thus, according to their multiple regression
analysis, self-employed trauma and orthopedic surgeons should be supported more.
Like their cardiac surgeons, this profession shares a similar viewpoint that the
pandemic’s impact on the cost of their operations stopped unnecessary procedures.
The specialized trauma and orthopedic surgeons are unmoved by financial con-
siderations. Results indicated that nearly one-third of the group would give up
practicing because of the pandemic, and nearly two-thirds might benefit from
increased healthcare funding.
On the other hand, telemedicine offers a silver lining, with 44% of participants
believing that it will rise in value in the future. While the survey was established in
rather extraordinary conditions, a lot of open questions remain about its psychometric
testing. In 1 month at his 22-year-old private practice, there was a 29% reduction of
money on every occasion due to the worldwide epidemic. He lost money at a private
practice which he started in Italy before the pandemic 2 years before it struck
entirely. According to the authors of this chapter, all surgical professionals in all
countries who want to limit their career to just cosmetic surgery had done so due to
the lack of options due to financial difficulty. Overall, it describes the first instance of
a partnership between Latin American and worldwide healthcare stakeholders to
show the substantial contribution of COVID-19 healthcare resources to the medical
service economy [24]. Finally, there is a discussion of COVID’s economic effect on
the Latin American radiation oncology field. The survey conducted in Latin-
American countries like Argentina, Costa Rica, and Peru (the three countries had
fewer than 3% of the total of radiation oncology treatment centers) showed that both
had seen a 20% revenue loss from closure inactivity. Less than 3% of the total
radiation oncology facilities were still operational during the pandemic. They admi-
nistered the ASTRO survey in Latin America to view the pandemic’s overall finan-
cial effect on an underutilized field of healthcare. In Latin American radiation
oncology, use of telemedicine as a tool of the future was predicted.
minimize the incidence of disease. Whenever they wear masks, risk to their coworkers
decreases. Since masks avoid involuntary hand and mouth gestures, they are helpful in
terms of personal well-being. The regular use of caps, gloves, mask, and glasses to
reduce the spreading of disease. The vaccines must be available to everyone, espe-
cially those at a risk of COVID-19 infections must be easily accessible.
Antimicrobial gels from the Centers for Disease Control may be made non-
contagious by primary low-level disinfectants. Usually, they are between 50%
alcohol and 60% gel solutions. This agent protects the alcohol from bacterial eva-
poration; therefore, there is plenty of time for the virus particles to come in touch
with it. There are significant considerations that have to be considered: focus and
touch time. Alcohol in water just reduces the total alcohol content gradually, as it
evaporates rapidly, which prevents more extended touch periods. An intact skin
barrier cannot keep viruses out of the mucus membranes but can provide defense
from transferring to them. Thorough handwashing of soap and water at the start of
the day reduces the infective bioburden brought in from the outside from interac-
tion with contaminated surfaces and suspended droplets and aerosols. Daily hand
hygiene after interaction with all aspects of the patient’s treatment to decrease the
risk of disease transmission can necessitate special hand wipes. One may minimize
the chance of being sick by washing hands, arms, face, or spitting in that manner.
Because it is safer for staff, it also helps prevent the movement of infective bacteria
to workspaces such as desks, computers, mice, and touch screens. The surface of
the tools, including drawer handles and bars, should be disinfected according to the
number of employees who use them. Plastic enclosures are more efficient at
keeping liquids away from compromising electronic connections through main-
taining color-coded containers and gel dispensers in work environments; con-
formity with safety disinfection procedures is supported by visual cues.
Mechanical ventilation controls lower the particle level of HVAC (reduces the
concentration of bacteria in the airflow). The usage of higher airflow speeds (above
612 exchanges/hour) at sites over 612 volumes helps remove infectious aerosols. Still,
it prevents desiccate or sedimentation of virus-laden or other aerosol particles. When
working in the ventilation and HVAC environments, work only in the fresh air inflow
zones and stay away from the return air inflow zones. Break rooms offer unique
obstacles. Lower volumetric airflow has the effect of making aerosols more persistent.
During meals, masks are off. This makes it easier for the droplets and particles to be
dispersed, facilitating the spreading of the infection(s). The locations that incorporate
some high-frequency touch surfaces, such as the door handles, the coffee cup, and the
tabletops. Cultural considerations should take precedence. Fatigue leads to a loss of the
ability to do repetitive and complex activities. It degrades for many hours; however,
fully after 12–16 h. The standards for infection prevention and prescribed practices are
likely to become less stringent as an individual’s level of exhaustion increases. There is
a greater likelihood of infection because of illness among people who live and function
close to each other. Controls on productivity are enormous. When job tasks are made to
anyone at random, the spread of personal creative ideas is maximized. Studies done by
Mascha et al. showed that consistency in team tasks would help minimize staff cross-
immunity. They recommend 1 week of work and 1 week of vacation for each
66 The Internet of medical things
People started being sick as the nCoV (Chinese New Year, January 2020) traveled
quickly through China. On December 29, 2019, the first four nCoVs were
announced. Wuhan is the biggest city in Hubei province, which is located in the
middle of central China. The symptom was strange pneumonia of no apparent
cause [26]. Early evidence shows the business touch. More and later, it was dis-
covered that disease could be spread from person to person, often by human touch.
The specific illness triggered by 2020-NP was called by the WHO on February 11,
naming it COVID-19 on the follow-up date [27]. People with preexisting asthma,
CVD, hypertension, chronic kidney disorder, and/an obesity are particularly vul-
nerable to the findings of the CDC proved that all who have these conditions have a
higher risk of serious illness from COVID-19. Although SES may increase the risk
of CVD, particularly among those who have asthma or hypertension and many
others, which are also risk factors of CKD, including depression, diabetes, and
obesity (SES). CDC estimates that 94% of COVID-19-treated patients succumbed
[28]. These particular circumstances increase the likelihood of SARS-2 infection,
and as a result, people in these areas must have the support and services to survive.
This study explains the importance of SDOH during pandemic (COH). Precausal
precautions should be taken to keep the transmission within abounds.
Health and
healthcare
Social and
Economic Social community
stability determinants of context
health
Neighborhood and
Education built
environment
structure, health, education, housing, food, general safety, recreation, and nutrition
(SDOH). In economic hardship, the number of people who have activity-impairing
chronic conditions goes up. Many people become unemployed and are afflicted by
homelessness, drug misuse, or become sick due to the effects of unemployment. An
explanation of SDOH is that there are financial resources, such as a salary, life
experience, and various forms of social help. Simply placed, in short, you are the
way you were born, learn to be, that is, the way you live and function your whole
life [31]. They see the entire guy. These various diseases have a widespread effect
on the well-being of people and the welfare of whole populations. Because of these
socioeconomic differences, SES can be equated with social inequality (SES). The
better on the socioeconomic scale, the worse the health consequences. Lifespan is
shortened for people on the bottom half of the social scale. If you look at things
socioeconomically, health problems build on the inequities already befalling a
community section.
Many characteristics are interrelated and play a major function in the eruption
of the COVID-19. For example, an individual’s education affects his or her eco-
nomic stability and what neighborhood he or she is eligible to live in, which affects
his or her social well-being. Consequently, socioeconomic variables are likely to be
important in infection and mortality At the time of data collection, the researchers
discovered certain counties in New York, including the Bronx, Brooklyn, and
Queens, which also had a significantly higher death rate among people of low SES.
This case is another example to consider when considering a child growing up in a
low-income home. It is difficult for low-income families to provide for their chil-
dren, which could lead to low-quality schools.
There would be no equality in education, no matter where this kid resides, as
long as it is different from a wealthier suburb or a more decadent school system
because government schools are financed by local, national, state, and federal
resources [32]. Any funding for our social projects will come from personal income
and property taxes. The affluent areas and districts would raise more money; there-
fore, they have more money to spend. Districts that serve low-income students get
smaller salaries and inadequate services, while poorly educated teachers receiving
small wages are included in low-income schools. It does not matter whether you get
average or below-average grades as well as long as you plan on pursuing an above-
average academic career in college. And if it is from a low-income family, the kid
would always have a few high-paying career prospects. This places the infant in the
same financial situation as his or her parents. People who live in poverty-ridden
communities have higher chances of remaining that way. The five factors can be seen
as a kind of a cycle that affects an ever-changing scenario, even today.
home [38]. This compares to 83% of participants who are in the top wage bracket
and who have access to grocery stores or farmers’ markets in their neighborhoods.
The families who lived in food deserts had more impact on their dietary choices,
blood pressure, and incidence of kidney disease than those who lived in other
locations [38]. It was discovered that people who lived in food deserts and had low
wages had lower serum concentrations of carotenoids. There are ways to get an
idea about how many fruits and vegetables you consume using carotenoids. It was
also discovered that individuals in food deserts have lower average protein,
potassium, sodium, and magnesium consumption but higher average wages. The
quantities of these minerals will reveal one’s acid load in the system. High pH
indicates an alkaline diet [39]. High protein foods (meat, dairy, etc., in particular)
trigger a rise in body acidity. Fruits and vegetables increase the number of organic
compounds in urine. When you have a diet of a lot of acidic foods, you are likely to
have metabolic acidosis and asthma, kidney disease, and other issue-related pro-
blems. Studies have shown that a heavy acid load contributes to obesity [40].
will need ventilator support [45]. Even if some people who are obese, the breathing
muscles can be weaker because of fat on the neck or in the belly, as previously
discussed, and lung volume may be affected because of fat that restricts the ability
to expand the chest and take a full breath [42,45]. The findings also showed an
improvement in the risk of COVID-19 as BMI rose [45].
4.6.1 Determination
People’s perceptions of unfairness and injustice can feed bigotry. Actions benefit
the well-off at the expense of the poor. In both the human and institutional levels,
healthcare levels, discrimination exists. Individual discrimination involves
encounters between patients and their healthcare professional attributable to eth-
nicity, gender, or some other discrimination that leads to a diminished ability to
access care. Healthcare services and the patient’s well-being may be impeded by
experiences that lead to illness. Hierarchical inequality can be seen in the context of
residential segregation along ethnic lines, gender lines, wage inequity, and so on.
The particular form of systemic discrimination (name, skin color, gender, sexual
orientation, national origin, socioeconomic class, age, religion, etc.) may impact
individual people as well as groups. Somehow, residential apartheid finds itself
contributing to the racial differences seen between African and Caucasian com-
munities. African-Americans live in concentrated poverty. Education systems in
these areas are inadequate, unemployment is very high, health insurance is una-
vailable, violence is a constant problem, and residents have no social support. Still,
life for the residents is complex [46]. This is because it is hard to maintain social
isolation in communities that are poor. Any relatives will not be able to keep their
distance from each other. In extreme poverty, minorities and blacks are more likely
to keep positions where it is not possible to telecommute [47]. This COVID-19
emergency has placed many minorities in the position of either needing to work and
putting food on the table, or remaining at home and keeping their families safe,
especially the Latino community because of a sheer number of jobs in the ware-
house work, manufacturing, maintenance, and custodial employment are common
among minorities [47]. While 35% of those infected with the virus do not identify
themselves as either race or ethnicity, it is readily apparent in the current statistics.
Many New York City residents have died from it than any other racial race [46]. In
New York, 29% of the population are made up of Latinos. Of all the COVID-19
people who die in New York, about 34% are people of color—22% of the area.
However, people of African descent in the United States have a 2.4-time greater
risk of death due to this infection than those of other races. The facts for the US
population, broken down by territory, are startling. According to the US Census
Bureau, African Americans are 13% of the overall population, although they have
suffered 32% of the COVID-19 deaths. On the other hand, white people are much
more likely to be at risk because they live in the United States than those in the rest
of the world [48].
Impact of world pandemic “COVID-19” 73
unhealthful choices (higher saturated and trans-fat and higher calories). Individuals
who do not have ready access to fresh fruits and vegetables tend to have inadequate
diets. Because African-American and Latino neighborhoods are more likely to have
a higher concentration of fast food and grocery stores, they are also more likely to
be, on average to be less attractive to someone wishing to establish a new business
in them. However, the ethnic groups often may not have as good health results as
racial groups. Why? People living in food deserts have an increased risk of obesity,
which is mentioned elsewhere. Income is another essential factor of balanced food
consumption. There are solid and numerous studies that indicate people on low
incomes depend on lower nutrient-dense food. Processed goods tend to be more
costly, but they are not always unhealthy. People who cannot afford to buy fresh
foods purchase the packaged ones instead [51]. It is essential to consider food
deserts during a pandemic and that people do not have accessibility to nutritious
foods, particularly. When food is in short supply, it would be impossible for those
who do not have sufficient access to good nutrition or food to eat their fill. People
would have to make more frequent visits to grocery stores to acquire their food,
thereby increasing their chance of infection. Disadvantaged populations living in
food deserts may face a greater risk of malnutrition retailers hoarding nutritious
foods because of the COVID-19 epidemic. Substandard nutrition may make this
more of an issue in food-deprived areas than in wealthier areas.
you live in densely populated areas and spread it. Poorly maintained low-income
households also inhabit public housing. Too many infestation problems of cock-
roaches, rodents, rats, and other vermin are conducted and discovered. Some things
often seen in these conditions were mold, the absence of air conditioning, smoking,
and cigarette use. Children who grow up in a housing project are twice as likely to be
born with asthma than children who grow up in a single-family household. People
living in poverty could be susceptible to COVID-19.
4.7 Academy
35 different languages [54]. The capacity to translate into languages includes Arabic,
Bengali, Chinese, Dutch, Hindi, and several others. The virus details, ways to avoid
catching the virus, and any remedies for illnesses are available on the fact sheets.
This has created an open forum for discussion also for non-native speakers. It is vital
to increase public knowledge of the infection and promote safety measures to stop
the spread of the disease.
60,000
March April May June
10,000
20,000
0
USA
15,000
March April May June
Number of new COVID-19 cases/day
10,000
5,000
0
New york
4,000
March April May June
3,000
2,000
1,000
0
Georgia
400
March April May June
300
200
100
0
Idaho
Figure 4.2 The wax and wane in new cases of COVID-19 per day in USA, New
York, Georgia and Idaho. The graphs were generated using the online
data form CDC and John Hopkins web sites
78 The Internet of medical things
eliminate COVID-19; the national economy would have to be shown what it has
in store.
The new COVID-19 is a growing threat to the global economy, public health, and
national security. As of June 30, 2020, there have been ten million reported cases,
causing about 500,000 deaths in 215 countries (as per reports from the WHO and
Johns Hopkins) [61,62]. I predict that the COVID-19 crisis would trigger sig-
nificant problems in both infectious disease and mass public health outbreaks and
extensive food poisoning in the community. According to an Organization for
Economic and Co-operation Development (OECD) estimate, the global GDP
growth rate is projected to decline to 2.4% by 2020. The possibility of instability,
and a significant decrease in global economic activity due to sharp drops in
production and spending and supply chain disturbances has risen. This is a huge
challenge for many countries as well as for governance and cooperation around
the world. Many people being sick could have a major effect on China’s econ-
omy [63–65]. As of April 17, China’s gross domestic product had risen 13.5%,
investment, and usage by 24.5%, and all dropped by 15% compared to the year
before; however, Chinese people’s unemployment increased to 6.2% for the first
time in the first quarter, according to the figures published by the National
Bureau of Statistics of China.
Entrepreneurship is fundamental to the country’s economy as a whole.
Because of this, for companies, it is imperative to consider the current state of firms
and their coping mechanisms and how the effects of the COVID-19 pandemic on
them are. As soon as the disease epidemic has run its course, researchers will begin
to study corporations’ innovation plans and responses [66–68]. Other sectors have
also been studied to identify possible solutions to the pandemic. However, these
findings have not been supported by observational evidence.
Firms in Guangdong Province were examined to see how they were affected by
the COVID-19 pandemic and if any policy responses could be prepared (the pro-
vince with the highest GDP in China). To explore why the pandemic impacted
corporations, they went on to look into whether the disease has affected them, and
what they expected in the future. Any supportive measures can be implemented.
4.9.4.1 Gibberish
Redeems to concentrate on demand patterns in the public sector and behavior
improvement in the public sector. Pandemics would lead to less GDP and jobs, con-
sonant with our research on the COVID-19 pandemic. This research, however, just
looked at aggregate trends without addressing the companies, which better captures
the effect of the COVID-19 pandemic on the economy. Firms would be significant
because of the effect of the pandemic. Returning to full manufacturing capability was
difficult because of the disease. More than half of the businesses reported problems
with inventory; some of them also claimed they were completely out of material. One-
third of the businesses we surveyed are feeling operational stresses, so we would
estimate that they hire one-third of the workers and incur one-third of the costs. Also,
23.4% of businesses have done away with domestic requests.
Many businesses stated that market penetration was impeded because of client
face-to-face and location meetings being challenging to conduct. Furthermore,
nearly two-thirds of the companies posted increased operational costs, including
postponed orders, distribution taking more prolonged, and less-than-expected pro-
duction output. Much of them had to contend with the financial problems asso-
ciated with their separation. They were still being held on strict accounts for payroll
salaries, insurances, leases, financial obligations, and accounts. As shown by the
answers to question 4 (supply of raw materials, spare parts, other production and
processing materials), the business owners explained that the supply chain is
affected by epidemic-related delays at varying degrees across the regions. Due to
the spread of the disease, several export industries were threatened. Several busi-
nesses announced cancellations and postponed delivery of overseas orders.
online schooling and 5G. Any companies also replaced traditional store-based
methods with multichannel marketing. As an example, on the other hand, the shop
is closed, but community shopping and social marketing are taking place.
Reported cases and fatalities are projected to be 48,786 and 1,482 on November 7, 2020.
Coronaviruses are the first known viruses that were detected in Wuhan, China. These
viruses mutate rapidly and have a great mutation risk of infection. The main problem
caused by coronavirus infection is severely acute respiratory illness. The number of
people infected by this lethal virus has increased from around 219, increasing the inci-
dents. The WHO announced that there is a pandemic of COVID-19 because it has spread
all over the globe. North America and Europe have both been identified as having nearly
21,000 cases of COVID-19, while Europe has 12,000 registered in secondary institutions.
The four nations with the most COVID cases on Earth up to November 7, 2020 are USA
(9,504,758), India (8,462,080), Russia (1,733,025), and France (encompassing the same
number of cases), and Russia (9,503,758). About 45,000 people have died in China due
to the disease and over 91,000 have been diagnosed with it [69]. And the extension of
deadly diseases from another lower middle class Asian nation with a population of
161.3 million [70], Bangladesh is caught in the tragedy. Bangladesh was the first to
recognize Creative Origins on March 8, 2020. Until the end of the date, the total of
COVID-19 deaths has been set at 410,988 and the number increases to 5,966.
4.10.1 Methodology
This chapter also included qualitative and quantitative analysis methods in some of
the research approaches used in it. Data from secondary sources, writers’ forecasts,
details from the WHO, and related academic institutions (IEDCR) were used.
Official data estimates have been used to calculate the population’s vulnerability.
Thus, other polls and tests have been used to confirm these results. Such personal
impressions of the world around the United States will be supplemented with data
collected from a wide range of different outlets and anecdotal knowledge from all
over the country. Primary research done in PubMed and Google Scholar was done
on COVID-19. Healthcare, religious, and economic data on Bangladesh was
gathered from March to November 7, 2020. Keywords were used in the quest, e.g.
“COVID-19,” “Coronavirus outbreak.”
4.10.6 Discussion
To address the widespread proliferation of COVID-19, GoB has implemented
several prevention and control measures. In Bangladesh, MOHFW set out
“Bangladesh Emergency Health Preparedness and Response Strategy for COVID-
19.” The foremost concern of this strategy is to slow the progression of the illness.
To further improve the ICU use, the hospital built a new facility, hired, and
equipped 2,000 more physicians and employed 5,000 nurses. The GoB initiative
guaranteed surgical devices, supplies, and health-related services for COVID-19
patients and providers. However, they do not have a higher number of patients
relative to their counterparts in other [nearby] countries No well-equipped COVID-
19 research laboratories and no clinical or health services for these citizens in the
country of Bangladesh make finding a cure difficult. This, thus, shall ensure that
the GoB has proper research facilities and health treatment for this lethal virus. For
many impoverished citizens in Bangladesh, it was a life-or-or-death decision to
take the long-malaria test. As opposed to staying sealed, more people choose to
consider the latter to ensure their long-term survival. The shutdown caused millions
of Americans to lose their jobs and wages.
It was hard enough and arrived too late. As a result, a lot of people who were
not well-off have been much worse off. Social distancing is difficult because of the
socioeconomic, environmental, and economic challenges that citizens face in
Impact of world pandemic “COVID-19” 85
4.11 Conclusions
At the COVID-19 pandemic’s advancing around the globe, the COVID-19 death toll
is accumulating. Aside from the increasing amount of cases and fatalities, also on the
whole, the epidemic had a dangerous side impact on economies: In addition to
decreasing their numbers, it increased their fragility. Following the announcement of
pandemic status in March 2020, bans on global travel were put in place, with citizens
encouraging anyone who could avoid being infected to avoid public areas by work-
place exposure and ensuring that they were at home at all times. In late March of
2006, Wuhan, China, was where the first cases of the disease were discovered. In
Brazil and the LMICs, a significant number of people in the United States, but also
the United States, India, and other higher income countries, were hit by the epidemic.
Unpreparedness was a big part of the problem of healthcare services. In some
instances, healthcare staff experienced a health-productivity gap. COVID-19 failures
also spurred healthcare institutions to find new necessities for patients. Telemedicine,
psychological distancing, masking, and handwashing have also been seen to lower
the burden of the COVID-19 pandemic.
Pandemic influence has more to do with other people than health services.
The community that resides in poverty and is heavily polluted with sewage and
86 The Internet of medical things
garbage is being hit harder by this toxin. Additional assistance, such as a benefit
payment for the poor, is needed. During epidemic outbreaks, it is highly critical
for this group to have a careful eye on this demographic since they are more
susceptible to disease. You become more susceptible during epidemics because of
SES and ethnic classification. Dietary patterns and low wages are correlated with
lower SES. Latinos are more disadvantaged than Caucasian populations owing to
personal and structural inequality, and as a result, they are more prone to suffer
from ill-health effects. Therefore, it is clear that the pandemic has passed them
by, perhaps never yet suspected. Richard Clarke Cabot, an American physician,
thought of both of these aspects while doing his daily work. On the basis of these
observations, he concluded that people with worse health were more likely to die.
This research supports the conclusion that poverty, disparities, and SDOH
increase the transmission rate of illness. While inequalities in health and disease
cause disparities in morbidity and mortality, inequities in healthcare exacerbate
the problem. Current research on epidemics has paid no attention to the impacts
of social inequality on well-being at periods of an epidemic, which means they
missed the point entirely, because they failed to do both. Because of this, it is
essential to respond to any health emergency promptly. To avoid an epidemic of
illness, it is essential to be aware of all the well-being and healthcare causes.
Identifying specific health and healthcare causes, such as SDOH, will aid in
providing access to a comprehensive set of services to the socioeconomically
vulnerable. Education of the public on the virus seriousness of the disease and
awareness-inducing nature of the illness is equally necessary. If you are more
aware of the community that is at risk than others, it may affect the decision to
avoid others that are less. Taking into consideration, the many causes will help
one to stop an epidemic from increasing. Appropriate and timely clinical treat-
ments, together with well-directed and timely schooling, have proved to be suc-
cessful countermeasures to the COVID virus. Reducing disparity between well-
being and healthcare may be achieved by using SDOH. Such projects need an
interdisciplinary approach that comprises doctors, anthropologists, public health
officers, and researchers, along with others from the National Institutes of Health,
the WHO, and others, to thoroughly understand various causes of health dis-
parities that groups of people face. We will have to determine how the last one
goes because we can deal with potential plagues.
Based on COVID-19 pandemic knowledge, a variety of hypotheses may be
reached. Asking the correct questions is critical in dealing with such a problematic
issue. Thus, rather than predicting the task’s difficulty, it is more prudent to assume
that it is essential. Second, more stringent policies and efficient compliance are
necessary if an epidemic is to deter or slow down its dissemination. Third, taking
note of the shortcomings of the health sector revealed by this experience, effective
steps must be taken to enhance public health. Proper plans, including improved
equipment of health facilities and staff training, are crucial in this respect. To avoid
the further transmission of COVID-19 infections, stringent controls on cross-
infections in hospitals should be taken.
Impact of world pandemic “COVID-19” 87
References
[1] He X, Lau EHY, Wu P, et al. Temporal dynamics in viral shedding and
transmissibility of COVID-19. Nat Med. 2020; 26: 672–5. https://doi.org/
10.1038/s41591-020-0869-5.
[2] Li Q, Guan X, Wu P, et al. Early transmission dynamics in Wuhan, China, of
novel coronavirus-infected pneumonia. N Engl J Med. 2020; 382: 1199–207.
https://doi.org/10.1056/NEJMoa2001316.
[3] Heron M. Deaths: leading causes for 2017. National Vital Statistics Reports,
2019; 68(6).
[4] Congressional Budget Office. CBO’s current projections of output,
employment, and interest rates and a preliminary look at federal deficits for
2020 and 2021. Congressional Budget Office (2020). https://www.cbo.gov/
publication/56335 (accessed October 25, 2020).
[5] Wu Z and McGoogan JM. Characteristics of and important lessons from the
coronavirus disease 2019 (COVID-19) outbreak in China: summary of a
report of 72314 cases from the Chinese center for disease control and pre-
vention. JAMA – J Am Med Assoc. 2020; 323: 1239–42. https://doi.org/
10.1001/jama.2020.2648.
[6] National Collaborating Centre for Determinants of Health. Marginalized
populations. National Collaborating Centre for Determinants of Health
(2020).
[7] Yang Song, Min Zhang, Ling Yin, et al. COVID-19 treatment: close to a cure?
A rapid review of pharmacotherapies for the novel coronavirus (SARS-CoV-2),
International Journal of Antimicrobial Agents, 2020; 56(2): 1–8.
[8] Bhuyan HK, Chakraborty C, and Shelke Y, Pani SK. COVID-19 diagnosis
system by deep learning approaches, Expert Systems, July 2021 (Early
Published).
[9] Czeisler M-E, Lane RI, Petrosky E, et al. Mental health, substance use, and
suicidal ideation during the COVID-19 pandemic—United States, June
24–30, 2020. MMWR Morb Mortal Wkly Rep. 2020; 69: 1049–57. https://
doi.org/10.15585/mmwr.mm6932a1.
[10] McGowan ML, Norris AH, and Bessett D. Care churn – why keeping clinic
doors open isn’t enough to ensure access to abortion. N Engl J Med. 2020;
383: 508–10. https://doi.org/10.1056/NEJMp2013466.
[11] Satiani B and Davis CA. Practice management the financial and employment
effects of coronavirus disease 2019 on physicians in the United States
(2020). doi:10.1016/j.jvs.2020.08.031.
[12] McMichael TM, Currie DW, Clark S, et al. Epidemiology of COVID-19 in a
long-term care facility in King County, Washington. N Engl J Med. 2020;
382: 2008–11. https://doi.org/10.1056/NEJMoa2005412.
[13] Sullivan-Marx E. Aging in America: how COVID-19 will change care,
coverage, and compassion. Nurs Outlook. 2020; 68: 533–5. https://doi.org/
10.1016/j.outlook.2020.08.013.
88 The Internet of medical things
[14] Normile D, Cohen J, Enserink M, et al. As normalcy returns, can China keep
COVID-19 at bay? Infected travelers pose a continuing threat, but local
coronavirus transmission still occurs as well. Science. 2020; 368: 18–9.
https://doi.org/10.1126/science.368.6486.18.
[15] Varghese GM and John R. COVID-19 in India: moving from containment to
mitigation. Indian J Med Res. 2020; 151: 136–9. https://doi.org/10.4103/
ijmr.IJMR_860_20.
[16] Bhuyan HK and Ravi VK. Analysis of sub-feature for classification in data
mining, IEEE Transaction on Engineering Management, 2021 (Early
Published).
[17] Statista. COVID-19 deaths per capita by Country. Statista (2020).
[18] Lee VJ, Chiew CJ, and Khong WX. Interrupting transmission of COVID-19:
lessons from containment efforts in Singapore. J Travel Med. 2020; 27: 1–5.
https://doi.org/10.1093/jtm/taaa039.
[19] Singapore’s economic contraction in the second quarter was worse than
initial estimates, August 2020, https://www.cnbc.com/2020/08/11/singapore-
releases-second-quarter-2020-gdp-economic-data.html.
[20] Bhuyan HK, Chakraborty C, Pani SK, and Ravi VK. Feature and sub-feature
selection for classification using correlation coefficient and fuzzy model.
IEEE Transaction on Engineering Management, May, 2021 (Early
Published).
[21] Daniel G, Stefania F, and Johannes W. COVID-19: Without Help, Low-
Income Developing Countries Risk a Lost Decade, August 27, 2020. https://
blogs.imf.org/2020/08/27/covid-19-without-help-low-income-developing-
countries-risk-a-lost-decade/
[22] Suleiman A, Bsisu I, Guzu H, et al. Preparedness of frontline doctors in
Jordan healthcare facilities to COVID-19 outbreak. Int J Environ Res Public
Health. 2020; 17. https://doi.org/10.3390/ijerph17093181.
[23] Ahmed SAKS, Ajisola M, Azeem K, et al. Impact of the societal response
to COVID-19 on access to healthcare for non-COVID-19 health issues in
slum communities of Bangladesh, Kenya, Nigeria and Pakistan: results
of pre-COVID and COVID-19 lockdown stakeholder engagements.
BMJ Glob Health. 2020; 5: e003042. https://doi.org/10.1136/bmjgh-2020-
003042.
[24] Martinez D, Sarria GJ, Wakefield D, et al. COVID’s impact on radiation
oncology: a Latin American survey study. Int J Radiat Oncol Biol Phys.
2020; 108: 374–8. https://doi.org/10.1016/j.ijrobp.2020.06.058.
[25] Luedi MM, Doll D, Boggs SD, et al. Successful personalities in anesthe-
siology and acute care medicine: are we selecting, training, and supporting
the best? Anesth Analg. 2017; 124: 359–61. https://doi.org/10.1213/
ANE.0000000000001714.
[26] Li Q, Guan X, Wu P, et al. Early transmission dynamics in Wuhan, China, of
novel coronavirus-infected pneumonia. N Engl J Med. 2020; 382: 1199–207.
doi:10.1056/NEJMoa2001316.
Impact of world pandemic “COVID-19” 89
[27] Rothan HA and Byrareddy SN. The epidemiology and pathogenesis of cor-
onavirus disease (COVID-19) outbreak. J Autoimmun. 2020; 109: 102433.
doi:10.1016/j.jaut.2020.102433.
[28] CDC. Preliminary estimates of the prevalence of selected underlying health
conditions among patients with coronavirus disease 2019—United States,
February 12–March 28, 2020. MMWR Morb Mortal Wkly Rep. 2020; 69:
382–6. doi:10.15585/mmwr.mm6913e2.
[29] AMA. Social Determinants of Health, Health Systems Science Learning
Series. Chicago, IL: AMA (2020).
[30] CDC. Coronavirus Disease 2019 (COVID-19): People Who Are at Higher
Risk for Severe Illness. Atlanta, GA: CDC (2020).
[31] Chapman AR. The social determinants of health, health equity, and human
rights. Health Hum Rights. 2010; 12: 17–30.
[32] Watson K. Why schools in rich areas get more funding than poor areas.
Global Citizen (2016). Available online at: https://www.globalcitizen.org/
en/content/cost-of-education-in-us/
[33] ODPHP. Social cohesion. In: Healthy People 2020. Washington, DC: U.S.
Department of Health and Human Services (2020).
[34] Kluch Avas. Americans’ social distancing beliefs and activity (May, 2020).
Available online at: https://news.gallup.com/opinion/gallup/311261/amer-
icans-social-distancing-beliefs-activity.aspx.
[35] ODPHP. Discrimination. In: Healthy People 2020. Washington, DC: U.S.
Department of Health and Human Services (2020).
[36] Kelli HM, Hammadah M, Ahmed H, et al. Association between living in
food deserts and cardiovascular risk. Circ Cardiovasc Qual Outcomes. 2017;
10: e003532. doi:10.1161/CIRCOUTCOMES.116.003532.
[37] Mackenbach JP, Cavelaars AE, Kunst AE, and Groenhof F. Socioeconomic
inequalities in cardiovascular disease mortality; an international study. Eur
Heart J. 2000; 21: 1141–51. doi:10.1053/euhj.1999.1990.
[38] Banerjee T, Crews DC, Wesson DE, et al. Food insecurity, CKD, and sub-
sequent ESRD in US adults. Am J Kidney Dis. 2017; 70: 38–47. doi:10.1053/
j.ajkd.2016.10.035.
[39] Osuna-Padilla IA, Leal-Escobar G, Garza-Garcia CA, and Rodriguez-
Castellanos FE. Dietary acid load: mechanisms and evidence of its health
repercussions. Nefrologia. 2019; 39: 343–54. doi:10.1016/j.nefroe.2019.
08.001.
[40] Abbasalizad Farhangi M, Nikniaz L, and Nikniaz Z. Higher dietary acid load
potentially increases serum triglyceride and obesity prevalence in adults: an
updated systematic review and meta-analysis. PLoS One. 2019; 14:
e0216547. doi: 10.1371/journal.pone.0216547.
[41] Cooksey-Stowers K, Schwartz MB, and Brownell KD. Food swamps predict
obesity rates better than food deserts in the United States. Int J Environ Res
Public Health. 2017; 14: 1366. doi:10.3390/ijerph14111366.
[42] NIMH. Obesity Hypoventilation Syndrome. Bethesda, MD: NIMH (2013).
90 The Internet of medical things
[43] CDC. Defining Adult Overweight and Obesity. Atlanta, GA: CDC
(2020).
[44] Petrilli CM, Jones S, Yang J, et al. Factors associated with hospitalization
and critical illness among 4,103 patients with COVID-19 disease in New
work City. medRxiv. 2020: 1–25. doi:10.1101/2020.04.08.20057794.
[45] Simonnet A, Chetboun M, Poissy J, et al. High prevalence of obesity in
severe acute respiratory Syndrome Coronavirus-2 (SARS-CoV-2) requiring
invasive mechanical ventilation. Obesity (Silver Spring). 2020; 28: 1195–9.
doi:10.1002/oby.22831.
[46] Firebaugh G and Acciai F. For blacks in America, the gap in neighbourhood
poverty has declined faster than segregation. Proc Natl Acad Sci USA. 2016;
113: 13372–7. doi:10.1073/pnas.1607220113.
[47] Nania R. Blacks, Hispanics Hit Harder by the Coronavirus, Early U.S. Data
Show. Washington, DC: AARP (2020).
[48] APM Research Lab. The Color of Coronavirus: COVID-19 Deaths by Race
and Ethnicity in the U.S. St. Paul, MN: APM Research Lab (2020).
[49] Kawachi I, Kennedy BP, Lochner K, and Prothrow-Stith D. Social capital,
income inequality, and mortality. Am J Public Health. 1997; 87: 1491–8.
doi:10.2105/AJPH.87.9.1491.
[50] Finch BK and Vega WA. Acculturation stress, social support, and self-rated
health among Latinos in California. J Immigr Health. 2003; 5: 109–17.
doi:10.1023/A:1023987717921.
[51] ODPHP. Access to foods that support healthy eating patterns. In: Healthy
People 2020. Rockville, MD: U.S. Department of Health and Human
Services (2020).
[52] Wu X, Rachek C, Sabath BM, Braun D, and Dominici F. Exposure to air
pollution and COVID-19 mortality in the United States: a nationwide cross-
sectional study. medRxiv. 2020: 1–36. doi:10.1101/2020.04.05.20054502.
[53] ODPHP. High school graduation. In: Healthy People 2020. Rockville, MD:
U.S. Department of Health and Human Services (2020).
[54] Shafa Z. New effort aims to provide Covid-19 resources to non-English
speakers in U.S. STAT News (2020).
[55] Lopez MH, Rainie L, and Budiman A. Financial and health impacts of
COVID-19 vary widely by race and ethnicity (2020, May 5).
[56] iSPARC. Coronavirus economic stimulus payments: who gets it, how, & impact
on other benefits. Psychiatry Inf Brief. 2020; 17: 1–4. doi: 10.7191/pib.1148.
[57] Abrams EJ, Myer L, Rosenfield A, and El-Sadr WM. Prevention of mother-
to-child transmission services as a gateway to family-based human immu-
nodeficiency virus care and treatment in resource-limited settings: rationale
and international experiences. Am J Obstet Gynecol. 2007; 197(3 Suppl):
S101–6. doi: 10.1016/j.ajog.2007.03.068.
[58] Kloppenburg K. Governor Little issues stay-at-home order for Idaho (2020,
March 25). Available online at: https://www.krtv.com/news/coronavirus/
stay-at-homeorder-issued-in-idaho.
Impact of world pandemic “COVID-19” 91
bangladesh/news/detail/08-10-2020-increased-testing-capacityessential-step-
in-fighting-covid-19 (accessed November 3, 2020).
[74] Bangladesh sees 100th death of doctors from Covid-19. Dhaka Tribune (2020).
https://www.dhakatribune.com/health/coronavirus/2020/10/15/bangladesh-sees-
100th-death-of-doctors-from-covid-19.
[75] Islam R. The impact of COVID-19 on employment in Bangladesh: Pathway
to an inclusive and sustainable recovery (2020).
[76] Sen B. Poverty in the time of Corona: Short-term effects of economic
slowdown and policy responses through social protection (2020).
[77] Sultan M, Hossain MS, Islam MS, Chowdhury K, Naim J, and Huq F.
COVID-19 impact on RMG sector and the financial stimulus package: Trade
union responses (2020).
[78] Ferguson NM, Laydon D, Nedjati-Gilani G, et al. Impact of non-
pharmaceutical interventions (NPIs) to reduce COVID-19 mortality and
healthcare demand (2020). https://doi.org/10.25561/77482.
[79] World Population Prospects 2019, https://population.un.org/wpp/.
[80] Reuters. Bangladesh shuts down villages after tens of thousands attend cle-
ric’s funeral. New York Times (2020).
[81] Bradbury-Jones C and Isham L. The pandemic paradox: the consequences of
COVID-19 on domestic violence. J Clin Nurs. 2020; 19: 1–3. https://doi.org/
10.1111/jocn.15296.
[82] Islam A. COVID-19 Lockdown Increases Domestic Violence in Bangladesh.
DW (2020) https://www.dw.com/en/covid-19-lockdown-increases-domes-
tic-violence-inbangladesh/a-53411507.
Chapter 5
Artificial intelligence in healthcare
Deepa Joshi1 and Anikait Sabharwal1
1
School of Computer Science, University of Petroleum and Energy Studies (UPES), Dehradun, India
94 The Internet of medical things
gathered from are discussed. Second, we talk about the moral and lawful difficulties
of AI-driven medical care. Next, the structured and unstructured types of healthcare
data have been analyzed followed by the AI techniques applied to these types of data,
such as ML, natural language processing (NLP), and DL. We then analyze the crucial
disease areas such as cardiology, radiology, neurology, and cancer, where the health
issues can be alleviated by employing AI techniques. In conclusion, we further dis-
cuss the areas where AI-based techniques are applied in real life.
5.1 Introduction
The goal of AI is to teach, while also passing on human intelligence and instincts to
a computer. An amalgamation of mathematics, statistics, and science makes this
task attainable.
Famous AI strategies comprise ML techniques for structured data, for example,
neural network (NN), the traditional support vector machine (SVM), and advanced
DL, in addition to NLP that is utilized for the unstructured data. Cardiology, cancer,
and neurology are some of the principal disease areas that employ AI tools. Instead of
teaching, training and programming systems for certain tasks, ML emphasizes on
passing on the abilities concurrently. ML designs algorithms through which systems
can create and work on predictions for homogeneous situations and tasks. Under the
approach termed supervised learning, the machines are given data and are trained on
examples that have similar desired output so they can further work on unseen pro-
blems and data. Under unsupervised learning, just the input data is given to the
machine, and it is required to build an output without help and oversight [1]. From
the early days itself, ML algorithms were devised and were utilized to analyze
medical datasets. Today, ML is an instrumental base for providing key tools to astute
data analysis. Particularly over the most recent couple of years, the evolution of
technology gave moderately reasonable and accessible ways to gather and store data.
Current medical clinics are exceptionally provisioned with observing and other data
accumulation tools. Furthermore, the data is assembled and apportioned in huge data
frameworks. ML is right now at such an innovative stage that it is equipped to
analyze clinical data, and specifically there is a huge load of work done in clinical
analysis in compact dedicated diagnostic problems [2].
ML is being employed on varied kinds of diseases like diabetes and cancer,
falling under healthcare analysis. Cancer being one of the most fatal diseases has
distinct kinds that comprise breast cancer, stomach cancer, lung cancer, prostate
cancer, and the list goes on. Nearly 12% instances of cellular breakdown in the lungs
come every year, where 10% cases do not make it through. Likewise, for breast
cancer, 11% cases arise, of which 9% do not make it. For dealing with cancer in the
medical field analysis, it is crucial that accurate and prime quality data is produced.
In such an ambition-driven world, medical services should utilize the data in a par-
ticular way that consistently results in an ascent in the quality of medical care and a
significant decrease in the expense required for the therapy [3]. With time, especially
in recent years, there is an exponential rise in medical care research with ML. The
Artificial intelligence in healthcare 95
difficulty human beings face in making decisions and inferring data is due to the
assortment of medical data that involves omics data, clinical data, or even EHR data.
For that very reason, ML has been effectively presented and advanced in medical
care for finer comprehension of data and superior dynamic cycle [4].
The chapter has been presented in 5-folds. Initially, the distinct data sources
from where the medical care data can be accumulated are examined. Second, we
discuss the moral and legal hurdles of AI-driven healthcare. Furthermore, the
structured and unstructured kinds of medical services data have been investigated,
which is trailed by the AI strategies applied to these kinds of data, for instance, ML,
DL, and NLP. We at that point dissect the absolutely critical disease areas like
cardiology, radiology, nervous system science, and cancer, where the medical
problems can be lightened by utilizing and efficaciously employing AI methods. In
conclusion, we examine the fields where AI-based methods are applied in reality.
Examination
Medical Laboratory
imaging results
regulators and the ones who process data. This led to legitimate commitments and
liabilities. The United Kingdom is known to have instrumental mechanisms like the
Information Commissioner’s Office (answerable for authorizing as well as putting
Data Protection Act into motion) and Health Research Authority (liable for a
government foundation diving into health research) and Confidentiality Advisory
Group (a strategy for classified health data in if explicit assent is unavailable),
which were not consulted prior to the rise of data transfer, subsequently done to
employ “self-assessment information governance toolkit” utilized to support the
security of particularly specialized technical framework to deal with NHS data [7]. In
such cases, it is crucial that the care-providers are cautious when they share the data
with an outsider which is in an implicit relationship with the separate patient(s),
respectively. The care that is provided straight to the concerned authority or person is
regarded as “action concerned about the protection, examination and treatment of
disease and the mitigation of enduring of a recognized individual.” This requires a
notice to be provided to the patient or subject of concern. In case the explicit assent
and notice are not granted, then all the data, irrespective of being tagged or untagged,
falls into the public domain and is required to be produced by a statutory body.
Through this, a tab can be kept regarding the unlawful proprietary manipulation of the
information and accordingly the data processor can only exchange the data up to a
certain limit. The freely accessible public or peer-examined data sets, for example,
Messidor, will help unconstrained development of processes and algorithms, since
without assent such de-identified data sets ought to be viewed as a community
resource. The idea of putting them in the hands of the community is to employ and
enforce policing of the data exchanges at a level that cannot be attained by the gov-
ernment. In India, we need to have legal regulations, for example, segment 251 in the
United Kingdom, which calls the government and legal control for such exchanges
[7]. As a matter of fact, the discussion of ownership and custodial responsibility of
such information in our country seldom takes place [8]. There is no particular brief
regarding what way the information will be handled by the machine in the event that
we utilize an AI device on such sets. The effect of transfer learning and AGI is such
that humans often are boggled as to how the data is handled by the machine. People
apparently found certain lines of play in AlphaZero “completely alien” but highly
efficacious. Notable progress has been seen in Justice BN Srikrishna Committee, as
they strategically empowered their patients [8]. In any case, it is reckless to anticipate
that general protection should address immensely tangled bioethical worries in the
expansion and growth of clinical AI. The healthcare profession is required to demand
explicit customs and “Dos and Don’ts,” which whenever stuck to will guard it against
prosecution and litigation in the event of data break since India is not just the colossal
producer and the least expensive source of such data (as per confirmation made by the
authors in the Google Diabetic Retinopathy Project) [9], it will pose to be a massive
market specifically for the algorithms that are derived from it in future. The machines
require rules and regulations for effective recursive self-improvement, which calls for
a dire need of strengthened bioethical and computational ethics framework to guar-
antee optimum and safe functioning of the same. The phrase, “First, do no harm,”
should be strongly conveyed from clinical morals into the discipline of computational
98 The Internet of medical things
bioethics. We possess the characteristics of a typical parent in this case; past a phase
of development, we may not be in charge of these algorithms any further or may not
comprehend them by any stretch of our imagination.
(a type of learning in which the machine trains without the help of any labeled
data). With time, progress, having such an eminent computational force as its
backbone, AI is becoming supreme. Although considering this, it raises a major
concern about a situation with this force discovering its way into unacceptable and
dangerous hands, be it artificial or human. The earlier one evolves at the pace of
human advancement, giving an open door to evolve and reclaim our lives, however,
not so for the artificial elements, as we can see with the AlphaGo experience. This
calls for deeper caution in the fields of medicine and medical research since the
individual influenced by every choice is a conscious human being. AI cannot be
considered one single technology, but rather an assembly of them, and many of
them have a direct, yet distinct, relationship with the field of healthcare.
even the patient’s medical outcomes that comprise quantitative disease levels,
patient’s survival times, and disease indicators, like, size of tumor, are gathered for
the purpose of clinical research [12]. The subsequent features are preferred to
accelerate ML system’s effectiveness in resolving diagnostic tasks in healthcare:
great execution, the capacity to suitably manage noisy data (errors) and with
missing data, the clarity of knowledge of diagnosis, the capacity to elaborate con-
clusions, and the capacity of the algorithm to reduce the quantity of examinations
that are important to acquire reliable diagnosis [2]. ML algorithms may be sepa-
rated in three significant classes: supervised learning, semi-supervised learning,
and unsupervised learning. The one that is immensely notable for feature extraction
is unsupervised learning, while supervised learning is reasonable for prescient
modeling through progressively developing a few connections between the
patient’s characteristics (input) in addition to the interest’s outcome (output). The
crossbreed between supervised learning and unsupervised learning has been
effectively put forward as semi-supervised learning that is appropriate for situations
where the outcome is missing for particular subjects [12]. Figure 5.2 shows the
ubiquity of distinct supervised learning strategies in medical implementations.
AI applications in healthcare typically employ supervised learning for a ple-
thora of reasons and purposes. Unsupervised learning may be utilized as a feature
of the step of preprocessing to lessen dimensionality or recognize subgroups, which
therefore makes the subsequent supervised learning’s step way more effective.
Important strategies incorporate decision tree, random forest, naı̈ve Bayes, dis-
criminant analysis, NN, logistic regression, linear regression, nearest neighbor, and
SVM [13].
variables. The methodology of NNs in tackling such issues is not quite the same as
typical approaches. It takes the countless contributions of many such manually
written digits named as training samples and builds a framework to naturally derive
rules for distinguishing the digits by persistently learning patterns from those
training samples. Moreover, the precision of such models can be further refined by
appending additional training samples [13]. “Perceptron” is the prime belief that
drives the building of an NN. The concept of a perceptron was encouraged by the
human nervous system, where each network consists of distinct interconnected
neurons that effectively communicate with one another.
Over the period of time, healthcare has become an instrumental field where AI
techniques, including DL, are extensively employed and applied. DL has the ability
to inspect more mind-boggling nonlinear patterns in the data. With the evolving
technology, there is a grave rise in the amount and complexity of data, which calls
for the new fame of DL. An algorithm that has its base on DL functions in the
subsequent manner: input layer is utilized as a gateway for the input data to enter
the deep NN. Figure 5.3 mentions the sources of data for DL.
The design of the input data gets changed so the deep NN realizes the way to
handle and process the same, and potentially certain extra calculations are then
performed. At that point, the outcomes are provided to the initial hidden layer. A
network can contain a variety of secret layers, implying that this footstep carrying
out a calculation and providing the result to the following hidden layer is further
carried out on various occasions, until and unless the yield layer has been reached.
The final layer’s role is essential to offer a solution to the question that DL is
expected to answer.
For instance, in Figure 5.4, a prostate magnetic resonance (MR) picture is the
input that goes via the network with every one of its layers, and thus the output
layer conveys the response to the doubt “Is there a lesion in the prostate MR
image?”
Diagnostic imaging
Electrodiagnosis
Genetic diagnosis
Clinical laboratory
Mass screening
Others
Figure 5.3 The sources of data for deep learning. The data is created by looking
through deep learning in an amalgamation with the diagnosis methods
on PubMed [38]
Artificial intelligence in healthcare 103
Figure 5.4 A deep neural network utilizes various layers to protect and handle
input data (a prostate MR picture, in this case). Once the data goes via
the network, a solution is given by the algorithm to a particular
question, like, “Is there a lesion in the prostate MR image?” [38]
clinical data is in the structure of narrative text, like physical inspection and
assessment, clinical research center and lab reports, and employable notes that are
unstructured while possessing no limits for the PC program. With regard to this
unique circumstance, NLP focuses on separating and hence gathering helpful data
through narrative text in order to help medical decision-making. Two essential
components can be found in an NLP pipeline, namely (1) processing of text and (2)
subsequent classification. It recognizes a progression of catchphrases and keywords
related to diseases in the clinical notes dependent on the verifiable historical
databases through text processing [14]. Following that, a proportion of the key-
words are picked by analyzing their impacts on the categorization of the abnormal
and normal cases. At that point, the approved keywords advance the structured data
to help clinical dynamics of making decisions. The development of the NLP
pipelines is with a goal of aiding clinical making of decisions to notify arrange-
ments related to treatment, overseeing unfavorable effects, and much more. For
instance, Fiszman et al. presented that bringing NLP into the picture for perusing
the chest X-beam reports provides aid to antibiotic associate frameworks to caution
doctors for the conceivable requirement against infective therapy [15]. On a similar
note, NLP was employed by Miller et al. to oversee the lab-based antagonistic
effects on its own [16]. NLP pipelines are also instrumental in assisting diagnosis of
diseases. For example, Castro et al. recognized 14 disease factors related to cere-
bral aneurysms by carrying out NLP on the medical and clinical notes [17].
AI, which incorporates the domains of NLP, robotics, and ML, may be imple-
mented in practically any domain in healthcare, and its possible commitments to
medical education, clinical research, and applications of medical services appear to
be boundless. Flaunting its hearty capacity to coordinate, learn, and integrate from
104 The Internet of medical things
huge datasets of the medical industry, AI has the potential to serve parts in analysis,
diagnosis [18] customized medication [19], and decisions made in the healthcare
industry [20]. For instance, diagnostic calculations and applications based on AI
applied to mammograms are aiding the identification of breast cancer, filling in the
second judgment cum opinion for radiologists [21]. Likewise, progressed virtual
symbols are equipped to take part in significant conversations, having inferences
for the diagnosis and thus, treatment of diseases in psychiatry [22]. AI imple-
mentations additionally reach out into the physical sphere with mobile manip-
ulators, physical task support systems, and robotic automated prostheses, and
aiding the conveyance of telemedicine.
Cancer, neurology, and cardiology are few of the key disease areas that employ
AI tools [12]. Somashekhar et al. proposed that the IBM Watson for oncology must
be a top-notch AI system for aiding the diagnosis and analysis of cancer via a two-
staged approval study [23]. Esteva et al. carefully performed an analysis on clinical
pictures to recognize subtypes of skin malignant growth [24]. Bouton et al.
developed an AI structure to restore and subsequently reenforce the control of
development in patients with quadriplegia [25]. Farina et al. effectively tried the
power of an offline machine or man-made interface that uses the issued timings of
spinal motor neurons to control the prostheses of the upper limb [26]. Dilsizian and
Siegel inspected the normal utilization of the AI system to analyze and examine the
anticipated utilization of the AI system heart disease via a careful look at the car-
diovascular pictures [19].
Large health companies are now often merging, which allows for pronounced
health data accessibility [35]. Greater health data helps in laying the fundamental
groundwork for the implementation of AI algorithms. A huge portion of indus-
tries in the healthcare sector focus on the implementation of AI in the support
systems of clinical decision-making. ML algorithms adjust and take into con-
sideration more powerful feedback followed by the solutions as the collection of
data occurs [36]. Big data is becoming the talk of the town amongst a number of
companies to explore the possibilities of its incorporation in the healthcare
industry. The organizations look into the market opening via the domains of “data
storage, analysis, assessment, and management technologies,” the most important
aspects of the medical industry. Following are listed certain key examples of big
shot companies that have been instrumental in contributing to AI algorithms for
use in healthcare.
IBM’s Watson Oncology is building at the Memorial Sloan Kettering Cancer
Center and Cleveland Clinic. At the same time, IBM is actively functioning with
CVS Health on applications in AI in the treatment of chronic diseases, along with
Johnson on the analysis of scientific papers in order to locate associations for
advancement in development of drugs. In May 2017, Rensselaer Polytechnic
Institute and IBM started a project that they were jointly working on, granted
Artificial intelligence in healthcare 107
SoftBank Robotics (Pepper) are also actively working toward building similar robots.
Far reaching tech companies such as Google, Amazon, Baidu, and Apple, all have
dedicated research divisions in AI along with having smaller AI-based companies
acquired worth millions of dollars since the field of AI is constantly evolving.
Numerous automobile manufacturers are thinking in the same line and are beginning
to employ ML medical care in their cars.
References
[1] D. Joshi, S. Anwarul and V. Mishra. Deep learning using Keras. In Machine
Learning and Deep Learning in Real-Time Applications; 2020 (pp. 33–60).
IGI Global. doi:10.4018/978-1-7998-3095-5.ch002.
[2] I. Kononenko. Machine learning for medical diagnosis: history, state of
the art and perspective. Artificial Intelligence in Medicine. 2001; 23(1):
89–109.
[3] K. Raheja, A. Dubey and R. Chawda. Data analysis and its importance in
health care. International Journal of Computer Trends and Technology.
2017; 48(4): 176–180.
[4] A. Dhillon and A. Singh. Machine learning in healthcare data analysis: a
survey. Journal of Biology and Today’s World. 2019; 8(6): 1–10.
[5] D. Joshi and T.P. Singh. A survey of fracture detection techniques in bone
x-ray images. Artificial Intelligence Review. 2020; 53(6): 4475–4517.
[6] J. Bali, R. Garg and R.T. Bali. Artificial intelligence (AI) in healthcare and
biomedical research: why a strong computational/AI bioethics framework is
required? Indian Journal of Ophthalmology. 2019; 67(1): 3.
[7] J. Powles and H. Hodson. Google DeepMind and healthcare in an age of
algorithms. Health and Technology. 2017; 7(4): 351–367.
[8] V. Khullar. White Paper on Data Protection Framework for India [Internet].
Prsindia.org; 2018. [cited 2018 August 25]. Available from: http://www.
prsindia.org/uploads/media/Report.
[9] J. Krause, V. Gulshan, E. Rahimy, et al. Grader variability and the impor-
tance of reference standards for evaluating machine learning models for
diabetic retinopathy. Ophthalmology. 2018; 125(8): 1264–1272.
[10] HIT Infrastructure. Unstructured Healthcare Data Needs Advanced Machine
Learning Tools; 2018. [online] Available at: https://hitinfrastructure.com/
news/unstructured-healthcare-data-needs-advanced-machine-learning-tools
[Accessed 23 March 2021].
[11] HealthSnap, Inc. Challenges and Opportunities of Structured and Unstructured
Health Data | HealthSnap INC; 2018. [online] Available at: https://healthsnap.
io/structured-unstructured-health-data/ [Accessed 23 March 2021].
[12] F. Jiang, Y. Jiang, H. Zhi, et al. Artificial intelligence in healthcare: past,
present and future. Stroke and Vascular Neurology. 2017; 2(4).
[13] I. Goodfellow, Y. Bengio and A. Courville. Deep learning. First Edition.
Cambridge, MA: The MIT Press; 2016.
Artificial intelligence in healthcare 109
www.forbes.com/sites/bernardmarr/2018/07/27/how-is-ai-used-in-healthcare-
5-powerful-real-world-examples-that-show-the-latest-advances/#610224b95dfb.
[Accessed 25 March 2021].
[29] B. Kalis, M. Collier and R. Fu. 10 pPromising AI Applications in Health
Care. Harvard Business Review 2018. Available at: https://hbr.org/2018/05/
10-promising-ai-applications-in-health-care. [Accessed January 10, 2021].
[30] S. Singhal and S. Carlton. The era of exponential Improvement in health-
care? McKinsey & Company; 2019.
[31] J.K. Kulski. Next-generation sequencing –an overview of the history, tools,
and "omic" applications. In: Next Generation Sequencing-Advances,
Applications and Challenges. London: Intech; 2016.
[32] J.P. Hughes, S. Rees, S.B. Kalindjian and K.L. Philpott. Principles of early
drug discovery. British Journal of Pharmacology. 2011; 162(6): 1239–1249.
[33] L. Zhang, J. Tan, D. Han and H. Zhu. From machine learning to deep
learning: progress in machine intelligence for rational drug discovery. Drug
Discovery Today. 2017; 22(11): 1680–1685.
[34] M. Rangaiah. Artificial Intelligence in Healthcare: Applications and Threats
| Analytics Steps; 2020. [online] Analyticssteps.com. Available at: https://
www.analyticssteps.com/blogs/artificial-intelligence-healthcare-applica-
tions-and-threats [Accessed 25 March 2021].
[35] A.N. Pisarchik, V.A. Maksimenko and A.E. Hramov. From novel technol-
ogy to novel applications: Comment on “An integrated brain-machine
interface platform with thousands of channels” by Elon Musk and Neuralink.
Journal of Medical Internet Research. 2019; 21(10): e16356.
[36] G. Litjens, T. Kooi, B.E. Bejnordi, et al. A survey on deep learning in
medical image analysis. Medical Image Analysis. 2017; 42, 60–88.
doi:10.1016/j.media.2017.07.005.
[37] Scnsoft.com. The State of the Art in Health Data Analytics; 2017. [online]
Available at: https://www.scnsoft.com/blog/health-data-analytics-overview
[Accessed 26 March 2021].
[38] V. Fortunati. Deep Learning Radiology: The Secret of Convolutional Neural
Networks; 2020. [online] Quantib.com. Available at: https://www.quantib.
com/blog/convolutional-neural-networks-in-deep-learning-radiology
[Accessed 25 March 2021].
[39] B. Kalis, M. Collier and R. Fu. 10 Promising AI Applications in Health
Care; 2018. Retrieved 5 April 2021, from https://hbr.org/2018/05/10-pro-
mising-ai-applications-in-health-care.
Chapter 6
Blockchain in IoT healthcare: case study
Kripa Elsa Saji1, Nisha Aniyan1, Renisha P. Salim1,
Pramod Mathew Jacob1, Shyno Sara Sam1, Kiran Victor1
and Remya Prasannan1
Health sector is one of the most important industries in the world. With the recent
advent in the number of diseases, health is becoming the primary concern for each
individual. But with their busy schedules, they do not find enough time to act on
this concern. In this chapter, a health monitoring system has been proposed to sense
the vital body parameters and to store and share them in a secure environment. The
vital body parameters are sensed from the patient using sensors. The sensed data is
then stored into blockchain. The stored data can be accessed by the patient and
doctors using a web application. When a doctor needs to access the information
about a patient from another doctor, they only need to transfer the patient ID
between themselves to access the data.
6.1 Overview
Due to the recent advent in the number and kind of diseases, health has become the
primary concern for all humans. With the changing climate, sedentary lifestyle, and
other factors affecting the environment, diseases are finding a way to get in people’s
life stealthily. So, it is necessary to have a system that will be able to help people out
of this conflict. It is necessary to have a system that will help people out of this
predicament.
IoT or Internet of Things is a system of sensors and software used in the
collection and transfer data over a wireless network without any human commu-
nication. The sensors are used to collect data or information from an environment.
The sensed data are stored in a database via Wi-Fi, Bluetooth, Ethernet, etc. This
data is processed and depending on it, needed steps are taken.
Blockchain is a distributed ledger platform used to store transactions (blocks) in a
database (chain). The information stored in the blockchain is immutable. Blockchain
1
Department of Computer Science and Engineering, Providence College of Engineering, Chengannur,
APJ Abdul Kalam Technological University, India
112 The Internet of medical things
employs a peer-to-peer network that allows the transactions to imply at most two
parties (sender and receiver). It is a decentralized network so there is no chance to
interact with this directly through a third party. The data is securely stored crypto-
graphically. These technologies are getting implemented in many sectors for provid-
ing a secure environment, data protection, sensing data easily, etc. In this proposed
model, these technologies were incorporated to sense, monitor, store, and share health
parameters.
In the existing system, patients’ data is stored in electronic health record
(EHR) and electronic medical records systems. These systems are used to share the
medical information of the patient with health-care providers and organizations.
The various challenges faced by the user in this scenario are listed next:
● The users do not have access to their records stored in EHRs.
● When consulting another doctor, there is an overhead of paper-based medical
records.
● Paper-based records do not provide security, confidentiality, and integrity.
● The medical records in EHRs are not secure from third-party intrusion.
The proposed system addresses these challenges using IoT, blockchain, and a
web application. The model aims to provide secure storage along with monitoring
and sharing of vital body parameters.
There are various EHR systems available based on various technologies. Most of
them focus on providing a secure and efficient way of medical information storage
using blockchain or cloud-based architectures. Here we discuss the summary of
some relevant systems available for securing IoT healthcare.
Kulkarni and Bakal [1] proposed a system to sense the body condition of
humans and then sent it to the medical experts. This medical team reviewed the
critical parameters that have been sensed and took further necessary action.
Kalaiselvi [2] proposed a system that is a network of connected Internet-
enabled devices that communicate with each other to take care of health using
remote sensors. This system does work without the assistance of human-to-human
or human-to-machine contact, this system can transmit data over a network. This
information is sent to the mobile applications.
Yeotkar and Gaikwad [3] proposed a model that is a remote monitoring tech-
nique which the person can be monitored wirelessly in his location using wearable
sensor arrangement by using accelerometer, temperature sensor, pulse oximeter and
heart-rate sensor, and galvanic skin response sensor to monitor physiological
human body.
Ved et al. [4] proposed a cloud-based personal health record system aiming for
an authorized user to access and review the patient’s records from any location. It is
a web application that works on the J2EE platform, deployed on Amazon EC2 and
runs on Microsoft SQL Server 2005. The user interface is divided into role-based
Blockchain in IoT healthcare: case study 113
modules and for message passing and maintenance. The database layer includes the
DCM4CHE server and SQL server. For establishing security, password-protected
access is provided to all the users.
Joshi et al. [5] proposed a model for the health records to be accessed in a
delegated and secured manner using attribute-based encryption (ABE) for the man-
agement of access and data security in cloud-based EHRs. The EHR manager
application is a web application that is developed, using Python, based on the fun-
damentals of the model-view-controller architecture and is open-source. The system
consists of access broker (for authentication), EHR ontology (stores encrypted data),
ABE, and cloud service provider modules (for hosting EHR ontology).
Yang et al. [6] proposed a system to solve security problems and improve
storage efficiency. Cloud storage and blockchain technology are used to guarantee
the storage efficiency of electronic health records. ABE is used to assure the fine-
grained access control of the data in EHR which is part of security. The experi-
mental results showed the time for encryption and search to be independent to the
number of attributes.
Adlam and Haskins [7] proposed blockchain technological know-how has been
evolving and ought to perchance enhance the modern-day EHRs infrastructure. It is
suggested that EHRs often employ a role-based get entry to manipulate the model.
Permissioned blockchain technological know-how may want to enhance the
authorization model by influencing clever contracts and attribute-based get entry to
control. Hyperledger Fabric has been recognized as a permissioned blockchain
technological know-how ideal closer to use-cases that require privacy.
Mahore et al. [8] proposed a model based on cloud and blockchain which pro-
vides privacy and health-care data for statistical analysis to the researchers at the
same time. The model uses permissioned blockchain where the keys, certificates, and
all the participating entities are managed by the membership service provider. When
a patient visits a hospital, a cloud-based record is created and the metadata is stored
on the blockchain from which the researchers can obtain data. The patient has
complete control over his records, divided into sensitive and nonsensitive, and has
two key pairs for each. The sensitive data holding the information that can disclose
the patient’s identity is encrypted using public key, while the nonsensitive data that
holds the diagnosis data is encrypted using public key and stored on the cloud.
Ahmed et al. [9] proposed a distributed digital network system that acts as a
public ledger and uses bitcoin to operate money services, security services, IoT,
and various Internet applications. This system keeps the secure history of the
complete data for certain things. It has been explained as the shedding light of
blockchain, some of the typical algorithms used in the various blockchain, and
scalability issues.
Tanwar et al. [10] proposed a technology to strengthen virtualization by
implementing personalized healthcare in real-time using blockchain technology. The
blockchain can improve healthcare by using various solutions for eliminating the
prevailing limitations in health-care systems which includes tools and frameworks to
determine the effectiveness of the system such as Docker container, composer,
Hyperledger Fabric, Wireshark capture engine, and Hyperledger Caliper.
114 The Internet of medical things
Consult
Consult
Doctor 2 Doctor 1
MAX30100. The sensed parameters are read by NodeMCU. NodeMCU has inbuilt
Wi-Fi with the help of which this data will be sent to the blockchain.
Patient
Blockchain
Node MCU
Sensed data
Retr
ata
ieve
ed
riev
data
Ret
Consult
Sensors
Consult
Doctor 2 Doctor 1
MAX30100
IOTA
DS18B20 NodeMCU
AD8232
with other entities using the web application. The doctors can also consult another
doctor by only sharing the patient ID rather than the whole medical record.
The block diagram of sensors layer is shown in Figure 6.3. The sensors used in
the system for sensing the vital body parameters are MAX30100 (for pulse and
saturation), DS18B20 (for temperature), and AD8232 (for ECG). The micro-
controller unit NodeMCU is used to read the data that is sensed using the sensors.
The data is then stored into the blockchain platform, IOTA.
In Figure 6.4, the use case diagram demonstrates the various use cases of the
health monitoring system. The end users of the system will be users and doctors.
The user can view the sensed data, share the data with other users (doctors), and
view their updated data through the web application. The doctor can modify the
118 The Internet of medical things
Consult a doctor
medical records of the patient. The doctor can also refer another doctor by sharing
the patient ID, used for retrieving the patient’s medical records.
and to communicate with MAX30100 using I2C protocol, respectively. The pulse
oximeter and heart rate sensor are shown in Figure 6.5 (left).
MAX30100 sensor
AD8232 sensor
Figure 6.5 Pulse oximeter and heart rate sensor (left) and ECG sensor (right)
Temperature sensor
NodeMCU
6.3.2.4 NodeMCU
NodeMCU is an open-source IoT firmware and development kit that is designed to
develop IoT-based applications. It consists of multiple digital and one analog pin. It
also has an inbuilt Wi-Fi that enables to communicate the detected data with the
Internet. NodeMCU is shown in Figure 6.6 (right).
The circuit diagram of the system is represented in Figure 6.7. All the sensors are
integrated together to the central coordinator which is NodeMCU in this scenario.
Hardware connections of the sensor module are given next:
1. Connect VIN pin of MAX30100 to the 3.3V pin of NodeMCU.
2. Connect INT pin of MAX30100 to the D0 pin of NodeMCU.
3. Connect SCL pin of MAX30100 to the D1 pin of NodeMCU.
4. Connect SDA pin of MAX30100 to the D2 pin of NodeMCU.
5. Connect GND pin of MAX30100 to the GND pin of NodeMCU.
6. Connect 3.3V pin of AD8232 to the 3v3 pin of NodeMCU.
7. Connect GND pin of AD8232 to the GND pin of NodeMCU.
8. Connect OUTPUT pin of AD8232 to the A0 pin of NodeMCU.
9. Connect LO pin of AD8232 to the D5 pin of NodeMCU.
10. Connect LOþ pin of AD8232 to the D6 pin of NodeMCU.
11. Connect the left pin of DS18B20 to the GND pin of NodeMCU.
12. Connect the middle pin of DS18B20 to the 3v3 pin of NodeMCU.
13. Connect the right pin of DS18B20 to the D4 pin of NodeMCU.
6.3.2.5 IOTA
IOTA is an open, scalable distributed ledger that allows feeless transactions
designed for an IoT ecosystem. It is based on tangle, a DAG. In tangle, every
transaction points to two preceding transactions that are called as parents and get
validated by the child transaction. Tangle grows in the direction of the paths that
have the highest weights and the lightest branches never reach consensus.
Figure 6.8 depicts the transactions in a tangle.
In IOTA, all the nodes issue and validate transactions, i.e., a node validates
previous two transactions and then issues its own transaction. IOTA is asynchro-
nous. A transaction takes different amount of time to reach different nodes and
another node may be seeing a previous version of the tangle. The nodes always
share new transactions with each other to increase their probability of getting
verified and eventually reaching the consensus.
s
m x
j u
g p
b
z
c k n v
i
a
d f
y
o t
e l
h
r
D
be calculated from the root and search for masked messages (obtained using root
and a sideKey). Figure 6.9 depicts the data stored in the public mode of MAM.
6.3.2.7 MongoDB
MongoDB is a document-oriented NoSQL database that is suitable for unstructured
data. The data store is lightweight, easy to traverse, and allows fast execution.
MongoDB Atlas is used to store the MongoDB database that is being used with the
system. MongoDB Atlas is a cloud provider that lets users host their databases and
related APIs. Integrating it into the system allows it to scale exponentially and also
increases the availability of the system.
6.3.2.8 NodeJS
NodeJS is an open-source server platform based on JavaScript. It facilitates the
creation of scalable and fast network applications. It provides a runtime environ-
ment and various JavaScript modules for the development of web applications.
NodeJS is asynchronous (non-blocking) resulting in scalable and fast execution
of code.
6.3.2.9 Express.js
Express.js is a web application framework based on NodeJS which provides robust
APIs to build web apps and websites. Many modules are available in npm which
can be plugged into Express making it flexible and pluggable. It helps to manage
server and the various routes associated with it.
1st generation
0-th generation
(genesis) To next generation
<title>My secret
2</title>
<title>My secret 1</title> <div>Furthermore,
nextRoot nextRoot
<div>To tell the truth... ...
Decrypt
Given info.
Decrypt
Address Address
the data to work with the other components in the workflow. The data is sent to the
NodeJS server because the NodeMCU does not have enough power to directly
write the data into the IOTA.
Express.js is used at the NodeJS server to further enhance the networking
capabilities of the NodeJS server and also enable routing for the server. Express
allows the processing of the incoming of the POST requests and also helps with the
generation of the response and redirection of the data to other parts of the archi-
tecture. The extracted data is correlated with the base database hosted on the
MongoDB Atlas, and if a match is found, the sensed data is written into the IOTA
Devnet using the MAM package.
When the patient data is published for the first time, the root node will not be
present in the MongoDB. In this case, the root is stored in the MongoDB and after
storing the data, the latest state will be stored into a JSON file. The latest state will
be an object that consists of the seed and the configurations of the state. The latest
state is used to connect the transactions in the tangle. This means that all the data of
a patient will be connected using the latest state. Figure 6.10 shows the data blocks
stored in IOTA tangle using MAM.
callback function that was received. This process is done until the last block. Then
the second callback function is used to display all this data. These data trytes are then
converted to ASCII and then to JSON. Express.js is used to render this JSON data as
a table. A user who is identified as a patient can only retrieve their own data. When a
patient tries to retrieve the data by providing a patient ID, the requested ID and the
patient ID are verified. If a match is found, the data is retrieved. Since this application
is ran on development parameters, the public mode of MAM is used for better sta-
bility and predictability. Figure 6.11 represents fetching data from MAM explorer.
6.4 Conclusions
The proposed health monitoring and management system consists of a hardware
module, blockchain module, and web application. The sensors incorporated with
the NodeMCU detect the health data from a patient’s body which can be used for
determining their health conditions. The data is stored into an IOTA node with the
help of a server and MAM. The patient ID is used to retrieve the stored data using
the Web application. A patient can only retrieve his/her own data. After retrieving
the data, the patient can consult with the doctor or a doctor can refer another doctor.
When a doctor consults another doctor, the patient ID is shared using which the
consulted doctor can access the patient’s data. The advantages of the proposed
system are as follows:
● Easier to track and analyze the vital parameters.
● Users can access their information and can use it to consult with a doctor.
● The overhead of paper-based records is eliminated.
● The process of consulting another doctor is made easier.
● More economical and optimum utilization of resources.
Blockchain in IoT healthcare: case study 125
The future plans are to incorporate a device with more processing power in the
hardware to eliminate the use of server and publish the data directly to IOTA
tangle. Our next stage will try to publish the data blocks in the restricted mode of
MAM and provide access control on the data stored on IOTA tangle.
References
[1] N. J. Kulkarni and J. W. Bakal, “Real Time Vital Body Parameter Monitoring,”
in Fourth International Conference on Computing Communication Control
and Automation (ICCUBEA), 2018.
[2] G. Kalaiselvi, “A Comprehensive Study On Healthcare Applications using
IoT, “International Journal of Engineering Science Invention (IJESI),
pp. 41–45, 2018.
[3] H. S. Yeotkar and V. T. Gaikwad, “IoT Based Human Body Parameters
Monitoring by Using Wearable Wireless Sensor Network,” International
Research Journal of Engineering and Technology (IRJET), vol. 06, no. 07,
pp. 2458–2466, 2019.
[4] V. Ved, V. Tyagi, A. Agarwal, and A. S. Pandya, “Personal Health Record
System and Integration Techniques With Various Electronic Medical Record
Systems,” in IEEE 13th International Symposium on High-Assurance
Systems Engineering, Boca Raton, 2011.
[5] M. Joshi, K. Joshi, and T. Finin, “Attribute Based Encryption for Secure
Access to Cloud Based EHR Systems,” in IEEE 11th International
Conference on Cloud Computing, 2018.
[6] X. Yang, T. Li, R. Liu, and M. Wang, “Blockchain-Based Secure and
Searchable EHR Sharing Scheme,” in 2019 4th International Conference on
Mechanical, Control and Computer Engineering (ICMCCE), Hohhot, China,
2019.
[7] R. Adlam and B. Haskins, “A Permissioned Blockchain Approach to the
Authorization Process in Electronic Health Records,” in 2019 International
Multidisciplinary Information Technology and Engineering Conference
(IMITEC), Vanderbijlpark, South Africa, 2019.
[8] V. Mahore, P. Aggarwal, N. Andola, Raghav, and S. Venkatesan, “Secure
and Privacy Focused Electronic Health Record Management System using
Permissioned Blockchain,” in 2019 IEEE Conference on Information and
Communication Technology, Allahabad, India, 2019.
[9] I. Ahmed, Shilpi, and M. Amjad, “Blockchain Technology A Literature
Survey,” International Research Journal of Engineering and Technology
(IRJET), vol. 5, no. 10, 2018.
[10] S. Tanwara, K. Parekha, and R. Evans, “Blockchain-Based Electronic
Healthcare Record System for Healthcare Applications,” Journal of
Information Security and Applications, 2019.
[11] S. Ramachandran, O. Obu Kiruthika, A. Ramasamy, R. Vanaja, and S.
Mukherjee, “A Review on Blockchain-Based Strategies for Management of
126 The Internet of medical things
7.1 Introduction
Medical imaging has made significant advances in a variety of health-care uses in
recent years, including research, diagnosis, and education. Various imaging mod-
alities such as computed tomography (CT), magnetic resonance imaging (T1-MRI
and T2-MRI), positron emission tomography (PET), and single-photon emission-
computed tomography (SPECT) are used to provide more clinical data to medical
practitioners, which reflects different information about the human body [1,2].
However, due to technical limitations, a single modality image cannot give enough
information to meet specialists’ clinical needs. Combining information from different
modalities (CT-MRI, PET-MRI, SPECT-MRI, etc.) into a single image, which better
identifies the region of interest and enhances diagnostic accuracy, is an effective way
to overcome this problem. Despite the fact that there are numerous algorithms to
conduct this operation [3,4], dictionary-based sparse representation (SR) techniques
have grown in popularity in the field of image fusion. The first is a fixed dictionary
that has been built using analytical models such as the discrete cosine transform
(DCT), wavelet, etc. The second is a learned dictionary made from high-quality
natural images. In image fusion, Yang and Li [5] were the first to introduce sparse
representation. For multi-focus image fusion, they used a redundant DCT dictionary.
Aharon et al. [6], on the other hand, demonstrated that an adaptive dictionary learned
from input visuals beats a fixed dictionary. As a result, Yu et al. [7] suggest a fusion
1
Department of Electronics and Communication Engineering, Amrita School of Engineering, Amrita
Vishwa Vidyapeetham, Chennai, India
2
Department of Electronics and Communication Engineering, R.M.D. Engineering College
Kavaraipettai, Chennai, India
3
Department of Computer Science and Engineering, Velammal Institute of Technology Panchetti,
Chennai, India
128 The Internet of medical things
strategy based on a joint sparsity model. The dictionary is trained by simply adding
all of the source image patches, which takes a long time and renders the dictionary
ineffective. For simultaneous fusion and denoising of multimodal pictures, Yin and
Li [8] used a joint sparsity model. The experimental findings were adequate, although
the proposed method’s performance is significantly reliant on external data. For
sparse coefficient estimation, Yang and Li [9] introduced the simultaneous ortho-
gonal matching pursuit algorithm (OMP). The proposed method is tested on a variety
of image types, and the dictionary used in this method is very sensitive on the input
image. Wang et al. [10] combined non-subsampled contourlet transform (NSCT)
with SR to perform image fusion. A major benefit of this strategy is that it produces
comparable experimental findings with less computing time.
From the focus information map of multi-focus images, Nejati et al. [11]
suggested a global dictionary learning approach. The dictionary input training data
is created by randomly selecting an arbitrary number of patches, which may or may
not result in a highly organized dictionary. Liu et al. [12] used multi-scale trans-
forms (MSTs) and SR to offer an image fusion framework. For the dictionary
learning process, a collection of 40 high-quality natural pictures is used. The same
author later suggested an adaptive sparse representation (ASR) model for image
fusion and denoising at the same time. The histogram of gradients is used to learn
six compact sub-dictionaries. The disadvantage of this approach is that it has a
higher computational cost. Kim et al. [13] suggested a dictionary learning strategy
based on K-means clustering and principal component analysis to lower the com-
putational cost of SR-based fusion approaches. The learnt dictionary is compact
and informative, but with this method, the number of clusters must be preset.
As a result, existing SR-based fusion solutions either share a preconstructed
vocabulary that requires previous knowledge of the external pre-collected data or uti-
lize a learnt dictionary that requires prior knowledge of the external pre-collected data.
The global dictionaries provided by Aishwarya and Bennila Thangammal [14] have
been shown to have greater representation ability than existing SR-based approaches;
nonetheless, there are still several key difficulties that need to be addressed. To begin,
previous knowledge of the external data sets is required for the creation of the initial
training samples. In reality, collecting appropriate training samples may not be
achievable. Furthermore, the fusion performance is highly dependent on the original
training data set’s construction. Second, global dictionaries are built using the inherent
structures of high-resolution images. As a result, dictionaries of this sort are more
suited to single-sensor image fusion, such as multi-focus image fusion. Learning a
dictionary adaptive to the relevant source images, on the other hand, can be more
efficient in representing the complex, intrinsic structures of multi-sensor or multimodal
fusion applications. As a result of these techniques, the need for a discriminative over-
complete dictionary that reflects the complicated structures of given images has arisen.
The design of the dictionary, activity-level measurement of sparse coefficients,
and the design of the optimum fusion rule are the significant elements that deter-
mine the ultimate fusion quality of SR-based techniques. Most SR-based fusion
algorithms have focused on the former element so far, while the latter part has not
been adequately addressed. Due to the aforementioned difficulties, this chapter
Adaptive dictionary-based fusion 129
The main goal of this work is to create a comprehensive and compact over-
complete dictionary in order to improve fusion performance. An informative
sampling strategy based on MSF is proposed to generate such a dictionary.
Figure 7.1 depicts the detailed dictionary learning process. Assume the source
images IA and IB are registered with size N N. To begin, each source image
IA and IB is divided into image patches of size n n using the sliding window
T
approach. Let the patches be denoted by P ¼ ptc t¼1 ; c 2 fA; Bg. Generally, only
a significant percentage of medical image patches contain clinically valuable
information. Using all of the image patches from the source images makes the
dictionary very redundant and has a significant impact on the fusion result. The
computing time is also increased as a result of this learning process. In order to
address the aforementioned difficulties, only informative patches with higher
structural information are chosen for dictionary learning.
First, any patch ptc with an intensity variance of less than five ðvarðptc Þ < 5Þ
will be eliminated. This preprocessing step removes all the meaningless patches of
source images. The source image patches obtained after the preprocessing phase
N1 N2
are denoted by PA ¼ p jA j¼1 and PB ¼ p kB k¼1 . Training a dictionary with
informative patches is widely known for allowing richer data representation than
typical preconstructed dictionaries. The structural information of patches is
Informative sampling
Construct
training data set
K-SVD
Adaptive dictionary
weighed using MSF to select such informative patches. MSF [15] is a useful metric
for determining an image’s overall edge strength in gradient directions. The MSF of
an image I with size M1 M2 is calculated as follows:
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
MSF ¼ RF 2 þ CF 2 þ SDF 2 þ MDF 2 (7.1)
where RF and CF are the row and the column frequency which is given by
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u
u 1 X M1 X M2
RF ¼ t ½Iðx; yÞ Iðx; y 1Þ2 (7.2)
M1 M2 x¼1 y¼2
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u
u 1 X M1 X M2
CF ¼ t ½Iðx; yÞ Iðx 1; yÞ2 (7.3)
M1 M2 x¼2 y¼1
and frequencies along the main diagonal and secondary diagonal are given by
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u
u 1 X M1 M X 2 1
The threshold value is set low enough that both low- and high-frequency pat-
ches are selected for training with minimal redundancy. For the dictionary learning
process, let TD; fTDl jl ¼ 1; 2; 3; :::; Lg indicate the training data set. To deter-
mine whether the patch contains enough spatial features, a selection rule is used.
This is accomplished by comparing MSFA and MSFB of each patch in PA and PB
with Tc . For example, if MSFAj > TA ; select p jA . Otherwise ignore p jA :
Select p jA ; if MSFAj > TA
TDl ¼ (7.7)
Ignore p jA ; otherwise:
According to the previous criteria, the patch with fine details and superior
visual clarity is used in the dictionary learning process. This procedure is repeated
until all of the useful patches in PA and PB have been retrieved. The mean value of
each informative patch in the training data set is subtracted from zero before the
dictionary learning process to guarantee that only the edge structures of the patches
Adaptive dictionary-based fusion 131
are included for training. Finally, the K-SVD method iterates between the sparse
coding step and the dictionary update step D to learn an adaptive dictionary D.
Sparse representation
Coefficient fusion and reconstruction
Fused
vectors
IB
Fused image
The fusion coefficients are formed by selecting the sparse vector having
maximum activity level both in spatial domain and transform domain. The fused
Adaptive dictionary-based fusion 133
1 2 3
4 5
Figure 7.3 Medical image data sets: 1–3—CT and MRI images; 4 and 5—MR-T1
and MR-T2 images
134 The Internet of medical things
artifacts. The NSCT method’s fusion result in Figure 7.4(c) loses a lot of image
features and has a contrast reduction problem. Figure 7.4(d) shows the blocking
artifacts in the fusion outcome of multilevel local extrema. The outputs of GFF and
MST-SR fusion are more visible; however, they do not maintain certain local
information from the original images. Figure 7.4 shows that the fusion results of the
ASR method and the MIFMSF approach cannot be separated easily. As illustrated
in Figure 7.5, a particular portion of source pictures is expanded. Figure 7.5(c)
shows that the NSCT approach has lost the edge and texture information of the
original images, resulting in low contrast. The multilevel local extrema method has
Figure 7.4 Fusion results of CT and MRI image pair: (a) and (b) source images;
fusion results of (c) NSCT, (d) multilevel local extrema, (e) GFF, (f)
MST-SR, (g) ASR-128, (h) ASR-256, and (i) MIFMSF method
Figure 7.5 The magnified details of fused image: (a) and (b) regions of source
images. Fusion details of (c) NSCT, (d) multilevel local extrema, (e)
GFF, (f) MST-SR, (g) ASR-128, (h) ASR-256, and (i) proposed method
Adaptive dictionary-based fusion 135
Table 7.1 Average statistical evaluation of five standard medical image pairs for
different methodologies
substantial artifacts that blur the fused image’s edge region (Figure 7.5(d)).
Artificial traces (see the eye regions in Figure 7.5(e)–(h)) are introduced in the
fused image since the GFF, MST-SR, and ASR algorithms fail to capture the ori-
ginal information of source images.
When compared to the input images, the MIFMSF technique produces output
with greater sharpness, clarity, and information content. Furthermore, the proposed
MIFMSF approach is free of undesired degradations and blocking artifacts, making
it ideal for precise clinical diagnosis.
7.6 Summary
The flaws of existing SR-based fusion schemes are investigated, and a dis-
criminative dictionary with less computational effort is constructed without any
prior knowledge about the external data sets. From the experimental analyses, the
discriminative dictionary is constructed with approximately 32% of the total source
image patches. The new fusion rule modifications ensure that all of the critical
spatial information of the source images is well retained in the final fused image,
resulting in the best visual appearance. As compared to state-of-the-art SR-based
136 The Internet of medical things
References
[1] V. Barra and J. Y. Boire, A general framework for the fusion of anatomical
and functional medical images, NeuroImage 13 (2001), 410–424.
[2] S. Li, X. Kang and J. Hu, Image fusion with guided filtering, IEEE Trans.
Image Process. 22 (2013) 2864–2875.
[3] A. P. James and B. Dasarathy, Medical image fusion: a survey of the state of
the art, Inf. Fusion 19 (2014), 14–19.
[4] R. Singh, R. Srivastava, O. Prakash and A. Khare, Multimodal medical
image fusion in dual tree complex wavelet transform domain using max-
imum and average fusion rules, J. Med. Imaging Health Inf. 2 (2012),
168–173.
[5] B. Yang and S. Li, Multifocus image fusion and restoration with sparse
representation, IEEE Trans. Instrum. Meas. 59 (2010), 884–892.
[6] M. Aharon, M. Elad and A. Bruckstein, K-SVD: an algorithm for designing
overcomplete dictionaries for sparse representation, IEEE Trans. Signal
Process. 54 (2006), 4311–4322.
[7] N. Yu, T. Qiu, F. Bi and A. Wang, Image features extraction and fusion
based on joint sparse representation, IEEE J. Sel. Top. Signal Process. 5
(2011), 1074–1082.
[8] H. Yin and S. Li, Multimodal image fusion with joint sparsity model, Opt.
Eng. 50 (2011), 067007.
[9] B. Yang and S. Li, Pixel-level image fusion with simultaneous orthogonal
matching pursuit, Inf. Fusion 13 (2012), 10–19.
[10] J. Wang, J. Peng, X. Feng, G. He, J. Wu and K. Yan, Image fusion with
nonsubsampled contourlet transform and sparse representation, J. Electron.
Imaging 22 (2013), 043019, 1–15.
[11] M. Nejati, S. Samavi and S. Shirani, Multi-focus image fusion using
dictionary-based sparse representation, Inf. Fusion 25 (2015), 72–84.
[12] H. Liu, Y. Liu and F. Sun, Robust exemplar extraction using structured
sparse coding, IEEE Trans. Neural Netw. Learn. Syst. 26 (2015), 1816–1821.
[13] M. Kim, D. K. Han and H. Ko, Joint patch clustering-based dictionary
learning for multimodal image fusion, Inf. Fusion 27 (2016), 198–214.
[14] N. Aishwarya and C. Bennila Thangammal, An image fusion framework
using morphology and sparse representation, Multimedia Tools Appl. 77
(2017), 9719–9736.
[15] Y. Zheng, E. A. Essock, B. C. Hansen and A. M. Haun, A new metric based
on extended spatial frequency and its application to DWT based fusion
algorithms, Inf. Fusion 8 (2008), 177–192.
Adaptive dictionary-based fusion 137
[16] Q. Zhang and B. Guo, Multi-focus image fusion using the nonsubsampled
contourlet transform, Signal Process. 89 (2009), 1334–1346.
[17] Z. Hu, Medical image fusion using multi-level local extrema, Inf. Fusion 19
(2013), 38–48.
Chapter 8
Artificial intelligence for sustainable e-Health
Shrikaant Kulkarni1
8.1 Introduction
This work is aimed at eliciting how an MNC organizational model can replace the
UK National Health Services (UK-NHS) government model, in situations wherein a
public-sector model is difficult to develop. Why and how the MNC model can offer
solutions of its own kind to a viable and well-articulated EHR system. It is prescribed
that the quality assurance of HC and efficiency with which it is accessed to EHRs can
be enhanced if right solutions are derived to the interoperability issue among insti-
tutions, organizations, and computer systems. HC is in fact an area where rather than
efficiency effectiveness is of paramount importance, the efficiency with which HERs
are accessed is key because of the sensitivity of medical decision-making to time [1].
1
Adjunct Professor, Faculty of Science & Technology, Vishwakarma University, Pune, India
140 The Internet of medical things
HC is a service industry that expects abysmally low margin of error as against other
services [4]. In HC, an error would prove fatal, and cannot be reversed, similar to
an error from airline pilot, although redundant systems are inbuilt in an aircraft. In
fact duplicate systems in HC may not be encouraged due to cost effectiveness
problems, most important are accuracy, the framework of information architectures
within data flow highways to assure quality of data, and technology, constancy in
training and semantics for those feeding the data, information, and decisions. To
attain such goals, it is indispensable that an organizational structure that is an
umbrella for subsidiary units such as hospitals and HC units is required to coordi-
nate and acquire efficiencies in health records entry and providing in an accurate
and timely fashion. In the absence of such a supra organization, the subsidiary units
will stay discrete, or fragmented, leading to critical interoperability problems and
time lags in patient HC. The absence of suitable communication systems and
information sharing among HC peers and with patients is also considered as a
major challenge and is responsible for causing one in five medical errors due to
access to inadequate information which is a barrier in effective decision-making.
Another implication of fragmentation of computer systems is the duplication of
procedures, which enhances costs, apart from discomfort to patients [5]. Earlier
research shows that EPRs will check the medical errors to a large extent. However,
Artificial intelligence for sustainable e-Health 141
studies have suggested that the efforts to employ EHR systems slow down proce-
dures as the interface is user unfriendly, although the interface was done with care
keeping in view professional users in perspective, and who evaluated it prior to use
on a mass scale. There is still no solution in sight that takes care of the optimum
utilization of latest information and communication technology (ICT) input/output
systems [6].
EHRs- and ICT-associated problems have come to the fore. American Senator
Hillary Clinton said “We have the highly sophisticated medical system in the
world, still patient safety is compromised because of medical errors, duplication,
and lack of efficiencies. Tapping the full potential of information technology will
check errors and better quality in our health system.” It is also found that electronic
medical records should be favored over chapter-based ones [7].
● data quality, its availability and acceptability across nation by patients, medical
practitioners, and health workers is a must for EHR development; and
● one more major bottleneck in EHR application is that medical professionals are
not able to perform their jobs due to significant data entry needs to populate EHRs.
and facilities; however, only the insurance companies, the medical practitioners or
the facility, and the patients are the repositories of their records and data. If an
individual leaves an insurance company but not the doctor, then the doctor can
make use of that EPR; however, on changing the doctor and if some other insurance
company is involved, then there are a host of protocols to be followed to get
records, which frequently are converted in terms of chapter records.
This happens mainly due to legal and privacy issues, and since there are non-
automated or computerized systems that communicate with one another for seamless
data transfer in between two doctors or facilities that are related to two different
insurance firms. The interoperability issue in such a situation can be addressed only
by proper coordination; however, another factor that often becomes a barrier is the
legal system that governs the entities. The legal issues can vary from reluctance to
share proprietary information, to leaking of information for either use or misuse of
the EPR by the legal system, the lawyers of the patient, or of another company. Thus,
fragmentation occurs not merely to the legal environment and the nonavailability of
linked data architectures but also because of an organizational structure, either public
or private, that can carry out all transactions with such a legal entity as like UK-NHS,
and alike entities in other nations, but a best option is a supra organization in the
NGO area or private sector would be encouraged preferably [14].
well as limitations. There is a lot to learn from such supra organization, works using
legislation with huge funds to spend. However, this model may be worth replicating
in the public sector, a private-sector organization could come up with a coherent
HC system, with or without the efficiency level attained by the UK-NHS. However,
the model cannot be implanted in all countries because of variations in financial
and legislative constraints prevailing in different countries.
The local and national NHS’s IT systems require further upkeep or replace-
ment of incumbent IT systems to
● integrate;
● implement innovative national systems;
● integrate NHS CRS with concerned HC-based products and services, like E-
prescription, which enhances patient care by minimizing errors in prescrip-
tions, redundancy of data, employee time, and cost; and
● provide right infrastructure (N3) to enable the NHS with smart network ser-
vices and higher broadband connections to better patient care protocols by
fetching patient care records from anywhere and anytime, thereby cutting
down HC costs by offering patient care remotely and not only saving time but
also by providing patient care expeditiously [17].
An EHR system tries to accomplish what the industries have already achieved. HC
is complex and, thus, asks for a customizable model to fulfill its own and patient’s
needs. Moreover, lack or inadequate information in EHR prevents clinicians in
making quality decisions. Hence, more efforts are to be taken for coordinating input
and output. Therefore, an organizational model and structure fits into an environ-
ment where interoperability issues can be addressed when faced having two or
more complex systems [24]. The UK-NHS is such a kind of organizational model
that tries to get over interoperability problems via its writ of legislation and the law,
which can enacted as it is a government organization. However, in spite of a
healthy legal and political environment, there are host of interoperability problems
146 The Internet of medical things
ML can create new opportunities for sustainable HC and the investigators can focus
on automation in the following areas:
● analysis and prediction of COVID-19 cases;
● discovery of COVID-19 patients;
● current consolidated portal to support pandemics in the future;
● setting up of an innovative Pilot COVID-19 Data Warehouse for reference in
future;
● upgraded tools for collecting, preparing, and storing data thoroughly; and
● platform to support for building up a community of medical professionals for
pandemic situations.
Knowledge
Machine learning (ML) -based
Systems
Artificial neural
networks (ANN)
Figure 8.1 Relation between AI, machine learning (ML), artificial neural network
(ANN), deep learning (DL), and knowledge-based systems (KBS)
8.18.1 Diagnosis
The diseases are made out depending upon potentially noisy data. The clinicians
should be able to correlate the symptoms to the right disease(s). Such a task
involves reasoning with inadequate and inexact data, defective sensors, etc.
Diagnosis is normally followed by a treatment plan, which involves determination
of a sequence of actions to treat the disease [33].
8.18.2 Interpretation
The collected data are analyzed to interpret. The data at times are unreliable,
erroneous, or extraneous. Therefore, the system should have capability to remove
outliers in the information.
8.18.3 Monitoring
Signals are constantly interpreted and alarms are put into effect as and when
needed.
8.18.4 Control
Signals or data are interpreted, and the system is regulated, depending upon
abnormalities in intended response [34].
addressed by breaking them into number of smaller ones. The designer must have
ability to handle the relation between these smaller tasks appropriately [35].
8.19 Conclusion
The AI systems first of its kind were knowledge-enabled DSSs, using isolated static
datasets. Such systems on connecting to electronic health data, AI became fruitful
for teaching, though not prepared for clinical care. DL via multilayered connected
ANNs has potential clinical applications. Google teams and New York University
observed that DL models can improve accuracy in lung cancer diagnosis. Although
our expectations of AI in clinical medicine and in diabetes are heightened, there
have been incremental developments showcasing the applicability of AI.
AI can be useful in developing synthetic pancreas (closed-loop systems) and
in comprehending the relation between social parameters and physiological bio-
markers of health. Diabetes too produces a significant amount of retinal, renal,
the taxonomy generated for diabetes with the help of root and rule-based method.
End-stage renal disease, vascular, and other data are monitored over the time.
Such pathophysiological data with other data will give new insights into diabetes
management. An expert opinions and care can be accessed to patients from any
part of the globe. A diabetes patient will have an access eye care from anywhere
through the application of smart phone images. Glucose control algorithms can be
individualized for the patients depending upon glucose consumption, insulin
pharmacology, nutria-genomics, and exercise patterns. It is clear that we are on
the verge of revolutionizing HC, with the ability to extract, analyze, the data for a
timely and quality clinical advice. AI will play a key role in bringing about a
paradigm shift in HC. It demands rapid progress in knowledge representation and
reasoning.
The UK-NHS model has proved that it can manage HC of 6 crore people, by
overcoming the interoperability issues and the countries that follow such public-
sector-governed HC system can adopt the model, if they possess strong political
and economic will power to follow. Countries, constitutions, legal and political
systems, financial condition, etc. of which if are not suitable in employing a UK-
NHS, or similar kind of model, can opt for an alternative in the form of MNC
model irrespective of whether it is developed as a nonprofit NGO or a for-profit
organization. Government sector organization gets the support from government
and its legislators to pass laws that enables the working of an HC, EHR, and EPR
system where interoperability issues are required to be made out and are to be
overcome by the legislature. In private-sector corporation, the interoperability
issues can be overcome by the generation of an internalized market under the aegis
of an NGO or a corporation.
An MNC organizational model was designed in order to get over various inter-
operability issues among countries. Therefore, an MNC model, having own inter-
nalized market to control, is a fit case in overcoming EHR interoperability issues,
integrating the internally related IS architectures, upgrading them, and training the
152 The Internet of medical things
References
[1] H. Almubarak, R. Stanley, W. Stoecker, and R. Moss, Fuzzy color clustering
for melanoma diagnosis in dermoscopy images. Information 8(89) (2017).
[2] A. Angulo, Gene selection for microarray cancer data classification by a
novel rule-based algorithm. Information 9, 6 (2018).
[3] Avon Health Authority, Electronic Patient Records Electronic Health Records,
Avon Health Authority, Schofield, J, Bristol, Bristol 15, 389–406 (2000).
[4] A.R. Bakker, The need to know the history of the use of digital patient data,
in particular the EHR. Int. J. Med. Inf. 76, 438–441 (2007).
[5] C.A. Bartlett and S. Ghoshal, Managing Across Borders: The Transnational
Solution, Harvard Business School Press, Boston, MA (1989).
[6] B. Baumann, Polarization sensitive optical coherence tomography: a review
of technology and applications. Appl. Sci. 7, 474 (2017).
[7] A. Begoyan, An overview of interoperability standards for electronic health
records. In Society for Design and Process Science. 10th World Conference
on Integrated Design and Process Technology; IDPT-2007. Antalya,
Turkey, June 3–8 2007.
[8] K. Chung, R. Boutaba, and S. Hariri, Knowledge based decision support
system. Inf. Technol. Manage. 17, 1–3 (2016).
[9] Commission on Systemic Interoperability, Ending the Document Game, U.S
Government Official Edition Notice, Washington, DC (2005).
[10] R.C. Deo, Machine learning in medicine. Circulation 132, 1920–1930
(2015).
[11] J. Dunning, Multinational Enterprises and the Global Economy, Addison-
Wesley, Wokingham (pp. 3–4) (1992).
Artificial intelligence for sustainable e-Health 153
1
Department of Computer Science and Engineering, Siksha ‘O’ Anusandhan (Deemed to be University),
Bhubaneswar, India
2
K L Deemed to be University, Vaddeswaram, Andhra Pradesh, India
156 The Internet of medical things
effectively and in efficiently early stages. With this program, the reconstruction and
treatment of BC would be more successful. Controller-assisted devices transmit all
sensor values as packets to the mobile app via Bluetooth. We predicted and alerted
the shift in patient status with the aid of the machine learning algorithm.
9.1 Introduction
BC has been the world’s utmost severe and wide-ranging syndrome, with 2 million
new cases occurring during 2018, according to the American Research Institute [1].
By improving clinical outcomes, and increasing cost-effectively patient care, the
health-care sector contributes to two key objectives. In order to increase clinical
results, real-time measurement and treatment of diseases are considered key fea-
tures. Patient care needs to be improved to improve patients’ quality of life and user
experience. IoT offers a cost-efficient way of achieving the two objectives for the
medical devices health industry. Connected health programs use scarce resources to
enhance health quality, which results in better clinical results. Measurable asso-
ciated advantages. Medical devices include reduced death rates, less visits to hos-
pitals, and less medical devices.
Hospital and ambulance admissions include reduced bedding and hospital stay.
Remote patient control results inefficient and prompt diagnosis and leads to
improved health-care management [2]. In addition, patients (and their relatives)
may become more visible and actively involved, by making them more visible in
their current health conditions.
BC cells in BC tissue are growing abnormally and the rate of cell diminution,
resulting in BC, is slowly increased. Typically, BC is a malignant cancer in the
breast. Most dividing cells are tissues known as cancer (malignant) and non-
cancerous (benign). Tumors are a piece of or build of tissue. In advanced medical
research development nations, 5-year primary BC survival rate is 80%–90%,
though secondary BC survival rate falls to 24% [3]. Specific invoice-based methods
were used to diagnose BC. Breast tissues are collected for examination in the
biopsy technique [4] and the result is highly accurate. Nevertheless, the patient has
discomfort in taking a biopsy from the breast. Mammograms [5] are also used to
treat BC and are used for BC diagnosis. Yet diagnosis of benign cancer is not easily
rendered by the mammogram technique. Magnetic reasoning imaging [6], a very
complex method that offers excellent outcomes with three-dimensional (3D) ima-
ges and shows interactive features, is also an invoice-based technique for diag-
nosing the breast. Such methods for invoice-based diagnostics are highly difficult
to implement and the tests do not diagnose BC efficiently and reliably.
Wearable technology offers continuous surveillance of BC women, unlike
traditional invasive mechanisms. Basically, two solutions exist: traditional methods
invasive in nature and one is the noninvasive method more frequently used. The use
of IoT on devices like iTBra simplified monitoring for BC even further. Such IT
devices monitor changes in a woman’s metabolic profile overtime and distinguish
between healthy and potentially cancerous profiles. The use of metabolic detection
An innovative IoT-based breast cancer monitoring system 157
you may want to see the doctor. The purpose of Royea was to make this established
technology a portable product that could be sold as a wearable tool, so that millions
of women could easily wear it and build an overall analytical database for early
detection. This is modified by the iTBra. “Rob and his company have been able to
attach sensors on the breast in search of cell-level changes in cancer and don’t have
to be seen by a doctor with his eye,” explained Kramer.
This structure has their own advantages, but do not have an early symptom
identification solution in terms of real-time predictions. The aim of this chapter is,
therefore, to overcome this restriction in order to create a complete system that
measures all BC characteristics necessary. In addition, the patient may be alerted
before the disease worsens, thus decreasing the women beings mortality rate, by
using the prediction algorithm on the real-time data. This inactive alternative is a
good example to constantly track and avoid patient complications when combined
with the real-time information allows prediction by the SMO classification with
SVM-RFE FS and allows recommending suitable remedies.
Azar and El-Said [5] have suggested a BC screening technique. They
employed three classification methods: radial basis function kernel (RBF), prob-
abilistic neural network (PNN), and multilayer perceptron, respectively. The results
of the classifiers were evaluated by applying performance assessment methods,
including precision, specificity, and sensitivity. Precisely 97.80% and 97.66% of
training and assessment classifications, respectively, were obtained and the accu-
racy was achieved. In addition to genetics, AliăźKović and Subasi proposed the two
Wisconsin brain cancer datasets [9].
Using the fuzzy-genetic algorithm system that hits 97.36% accuracy, Akay [10]
suggested using the F-score system for BC screening and endorsing the vector
machine (SVM) and had strong output results. The K-means Zheng et al. [11]
algorithm used for the aggregation and extraction properties and for the classification
of brain and malignant breast tumors in combination with SVM. In classification and
low processing time, the technology proposed achieved great precision [12]. Used in
conjunction with various classifiers for Ramadevi Hybridized Primary Component
Analysis (PCA) used in many datasets on BC and realized fair precision. Now [13]
the writer suggested a Pareto-based mimetic strategy, the Naive Bayes Classifiers
(NBC) Detection Network. The experimental results have shown that the technology
proposed has achieved reasonable grading precision and very low machine time.
For feature extraction, the K-means and SVM were used for classification.
Intelligent BC detection technique was developed by Onan [14]. For example, the
floppy-through was used to pick and select functions based on consistency. In BC
screening, he recycled the nearest neighbor algorithm. The technology for the opti-
mization of swarm particles and nonparametric estimates for the kernel density of the
BC density was developed by Sheikhpour et al. [15]. Rasti et al. [16] have developed
a diagnostic technique of BC using a combination of convolutionary neural networks
and have achieved 96.39% precision. Ani et al.’s IoT-built treatment and control [17]
93% accuracy has been reached by the ensemble classifier and system prediction
process. Yang et al. [18] proposed a wearable ECG cloud-based intelligent IoT health
system and proposed a very good system results for disease diagnosis.
An innovative IoT-based breast cancer monitoring system 159
9.3.2 Dataset
The data collection “WDBC” obtainable on machine learning source [20] has been
developed by Dr William Wolberg from the University of Wisconsin. This was
used as a dataset for the conduct of the proposed research to develop master
training programs for diagnosing BC. The dataset includes 569 items, 32 attributes,
and 30 characteristics are real-value characteristics. The diagnosis on the target
performance label has two classes for the malignant or benign individual. Total 357
benign and 212 malignant individuals are classified in class. Therefore, the dataset
is a matrix of 56932.
Algorithm of SVM-RFE
Training examples X0 ¼ ½x1 ; x2 ; . . . xk ; . . . xl T
Class labels y ¼ ½y1 ; y2 ; . . . yk . . . yl T
Initialize:
Subset of surviving features s¼[1,2, . . . n]
Feature ranked list r¼[]
Repeat untils¼[]
Restrict examples of training to successful indices
X¼X0(:,s)
Train the classification
a¼SVM-train(X, y)
X Calculate the dimension length(s) weight vector
w¼ ak yk xk
k Calculate parameters of rating
ci¼(wi)2, for all i
Find the function with the lowest classification criteria
f¼argmin(c)
Update feature ranked list
r¼[s( f)r]
Delete the function with the smallest classification criteria
s¼s(1: f1, fþ1: length(s))
Output:
Feature ranked list r.
As described earlier, for speed reasons, more than one function can be removed
by stage.
Ei ¼ Fi b
Let us mark the indexes of the two multipliers chosen to optimize in one stage
as i2 and i1 according to the Platt [8] pseudo code. An analysis of Platt’s
162 The Internet of medical things
information (1998) shows that we just need to learn about Ei1 Ei2 ¼ Fi1 Fi2 to
take a step from ai1 and ai2 .
For an efficient solution to the problem, the method used for the choice of i1
and i2 at each stage is essential. Platt developed a strong set of heuristics, based on a
variety of experiments. The two-loop method is used: the external loop is i2, and the
internal loop is i1 for the chosen i2. The external loop iterates over all patterns,
violating the optimal conditions—first only over those with Lagrange multipliers at
either the upper or lower boundary of optimality conditions to ensure that the
problem has in fact been solved. Platt maintains and updates an Ei cache for indices
i for no boundary multipliers for efficient implementation. The rest of the eggs are
calculated as necessary.
Now let us see how i1 is chosen by the SMO algorithm. The goal is to increase the
objective function significantly. Because all the possible i1 options to try to pick one that
best enhances the target function are expensive, the i1 index is chosen to optimize
j Ei2 Ei1 j. Since Ei is available in cache for non-boundary multiplier indices, only the
abovementioned i1 indices are used initially. If the option of i1 does not provide ade-
quate improvement, the following steps are taken. All indexes of no bound multipliers,
one by one, are evaluated as i1 choices from a randomly selected set. If enough progress
is not made, all indices are attempted, one at a time, from a randomly chosen index as i1
options again. The random seed collection, therefore, affects SMO’s running time.
Breast cancer
dataset
Preprocessing
Attribute
evaluation
Data partition
Classification
Malignant
Benign
Feature information:
(1) id number
(2–31) Ten real-valued features are computed for each cell nucleus:
(a) radius (mean of distances from center to points on the perimeter)
(b) texture (standard deviation of grayscale values)
(c) perimeter
(d) area
(e) smoothness (local variation in radius lengths)
(f) compactness (perimeter2/area—1.0)
(g) concavity (severity of concave portions of the contour)
(h) concave points (number of concave portions of the contour)
(i) symmetry
(j) fractal dimension (“coastline approximation”—1)
(32) Diagnosis (M¼malignant, B¼benign)
0.5
0
0 0.5 1
0.5
0
0 0.5 1
RFE algorithms and tabled in Table 9.3. All the experiments use SMO parameters
like batchSize¼100, buildCalibrationModels¼false, C¼1.0, epsilon¼1.0E12,
numDecmalPlaces¼2, numFolds¼1, randomSeed¼1, and tolerance para-
meter¼0.001. Tables 9.4 and 9.5 record the output in a specific feature combination
of the SMO polykernel predictive model. The polykernel of SMO has 98.24%
correctly classified with 98.2% TPRate, 2.8% FPRate, 98.3% precision, 98.2%
recall, 96.3% MCC, and ROC area (plotted in Figure 9.13) of 97.7%.
0.5
0
0 0.5 1
0.5
0
0 0.5 1
algorithms and tabled in Table 9.3. SMO parameters, for example: batchSize¼100,
constructionCalibrationModels¼false, C¼1.0, epsilone¼1.0E12, numDecmalPlaces¼2,
numFolds¼1, RandomSeed¼1, and parameter¼0.001 tolerance shall be used for
any experiment. A certain function combination is shown in Tables 9.4 and 9.5 to
demonstrate the efficiency of the SMO RBFKernel prediction model. The
RBFKernel SMO includes 93.85% correctly classified with 93.8% TPR, 10.2%
FPR, 94.3% precision, 93.8% retrieval, 887.1% MCC, and 91.8% of the ROC area
(plotted in Figure 9.15).
9.5 Conclusion
The IoT, which has become a means of processing both actual data and past
knowledge by the evolving part of cloud computing and data mining, has seen the
transition to life in recent years. A diagnostic method is built in this chapter ana-
lysis for the diagnosis of BC. SMO was used to detect BC in the design of the
device computer prediction model. For the proper and relevant selection of the
correct target type of malignant and benign persons, the SVM-RFE attribute
assortment algorithm is used. New sets of characteristics from the Diagnostic BC
dataset were developed by the SVM-RFE algorithm. The dataset was intended for
training and validation. In addition for model performance assessment, techniques
for measuring performance such as correctly and incorrectly classification accu-
racy, TP rates, fp rates, precision, retrieval, MCC, and ROC area were also used.
The WDBC dataset was developed to check the proposed program using 32
features with 30 existent importance features and 569 instances on the UC Irvine
data mining depository. Weka machine learning libraries are used to incorporate
and improve the proposed framework. The test results indicate that the method
suggested efficiently classifies the malignant and benign. Various changes to the
BC characteristics may result in an increase in malignant and benign predictions.
These results indicate that the new diagnostic method can be used to assess BC
reliably and to be incorporated into health care in addition. SVM classifications
were evaluated on a rating subset of features using several kernels, such as the
Normalized PolyKernel, linear, Puk, and RBF. SMO PukKernel-linear efficiency
is higher than other SMO-kernels according to Tables 9.4 and 9.5 and Figure 9.16
graph also indicating. The novelty of the study is planned to distinguish BC and
stable people as a diagnostic device. For the diagnosis of BC, the program used
the FS algorithm SVM-RFE, SMO, training methods, and output assessment
methods.
Machine-based decision-making is more effective for improved detection of
BC. According to Table 9.5, the proposed device performance (SVM-RFE-SMO),
compared with the classification results of other studies proposed, is excellent
and achieves 98.9% classification accuracy. More technology will be used in
future to boost the performance of BC diagnostic framework with the addition of
assortment algorithms, optimization, and dedicated neural network classification
approaches.
178 The Internet of medical things
360
340
320
300
280
260
240
220
200
180
160 NormalizedPolyKernel
140
120 PolyKernel
100
80 PukKernel
60
40 RBFKernel
20
0
ea
TP
FP
FN
s TN
as d %
te
Pr te
on
ll
RO C
ca
C
ra
ra
ar
isi
d
M
Re
ly sifie
ie
TP
FP
C
ec
sif
la
cl
c
In ctly
ct
rre
rre
Co
co
References
[1] American Institute of Cancer Research, “Breast Cancer Statistics,” 2018,
https://www.wcrf.org/dietandcancer/cancertrends/ breast-cancer-statistics.
[2] A. J. Jara, M. A. Zamora, and A. F. G. Skarmeta, An Internet of Things–
Based Personal Device for Diabetes Therapy Management in Ambient
Assisted Living (AAL). Springer, London (2011).
[3] M. Islam, H. Iqbal, R. Haque, and K. Hasan, “Prediction of breast cancer
using support vector machine and K-nearest neighbors,” in Proceedings of
the IEEE Region 10 Humanitarian Technology Conference (R10-HTC),
vol. 23, pp. 1–5, Dhaka, Bangladesh, 2017.
[4] A. M. Ahmad, G. M. Khan, S. A. Mahmud, and J. F. Miller, “Breast cancer
detection using Cartesian genetic programming evolved artificial neural
An innovative IoT-based breast cancer monitoring system 179
Information technology can play a vital role in future smart health-care systems.
Using information technology, health-care services can be improved. This
improvement includes shifting the specialization or department-centric health-care
services to patient-centric health-care services. This shift is necessary to have better
patient experiences and providing specialized health-care services to many patients
with lesser resources. In information technology, the Internet of Things (IoT) is an
advanced approach for ensuring this system. In IoT, industrial IoT (IIoT), Internet
of nano things, Internet of robots, Internet of patients, and Internet of medical
things (IoMT) are some of the concepts important to understand the functionality of
smart health-care system. In addition to IoT or its variants, other IoT-associated
solutions that include blockchain technology, parallel and distributed computing
approaches (cloud/fog/edge), virtualization, cybersecurity, automated software
development, and smart infrastructure development have shown great enhance-
ments in recent times. The objective of this work is to explore different
information-technology-based solutions that can make a patient-centric smart
health-care system feasible in nearby times. In this work, recent developments of
IoT that are used to interconnect health-care objects and made technical revolutions
will be explored initially. Thereafter, IoT association with other technological
approaches will be explored. IoT association with other technical aspects is
necessary to understand the IIoT-based health-care solutions. The survey work will
start with the integration of IoT technology with blockchain technology to keep the
patient data confidential and transparent to authenticated parties. The IoT devices
1
Department of Systemics, School of Computer Science, University of Petroleum and Energy Studies,
Dehradun, India
2
Department of Computer Science and Engineering, Jaypee Institute of Information Technology, Noida,
India
182 The Internet of medical things
integrated with the patient help in collecting the data which will be stored at some
central repository. Here, blockchain maintains data security via cryptography pri-
mitives and protocols. The cryptography primitives will provide confidentiality,
integrity, authentication, access control, and non-repudiation properties, whereas
blockchain technology ensures immutability, transparency, distributed computing
and ledger, and data security. In IoT networks, security is a major concern. IoT
networks contain both resourceful and resource-constraint devices. Both types of
the device-based networks require different security solutions. In any smart health-
care system, both types of device-based solutions are required. This work will
discuss feasible cybersecurity solutions for IoT networks useful in smart health-
care systems that are developed in recent times. The feasible cybersecurity solu-
tions will consider health-care infrastructure as cyberspace for collecting, storing,
and analyzing data. In a similar but larger infrastructure, a vast amount of parallel
and distributed computing approaches are required. In this work, we will be dis-
cussing various approaches that are used in recent times for parallel and distributed
computing in health-care system. This includes cloud, fog, edge, and message
passing interface (MPI)-based approaches. All these approaches are used in recent
times at a large scale for IoT-based applications. These approaches will be required
to understand and compute various healthcare-related tasks using IoT networks and
provide efficient outcomes. Thereafter, data-related technological aspects that
include the usage of artificial intelligence (AI), and machine learning (ML) will be
explored. These aspects will address the recent developments that use AI- and ML-
based approaches and are found to be useful in the smart health-care system. In this
technological aspect, all associated approaches, including deep learning, reinfor-
cement learning, federated learning, neural network, virtual reality (VR), and
augments reality aspects, derived during recent times will be explored.
Virtualization has played an important role in reducing the number of resources and
improving the services. Thus, a few of the recent aspects in the direction that are
associated with smart health-care systems are planned to be discussed here. In a
smart health-care system or Healthcare 4.0-based technological solution, data
privacy and security are other major concerns. The traditional security primitives
and protocols are easily breakable with quantum computers. Post-quantum cryp-
tography aspects are required to handle future security aspects. Thus, this will
discuss the importance of quantum computations for IoT networks. Further, how
post-quantum cryptography will be helpful, recent developments for the health-care
system and applications to automated patient-centric services will be discussed.
Lastly, robotics and drone-based technological solutions that are recently discussed
in the scientific community and important for Healthcare 4.0 or smart health-care
system will be explored. This study will briefly explain the recent medical opera-
tions and experimentations that are conducted successfully using robots and drones.
All the information-technology-based solutions require smart infrastructure. Thus,
requirements of smart infrastructure in the smart health-care system which include
smart electric supply system, smartphone-based mobile applications, smart medical
and pharmaceutical system, smart ambulance and other transportation systems, and
smart medical appliances will be explained.
Patient-centric smart health-care systems 183
10.1 Introduction
Software systems with advanced technologies, including AI, ML, drones, cognitive
computing, parallel and distributed processing, smart and intelligence processing,
and interconnection of devices, have played important roles in various applications
with different forms of their implementations. There are technological aspects like
Healthcare 4.0, Industry 4.0, Agriculture 4.0, and Agriculture 5.0 that integrated
major technological aspects under one umbrella for automation. Thus, the form and
technological aspects vary from application to application. In a smart and intelli-
gent health-care system, these technical aspects have played and will continuously
play vital roles. The major advantaged that information-technology-based aspects
can give to various applications include (i) health-care automation; (ii) patient-
centric health-care services deployment rather than specialization or department-
centric approaches; (iii) remote monitored, control, and analysis-based effective
system; (iv) health-care datacenter creation for effective analysis and handling
futuristic situations; (v) industrial health-care system with large-scale inter-
connectivity; and (vi) secure and trustworthy health-care services. With these
aspects taken into consideration, the major and important functionalities of
healthcare-related applications are discussed in recent times concerning fast and
secure processing approaches toward automation. The effective and efficient
approaches can only be important to implement if a detailed analysis shows its
prediction for a real-time system. The use of AI and ML approaches has made this
possible with the help of advanced resources. Thus, there is a strong need to discuss
and explore such recent advancements in the medical and health-care sector.
This work has surveyed the recent developments in the health-care sector for
handling medical services to patients. In recent developments, technological
aspects are covered in detail. Among technological aspects, those approaches are
found which can handle the COVID-19 or its related situations effectively. The
survey of the proposed system includes the interconnection of medical devices for
patient-centric services. It has been observed that most existing health-care services
are specialization-based services with the least interconnectivity. Lack of con-
nectivity and automation make the medical services slower for patients. This
increases the overall cost and time for health-care operations as well. Thus, there is
a strong need to shift from a specialization-based health-care system to a patient-
centric health-care system. In recent times, various initiatives have been taken
using advanced technologies. This work discussed these technological systems
briefly.
This work is organized as follows. Section 10.2 covers the healthcare-related
advancements using IoT-based technology. Similarly, Section 10.3 covers the need
and advancement of blockchain technological aspects for health-care systems and
services. The interconnection of medical systems, patients, equipment, and pro-
cesses needs data security. Section 10.4 covers health-care cyberspace and asso-
ciated cybersecurity aspects. These aspects are necessary to handle a patient-related
health-care system. Large-scale implementation of health-care services is not pos-
sible without the use of parallel and distributed computing architectures. Thus,
184 The Internet of medical things
important, and recently proposed parallel and distributed computing approaches for
healthcare are discussed in Section 10.5. Section 10.6 discussed the recently pro-
posed AI and machine-learning-based approaches for effectively handling health-
care services and pandemic situations. Section 10.7 discusses the importance of
virtualization for the effective utilization of resources in handling smart health-care
systems. In the future, the quantum computational aspect is necessary for handling
large health-care data and fast processing. Thus, Section 10.8 discusses the
importance of quantum computation and its role in a smart health-care system.
Post-quantum computing and cryptography is the solution to those problems that
identify loopholes in quantum-based breakable solutions. In the health-care system,
data privacy and security is very important. Thus, Section 10.9 discusses the role of
post-quantum computing and cryptography to effectively handle futuristic pan-
demic situations. Section 10.10 discusses the recently implemented and/or pro-
posed drone and robot-based approaches for medical and smart health-care
services. Finally, the conclusion and futuristic aspects are drawn in Section 10.11.
In recent COVID-19 timings, it has been observed that IoT networks played
important roles in handling patients. Many wearable devices can help everyone to
monitor their health-related activities and plan the future. Singh et al. [1] identified
the integration of wearable devices and IoT networks to patient-centric smart
health-care systems. It has been identified that the proposed wearable and IoT-
network-based system can reduce the chances of medical mistakes that include
doctor, medicine, and operational mistakes as well. Further, the proposed system is
capable of handling patients with lesser expenses, efficient machinery for superior
treatment, effective medical control, and enhanced medical diagnosis. An IoT-
based system for handling medical situations (like COVID-19), the important
processes include (i) health data monitoring in remote location, (ii) virtual man-
agement of meetings and conferences with sensor- and camera-based devices, (iii)
data collection, storage, analysis, and visualization for effective control and mon-
itoring, and (iv) case report preparation and presentation for effective display and
handling at various levels. Kamal et al. [2] have discussed the importance of the
patient health-care system in COVID-19 handling, its present situation, challenges,
and opportunities. This is a small comprehensive survey explaining the systems of
integrating wearable and sensor devices attached to a patient’s body. With this
integration, patient monitoring is much easier and automated. Likewise, the
Internet of cameras can be constituted which constantly monitor the patient and
hospital activities. In this monitoring, image-based analysis is possible which can
constantly monitor patient movements, emergencies, remote monitoring, and sur-
veillance. In similar cases, a camera-based interconnected system is helpful to
handle hospital situations. The administration and surveillance are much easier and
can be done with minimum manpower.
Patient-centric smart health-care systems 185
There has been a rapid increase in health-care data in recent years since medical
care has become the most common and necessary part of human lives. To make
medical analysis and prescription comfortable and suitable for patients as well as
doctors, the health sciences are now making their way toward the IoT. The
Patient-centric smart health-care systems 189
implementation of the IoT in healthcare has made remote monitoring easy. Patients
can now be monitored remotely without the need to visit hospitals periodically. All
the data is shared directly with doctors, and the prescription can be made accord-
ingly. But this personal data of patients needs protection as sharing and storing data
using IoT is not secure and can be hacked easily. So, implementation of the
blockchain in smart health-care systems can provide security to the data of the
patients [10,11].
patterns, and heart rate transparently and safely. Patients can use the health app or
ask for information from doctors or hospitals. Today, using the HealthBank
approach, not only patients are able to place their information on this platform, but
it is also available for medical research purposes. This innovation has expanded a
new direction of vulnerability and security for data from IoT-based networks.
Implementation of traditional cryptographic solutions on IoT networks is not a
future aspect as IoT devices have many constraints such as performance, memory,
and computational power. In this work [15], the authors have proposed a novel
structure, a unique architecture that integrates blockchain networks with IoT
devices to achieve privacy and security threats to information integrity. Integration
of smart contracts is very important in this work, by which we can handle author-
ization, access control, data management, and device authentication. The overall
scalability of the network can be improved using an off-chain data storage
mechanism in the framework, this model can be integrated into any existing IoT
application with minimal modifications.
patient and his or her well-being at the center of changes in legislation, policy, and
regulations. Rather, attempts are being made to balance patient results, financial
benefits, and liability concerns. Increasingly, an increasingly acute danger to
patients is the failure to recognize and handle cybersecurity problems associated
with their own well-being. In [18], the argument for patient-centered approaches to
not only medical treatment but also health-care cybersecurity is discussed. The
development of the IoT is still in its infancy and it is important to solve several
similar problems. The IoT is a coherent notion of combining all. IoT has a tre-
mendous opportunity to make accessibility, honesty, availability, scalability, anon-
ymity, and interoperability more available to the public. How to secure the IoT,
however, is a difficult challenge. Device security forms the basis for IoT growth. In
[19], IoT reviews cybersecurity. Protecting and incorporating heterogeneous smart
devices and information technology systems are essential concerns (Information and
Communications Technology). The architectural framework in relation to cyberse-
curity in IoT comprises various features and principles, some protocols related to the
architecture, synchronization and wireless networking, heterogeneous and pervasive
frameworks, lightweight technologies, authentication, etc. The architecture involves,
from a technological perspective, integrity, scalability, reliability, secrecy, avail-
ability, and interoperability between heterogeneous smart devices.
Numerous risks are emerging in the IoT paradigm, including unauthorized or
inappropriate operation, malicious code updates, control bypassing, and data integrity
tampering. In IoT implementations, information may be revealed or lost.
Consequently, safeguarding sensitive documents, passwords, and certificates is cru-
cial. Provided that, unregulated IoT apps and systems reveal embedded patented
algorithms that can easily be pirated or studied, intellectual property can be breached.
To stop the discovery of hidden bugs, it is advised that hackers find it more difficult
to reverse-engineer, study, or manipulate the code in general. To ensure continuity
and authenticity of data, integrity processes are used. Hash functions and digital
signatures are used to ensure the confidentiality of documents. In addition, at the
storage level as well as on the network path in the IoT environment, data protection
must be secured. The anonymity program is the concealment of sources of data. In
addition, this program helps in the preservation of personal privacy and secrecy. In
the IoT, non-repudiation [20] helps to guarantee that a bargaining side is unable to
challenge on official papers the authenticity of its signature. Finally, freshness sig-
nifies the presence of facts and the lack of past texts. Figure 10.1 shows the inte-
gration of IoT and cybersecurity and related factors necessary to understand.
Device or network protection is a key issue, and best practices such as restricting
external computer connections, restricting the usage of the Internet directly to some
important endpoints and devices, ensuring that only designated networks are author-
ized, using keys for safe booting and safe firmware are used in numerous ways to
secure them, enforcing device authentication at each link establishment. Virtual
Private Networks are external networks that allow only partners to access them, which
they pledge to, keep private and have assured honesty. This infrastructure, on the other
hand, is not visible for dynamic global information sharing and is not safe for non-
extranet third parties. DNS Security Extensions use asymmetric cryptography for asset
192 The Internet of medical things
Supporting Implemention
Possible threats Specifications points
technologies
Anonymous Risky
Confidentiallty Applicability
routing operations
DNS Reverse
Availability System or
extensions for
engineering services
defence
record signing to ensure the root validity and integrity of information received or
delivered. Data is wrapped in multiple layers of encryption using the public keys of
onion routers on the transmission network, for example, using onion routing tech-
nology to encrypt and mix the Internet traffic of several senders.
IoT nodes store and process private data, they are turning into a gold mine for
cybercriminals. As a consequence, security, especially the ability to detect com-
promised nodes, as well as the ability to catch and preserve evidence of an attack or
malicious activity, emerges as a top priority for effective IoT network deployment.
In the IoT world [24], anonymity, access control, safe networking, and secure data
storage are also becoming significant security issues. Furthermore, during an
inquiry, every new device we develop, every new sensor we deploy, and every byte
that is organized in an IoT environment can be scrutinized. The rapid growth of IoT
devices and software has resulted in the introduction of various unstable and
insecure nodes. Furthermore, traditional user-driven protection architectures are
inefficient in object-driven IoT networks. Therefore, to protect IoT networks,
obtain, store, and analyze residual information from IoT settings, we need
advanced equipment, techniques, and procedures. Health programs and resources
are more important, innovative, and linked than ever before. While enhancing
clinical quality and changing patient delivery, thus improving human life, protec-
tion of health-care data and devices is becoming a growing concern. Health
equipment and services have grown more intertwined, opening them to emerging
cybersecurity risks. It makes the health-care industry the most vulnerable to serious
safety threats. As mentioned earlier, the problem is compounded by the health-care
services and infrastructures allowed by Cyber-Physical Systems (CPS)-IoT [25] that
are vulnerable to a number of emerging cyberattacks. CPS-IoT systems are cate-
gorized as essential safety and protection systems and present extended attack surface
characteristics of fragmentation, interconnectedness, heterogeneity, and cross-
organizational existence. Cybersecurity threats have the ability to breach users’
privacy, inflict bodily harm, financial losses, and pose a danger to human life, so
avoiding them is vital. With millions of medical documents [26] stolen worldwide,
studies illustrate the increase of attacks and the rise of medical identity fraud.
open topics with many of the benefits that discuss big IoT concerns, such as
usability, portability, interconnection, data storage, and privacy. IoT provides a
systemic basis for a range of high-tech health-care technologies, including real-
time patient tracking, environmental and interior safety data, and ubiquitous and
widespread access to information that helps both health professionals and patients
while also providing essential features for health-care networks like usability,
mobility, and extensibility. Continuous scientific advances allow IoT devices to be
created by countless sensing, data fusion, and logging facilities, leading to many
advances in better living environments (ELEs) [28]. To transfer data between
sensors and servers, smart healthcare relies on a combination of short- and long-
range communications networks. Wi-Fi, wireless metropolitan area network
(WiMAX), Zig-Bee, and are the most popular Bluetooth short-range networking
networks, which are mostly used for short smart health-care communications such as
body area network. In smart healthcare, a series of technology such as mobile com-
munication (global system for mobile communication) and General Packet Radio
Service, Long-Term Evolution (LTE) Advanced, and LTE are used to relay data
from a local server to a ground station. Security requires smart healthcare. Even
though smart healthcare relies on the Internet to connect various devices, security is a
major concern. Complex security mechanisms and algorithms are difficult to
implement because of the restricted presence of IoT computers (limited processing
and battery life). As a consequence, 70% of IoT applications [29], including intelli-
gent health-care equipment, are vulnerable to cyberattacks. Consequently, there are
regular threats and security and privacy concerns. From hospital to patient-centric
healthcare, there is an increasing sense of urgency. IoT is designed to be a powerful
enabler by ensuring that computers and cloud storage, as well as acting agents such
as patients, hospitals, research laboratories, and emergency services, are all in har-
mony. Given the benefits of IoT eHealth [30], a variety of challenges must be
resolved. Data storage, scalability, interoperability, device–network–human inter-
faces, usability, and privacy are among others. In terms of variety, length, and
velocity, we can see that knowledge is getting more dynamic. Uniformity will allow
for interoperability between different devices and records. All people, from children
to the aged, should be able to use interfaces that are clear and genuine. Safe infra-
structure, network protocols, less vulnerable to threats in the cloud, and adequate
training for people handling secure data are all required to increase the security of
devices, networks, clouds, and agents.
This section discusses the parallel and distributed computing-based approaches pro-
posed for the health-care system in recent times. Here, those frameworks are studied
and explored in detail which is recent and found to be effective for handling pan-
demic situations. Details of recently proposed frameworks are discussed as follows.
196 The Internet of medical things
or support data visualization for interpretation, and many more are important for
any system. In health-care systems, AI and ML have played vital roles and it is
expected to continue as well. This section will explore the recent advancement of
the use of AI and associated technologies for COVID-19 and its variants.
Figure 10.2 shows the important technical aspects that will be explored in detail.
Al for robotics and
drone-based system
to effectively
implement
Expert system for health services IoT system with Al for healthcare
healthcare automation and industrial level
implementation
Artificial intelligence
Al-based approach to
improve healthcare
service life cycle
Figure 10.2 AI-based technological aspects for the smart health-care system
Patient-centric smart health-care systems 199
All of these technical aspects are associated with AI-based developments. In brief,
the use of AI-integrated technological aspects is explained as follows.
● AI technology with robotics is used to perform various experimentations in the
sanitization, thermal imaging, monitoring, and surveillance. All of these
activities are contactless and important to handle pandemic situations. In past
pandemic situations, these activities helped to identify the patients with
COVID-19 symptoms, sanitization, and other healthcare-related activities. A
large number of developments in developed and developing countries are
made to address the challenges of handling COVID-19-related situation
handling using drones. Likewise, robotics is equally important. In the past
decade, efforts are made to integrate human intelligence behavior to machines
using robotics for performing health-care operations.
● IoT system generates a lot of data as well. The best way to handle this data is
using AI and ML. The classification or categorization of data, its interpreta-
tions, and visualizations are easily possible using ML approaches.
● Medical data analysis can be performed in various sections, including clinical
trials, patient records, operations success rate, and susceptible, infectious,
recovery model. This analysis and their observations are important to respec-
tive sectors. All of this creates more important when there is a need to speed up
the clinical trials and obtain the results in comparatively lesser time duration.
For example, it has been observed that to successfully obtain COVID-
19 medicine, the clinical trials are performed at a much fast rate.
● In healthcare, the patient service life cycle starts from symptoms identification to post-
discharge treatments. All of these aspects need constant monitoring and inspection that
is not possible using manual processing. Automated processes with the use of tech-
nology can make this feasible. In this aspect, AI technology is very useful.
Dataset A B C D E F G Remark
[35] Y N N N N N N This is an image-based dataset to study
the impact of COVID-19 in recent
times. Here, a comparative analysis
of COVID-19 and the non-COVID-
19 dataset is made to have a better
analysis. This dataset consists of
X-ray and CT chest images
[36] Y N N N N N N This is a dataset of COVID-19 X-ray
dataset. This dataset is divided into
training and test datasets. This data-
set is useful for COVID-19 virus
detection, preliminary diagnosis, and
contributes toward virus control
progress
[37] Y N N N N N N This dataset is image-based one in
which images are classified into six
categories: viral, bacterial, fungal,
lipoid, aspiration, and unknown.
Each of these categories includes
further classification. For example,
viral images are classified as
COVID-19 (SARS-CoV-2), SARS
(SARS-CoV-1), MERS-CoV, Vari-
cella, Influenza, and Herpes. Like-
wise, other categories are classified
[38] Y N N N N N N This dataset contains augmented X-ray
images. In this dataset, images are
classified as all augmented images
and COVID-10 X-ray-augmented
images
[39] Y N N N N N N This dataset contains CT images of a
large number of COVID-19 patients.
This dataset has the classification of
images with COVID-19 and non-
COVID-19 patients. These images
are used by Tongji Hospital, Wuhan,
China for diagnosis and treatment of
patients
[40] Y Y N Y N N N This dataset contains X-ray images
with detailed symptoms. Some of the
symptoms are used to calculate the
COVID-19 score. This analysis is
performed to identify the chances of
COVID-19. Dataset is divided into
three major categories: training,
validation, and testing. Using these
categories, the whole dataset is
(Continues)
Patient-centric smart health-care systems 201
Dataset A B C D E F G Remark
divided into two parts. The first part
contains information about those
images that are X-ray images for
distribution. On the other hand, the
second part contains those X-ray
images that are collected to classify
patient distribution
[41] Y Y N Y N N N This is another dataset over COVID-19
and non-COVID-19 symptoms. This
dataset is regularly updated and the
last update was performed on Janu-
ary 28, 2021. This dataset also con-
tains CXR images having positive
and negative COVID-19 cases. Here,
pneumonia is the most common
feature considered for analysis that
can indicate COVID-19 as well.
Here, chest radiography is applied to
identify COVID-19 cases with dif-
ferent patients. This image-based
analysis helps in identifying critical
factors necessary to understand
COVID-19
[42] Y Y N Y N N N This is another image-based dataset that
contains categories like left lung,
right lung, cardiomediastinum, air-
ways, ground-glass opacities, conso-
lidation, pleural effusion,
pneumothorax, endotracheal tube,
central venous line, monitoring
probes, nasogastric tube, chest tube,
and tubings. Images of all these
categories are used in experimenta-
tion to identify the COVID-19 symp-
toms
[43] Y N N N N N N This is another image-based dataset that
contains lung images for identifying
COVID-19-affected patients
[44] Y Y N Y N N N This image-based dataset focuses on
multiple symptoms for identifying
COVID-19 cases. Here, the dataset is
divided into normal patients’ images,
COVID-19-affected images, pneu-
monia symptom images, lung opacity
images, and metadata of all cate-
gories. Thus, this data is very useful
in COVID-9 case identification and
treatments
(Continues)
202 The Internet of medical things
Dataset A B C D E F G Remark
[45] Y N N N N N N This is another sort dataset with
COVID-19 symptom identification
using comparative analysis of pneu-
monia and normal patient
[46] Y N N N N N N This dataset contains CT images that
can be more than 40 COVID-19
patient’s data for analysis and study
[47] Y N N N N N N This dataset is not publically available
for study
[48] Y N N N N N N This dataset contains COVID-19 fea-
ture-based images that were used in
past for identifying COVID-19 pa-
tients with observable features
[49] Y Y N Y N N N This is an image-based dataset that
classifies the COVID-19 symptoms
into various categories, including
common, less common, and rare. In
the common category, symptoms
include fever, cough, anosmia, fati-
gue, sputum, and shortness of breath.
Among less common features,
myalgia/arthralgia, headache, sore
throat, chills, Pleuritic pain, and
diarrhea are taken for identification.
Likewise, nausea, vomiting, abdom-
inal pain, GU bleeding, nasal con-
gestion, palpitation chest tightness,
hemoptysis, and stroke are taken in
the rare category. In conclusion, this
data is very useful for image and
feature-based analysis
[50] Y N N N N N N This image-based dataset is labeled
with COVID-19 CT scans with left,
and right lungs, and infections.
Dataset is prepared to have a com-
parative analysis of COVID-19 and
non-COVID-19 symptoms with lung
disease features
[51] Y N N N N N N This dataset is prepared with lung and
infection segmentation with limited
annotations, segment COVID-19 CT
scans from non-COVID-19 scans,
and both COVID-19 and non-COV-
ID-19 scans. The annotation and
labeled dataset is useful for easy
classification
[52] Y N N N N N N This is ultrasound-based dataset for
COVID-19 patient identification.
(Continues)
Patient-centric smart health-care systems 203
Dataset A B C D E F G Remark
The dataset is helpful for noninva-
sive, cheap, portable and available
for almost all medical facilities
compared to other challenging and
time-consuming options for COVID-
19 identification
[53] Y N N N N N N This is a chest CT-scan-based dataset
for COVID-19 identification and
adds a provision to apply artificial
intelligence in radiology for data
analysis based on advanced innova-
tive technologies
[54] Y N N N N N N This is another CT scan image-based
public dataset available for SARS-
CoV-2 CT scan. This dataset con-
tains 1,252 images for SARS-CoV-2-
infected COVID-19 patients and
1,230 non-SARS-CoV-2-infected
COVID-19 patients. All of this data
is collected from Sao Paulo, Brazil
hospitals. This dataset can be used
for COVID-19 identification and
comparative feature analysis with
non-COVID-19 patient data
[55] Y N N N N N N BIMCV-COVID-19þ is a large dataset
with chest X-ray and CXR (CR,
DX), and computed tomography
images. In this image-based analysis,
COVID-19 patients can be identified
with radiographic findings, patholo-
gies experimentation, polymerase
chain reaction (PCR) prediction,
immunoglobulin (IgG), and immu-
noglobulin M (IgM) diagnosis. In
addition to this, various other test
data is available for analysis. In this
dataset, high-resolution images are
stored with the medical imaging data
structure that includes semantic seg-
mentation of radiographic findings.
In addition to images, patient’s
demographic evolution, method of
projection, and acquisition are also
available for studies
[56] Y N N N N N N This dataset is labeled chest X-ray and
CT-images-based dataset for
COVID-19 identification. This data-
set is weakly labeled. This dataset is
(Continues)
204 The Internet of medical things
Dataset A B C D E F G Remark
a collection of medical images and
can be used for comparative analysis
of clinical symptoms and clinical
findings of COVID-19 and other
influenza demonstrations
[57] Y N N N N N N This dataset is not publicly available
but is used with AI and machine-
learning-based systems for identify-
ing COVID-19 patients
[58] Y N N N N N N In this work, deep learning is applied
for an image-based dataset for iden-
tifying COVID-19 patients
A: COVID-19 image dataset, B: COVID-19 text dataset, C: COVID-19 audio dataset, D: other health-
care datasets, E: other pandemic datasets, F: medicine dataset, G: patient dataset.
many efficient ways. The use of technology can help in reducing the medical
treatment costs that are increasing day by day. The costly medical treatments are
difficult to select for many especially people living below the poverty line. In the
case of scarcity of resources, and higher medical costs, an AI-based system can
help the patient to self-monitor themselves and reduce the cost to a certain extent.
AI-technology can behave like human intelligence, pattern recognition, data ana-
lysis, anomaly, and outlier detection and prediction. These features can help the
medical system to handle more complex challenges that a human can handle. AI
technology is much rich in handling the problem and finding the solutions. For
example, AI-technology can widely be used for analyzing the COVID-19 symp-
toms in recent times, accelerate the scientific discovery processing to handle
COVID-19, its variants, and future pandemics. Usage of AI technology is not just
helpful for medical staff but it helps administrative staff as well. To administrative
bodies, this system can automate the administrative jobs, accelerate and amplify the
system transparency to staff roles and responsibilities, and remove fabricated ser-
vices. In the medical system, data is very important. To handle the data efficiently,
error detection and removal methods need to be focused upon. This error can occur
in data collected for diagnosis, prognosis, and therapy. In medical diagnosis, patient
data collection, and interpretation using doctor’s knowledge and experience for-
mulate it to a proper diagnosis and therapeutic plans for a patient. In the present
system, all of these plans are prepared by a physician. However, it is expected that
this system can be automated with AI in nearby times. This automation can help the
patients to give their inputs and get the diagnosis in output. An expert system is a
well-known example of this type of treatment. An expert system’s inference engine
can handle complex problems with multiple If-then type rules and transform the
input to actionable outputs. Thus, expert systems are very much useful to doctors in
Patient-centric smart health-care systems 205
Over the clinical set of data, natural language processing can be applied for
generating more medical data that can be further analyzed. This is an iterative
cycle that operates until the desired output is achieved.
● Hospital-level analysis includes hospital services (medical and nonmedical)
that can also generate a large amount of data for analysis. The medical services
include the type of treatment, medicines, operations, equipment, and other
medical support. Among nonmedical operations, administration, payment,
insurance, patient monitoring, and other duties are counted. To automate the
complete medical setup, both medical and nonmedical services have to be
precise that is possible with the least error-based ML approach. For example,
the medicine supplier system should have a record of how, which, and what
medicines are delivered to different hospitals. Analysis of this data can indicate
the future requirements, hospital practices in experimenting with the medicines
and their success ratio, a list of authentic medicine suppliers and distributors in
a specific geographical region, and many more.
● In the patient-level analysis, patient diagnosis and treatments can be analyzed
for a depth study. In this analysis, patient test records that include serology,
PCR, and rapid in-clinic antigen tests data are recorded for COVID-19. This
analysis if analyzed through machine- and deep learning aspects then it is very
much useful for patients to get timely diagnosis and treatment.
case they for virtual machine migration. Smart health-care applications store
patient’s medical records, analyze these records, and monitor the patient’s condi-
tion regularly. Integration of cloud services made the availability of patient records
easy, fast, and to a longer distance. This work has used the Ant colony optimization
algorithm to improve the virtual machine selection for generating a sequence of
random optimized solutions. Here, a large number of iterations are performed that
execute the medical record processing at a much faster rate. Virtual machine
migration using particle swarm optimization and virtual machine selection algo-
rithm minimized the number of virtual machines to be migrated and reduces energy
consumption. The objective of smart health-care application development and
integration with virtualization is to minimize the number of virtual machine
requirements to be migrated, reduce energy consumption, and minimize response
time by handling a large number of users. Demirkan [61] has proposed a framework
to include three-layer architecture for a smart health-care system. In the proposed
framework, three-layer architecture is having an option of the service-oriented
framework to include virtualization of resources. The conceptualized data have
driven model in mobile and cloud-enabled smart health-care system with virtuali-
zation give better performance compared to other approaches. The health-care
organization provides cost-effective solutions to health-care services with IT setup
costs and reduced risks. Figure 10.4 shows the smart health-care system blocks with
virtualization. In this system, inputs can be provided through health-care service
providers with different specializations. These specializations include services to
handle patients, patient data, equipment suppliers, and other services like medicines
distribution and staff arrangements. In an IT-based smart health-care system, ser-
vices operate with different data, applications, and software. In service-oriented
service provider
Health-care
Service-oriented architecture
data, application and software services
service consumer
Health-care
architecture, data, application, and software can be divided into two categories:
software services and components. In software services, all technical aspects rela-
ted to individual software execution can be explored, whereas components include
the integration of one or more software services to achieve health-care objectives.
For example, patient treatment requires a historical data retrieval system and a
patient monitoring system. In another example, data may be required to share with
multiple systems. This feature can be achieved through storage virtualization. In
storage virtualization, medical records can be transferred easily and at a higher rate.
Further, if application virtualization is shared with storage, the advantage of a
complete system can be taken. A smart health-care system with virtualization can
give various advantages. Some of the advantages are explained as follows:
● Big-data-enabled intelligence and acknowledgment management systems and
services are very useful to the health-care system because the health-care
sector can collect, analyze, share, and visualize the medical structured or
unstructured data for patient treatment and actionable decisions.
● The patient-centric or electronic-record-centric health-care system is useful for
the future in keeping the data secure and safe. This data can be utilized for
better medical treatments and patient monitoring. This data can be used for
preparing patient profiles and can be interrelated for symptom-related experi-
mentation in medical treatment.
● Data linked with ambulance and other medical services can help the patients to
arrive and leave the hospital as per his/her needs. Availability of such data can
automate the patient arrival and discharge system.
● Availability of data can reduce medical costs, shorten patient stays, and can
implement medical treatments much easily. All this is possible when medical
staff is educated with advanced technology to handle the medical treatments
and can see and understand the variations in medical data.
In the area of integrating virtualization technology with healthcare, very few
initiatives have been taken. This area can be explored with the objective of (i)
identifying the direct and indirect risks associated with the collection, storage,
sharing, processing, and analysis of data at various levels; (ii) apply ML and AI
principles in prior identification of medical operations with the highest potential of
improving the patient’s condition; (iii) to understand the challenges in enhancing
the successful medical practices to the industry or large-scale implementations; (iv)
focus on research in smart health-care systems that offer opportunities to study the
different environments and propose models, frameworks, and methods to improve
and automate the health-care services.
security to end devices. Patient’s data security is prime concern in futuristic health-
care systems. Quantum-computation-based security system in smart healthcare can
ensure this at a minimum cost. (ii) Analyzing large data with minimum time
duration. Health-care system needs patient’s data to be analyzed at much faster rate
compared to traditional systems. This analysis can predict the disease at an early
stage and helps in patient-centric treatments as well. Quantum computing plays
transformative effects in many ways. For example, quantum computing can help in
developing supersonic drug design at much faster rate. In pharmaceutical industry,
pharmaceutical developments are lengthy and costly. Here, IT-technologies like
artificial intelligence, human organs-on chip, and silico trials play important roles.
For example, the treatment in the case of Ebola virus, and developing new drugs with
smart algorithm. According to [64], the important use cases of quantum computing in
healthcare include the following: (i) it is helpful in diagnostic assistance. Quantum
computing can diagnose the patients at early stage with more accuracy and effi-
ciency. This is possible because of the use of advanced technology and fast com-
putations; (ii) it is helpful in people healthy with personalized interventions and
precision medicine; and (iii) the careful analysis and processing can optimize insur-
ance premiums and price. Quantum computing can play important role in health-care
sector by analyzing the medical images, detecting edges in images and image
matching algorithms. Thus, quantum computing is very helpful in fast image analysis
and image-aided diagnostics. Using quantum computing, it is much easier to com-
bine multiple datasets, perform analysis, and give results at much higher rate.
Quantum support vector machine is very helpful in enhancing the classification and
performing diagnosis that classify cancerous cells from normal easily.
is tried to be integrated with IoT networks and blockchain for providing healthcare-
related services. Quantum computations integrated with blockchain technology
need the implementation of hash computations. In a blockchain network, the hash
function provides security to the network using consensus algorithms. The concept
of Hash-cash is also relevant for integrating immutability and transparency to
transactions.
Kumar et al. [34] proposed a drone-based system for thermal imaging, saniti-
zation, monitoring, surveillance, and medication. In the proposed system, data can
be collected and processing at local and remote places before to share with the
hospital. Here, a drone-based system is designed and used for health-care opera-
tions. To avoid collision among multi-drone flying strategies, single and multilayer
drone flying approaches are proposed. Here, an analysis of the proposed approach
is made using simulators. Results show that the proposed approach is efficient in
drone flying and handling health-care operations. Lum et al. [87] proposed a drone
and robot integrated system for surgical operation. Here, a drone device is used to
move to a remote area where it is difficult to provide health-care service. There, a
robot is an instruction to operate and perform medical surgery. Here, a drone device
is used to signal and arrange the medical facilities necessary for medical operations.
Câmara et al. [88] have discussed the importance of drones in disaster scenarios. In
such scenarios, drone devices are very useful in remote monitoring and surveil-
lance. Drone devices can collect the live on-ground status that can help in arranging
necessary services especially medical and first aid kits. In today’s scenario, drone
devices are available for touchless screens and secure reporting. These services can
make pandemic handling much easier and simpler. The usage of such devices can
reduce the cost of monitoring and extending the health-care service for needful
compared to costs observed during COVID-19 times. Kim et al. [89] addressed the
issue of providing medication and test kits to chronic disease patients. This
experimentation has used drones for this service and it has been found that drones
are very useful in this case. Here, drones are used for two major purposes: (i)
identifying the optimal number of drones required for handling medical services
and (ii) how to reduce the cost of drone-based medical operations where medication
is provided to the patient and test kits (that include blood and other samples) are
collected. In this approach, optimization approach is used for cost–benefit analysis.
Results show that drone-aided health-care services are very powerful for health-
care services and easy to apply. Graboyes and Skorup [90] explored medical
drones. Here, technical and policy challenges are discussed in detail. Flying drones
and collecting data can breach privacy and security. Thus, flying drones in many
countries are not allowed. Besides, the lack of drone-related policies does not
provide easy and effective solutions for integrating this technology with the real-
time application such as healthcare. Here, the solution is proposed to have an initial
step to make it feasible on-ground. In conclusion, implementing the proposed
approach in the real-time application needs to solve large challenges that are dif-
ficult to achieve in the present scenario. To address this challenge, infrastructure
level changes and transparent policies are required.
References
[1] Singh, R. P., Javaid, M., Haleem, A. and Suman, R., 2020. Internet of things
(IoT) applications to fight against COVID-19 pandemic. Diabetes &
Metabolic Syndrome: Clinical Research & Reviews, 14(4), pp. 521–524.
[2] Kamal, M., Aljohani, A. and Alanazi, E., 2020. IoT meets COVID-19:
Status, challenges, and opportunities. arXiv preprint arXiv: 2007. 12268.
[3] Amit Kumar M., 2019. IoT Healthcare, GitHub, https://github.com/mis-
hracodes/IOT-Healthcare.
Patient-centric smart health-care systems 217
[20] Yaqoob, I., Ahmed, E., Muhammad Habib ur Rehman, et al., 2017. The rise
of ransomware and emerging security challenges in the Internet of Things.
Computer Networks, 129, pp. 444–458, doi: 10.1016/j.comnet.2017.09.003.
[21] Nasiri, S., Sadoughi, F., Tadayon, M. H. and Dehnad, A., 2019. Security
requirements of internet of things-based healthcare system: A survey study.
Acta Informatica Medica, 27(4), pp. 253–258, doi: 10.5455/aim.2019.
27.253-258.
[22] Usmonov, B., Evsutin, O., Iskhakov, A., Shelupanov, A., Iskhakova, A. and
Meshcheryakov, R., 2017. The cybersecurity in development of IoT
embedded technologies. In 2017 International Conference on Information
Science and Communications Technologies, ICISCT 2017, vol. 2017 (pp.
1–4), doi: 10.1109/ICISCT.2017.8188589.
[23] Mohan, A., 2014. Cyber security for personal medical devices internet of
things. In Proceedings – IEEE International Conference on Distributed
Computing in Sensor Systems, DCOSS 2014 (pp. 372–374), doi: 10.1109/
DCOSS.2014.49.
[24] Conti, M., Dehghantanha, A., Franke, K. and Watson, S., 2018. Internet of
Things security and forensics: Challenges and opportunities. Future Generation
Computer Systems, 78, pp. 544–546, doi: 10.1016/j.future.2017.07.060.
[25] Abie, H., 2019. Cognitive cybersecurity for CPS-IoT enabled healthcare
ecosystems. In International Symposium on Medical Information and
Communication Technology, ISMICT, vol. 2019, doi: 10.1109/
ISMICT.2019.8743670.
[26] Coventry, L. and Branley, D., 2018. Cybersecurity in healthcare: A narrative
review of trends, threats and ways forward. Maturitas, 113, pp. 48–52, doi:
10.1016/j.maturitas.2018.04.008.
[27] Pan, J. and Yang, Z., 2018. Cybersecurity challenges and opportunities in the
new ‘edge computing þ IoT’ world. In SDN-NFVSec 2018 – Proc.
2018 ACM Int. Work. Secur. Softw. Defin. Networks Netw. Funct.
Virtualization, Co-located with CODASPY 2018, vol. 2018 (pp. 29–32), doi:
10.1145/3180465.3180470.
[28] Marques, G., Pitarma, R., Garcia, N. M. and Pombo, N., 2019. Internet of
things architectures, technologies, applications, challenges, and future
directions for enhanced living environments and healthcare systems: A
review. Electronics, 8(10), p. 1081, doi: 10.3390/electronics8101081.
[29] Ahad, A., Tahir, M. and Yau, K. L. A., 2019. 5G-based smart healthcare
network: Architecture, taxonomy, challenges and future research directions.
IEEE Access, 7, pp. 100747–100762, doi: 10.1109/ACCESS.2019.2930628.
[30] Farahani, B., Firouzi, F., Chang, V., Badaroglu, M., Constant, N. and
Mankodiya, K., 2018. Towards fog-driven IoT eHealth: Promises and chal-
lenges of IoT in medicine and healthcare. Future Generation Computer
Systems, 78, pp. 659–676, doi: 10.1016/j.future.2017.04.036.
[31] Featherstone, R. M., Boldt, R. G., Torabi, N. and Konrad, S. L., 2012.
Provision of pandemic disease information by health sciences librarians: A
Patient-centric smart health-care systems 219
[47] Rahman, T., 2020, COVID 19 Detection with X-Ray & COVID19 - Pytorch,
Kaggle, https://www.kaggle.com/portgasray/covid-19-detection-with-x-ray-
covid19-pytorch
[48] Società Italiana di Radiologia,2020, covid-19 Database, https://sirm.org/
category/senza-categoria/covid-19/
[49] Dr Daniel J. Bel , 30 Oct 2021, COVID-19 radiopaedia, https://radiopaedia.
org/articles/covid-19-4?lang=us
[50] Ma, J.; 2020, COVID-19 CT Lung and Infection Segmentation Dataset,
zendo, https://zenodo.org/record/3757476#.YX2FOlVBzIW
[51] Ma, J. et al., 2021, MP-COVID-19-SegBenchmark, gitee, https://gitee.com/
junma11/COVID-19-CT-Seg-Benchmark
[52] Born, J. et al., 2021, Accelerating Detection of Lung Pathologies with
Explainable Ultrasound Image Analysis ,GitHub, https://github.com/jannis-
born/covid19_pocus_ultrasound
[53] Stanislav E. S., 2020, Artificial intelligence in radiology, MosMedData,
https://mosmed.ai/en/
[54] Eduardo, P., 2020, SARS-COV-2 Ct-Scan Dataset, Kaggle, SARS-COV-2
Ct-Scan Dataset
[55] BIMCV-COVID19,2020, BIMCV-COVID19, Datasets related to
COVID19’s pathology course, BIMCV, https://bimcv.cipf.es/bimcv-pro-
jects/bimcv-covid19/
[56] Yang H. and Sohn E. Expanding Our Understanding of COVID-19 from
Biomedical Literature Using Word Embedding. Int J Environ Res Public
Health. 2021; 18(6): 3005. doi: 10.3390/ijerph18063005. PMID: 33804131;
PMCID: PMC7998313
[57] Banerjee, I., Sinha, P., Purkayastha, S., et al., 2020. Was there COVID-19
back in 2012? Challenge for AI in Diagnosis with Similar Indications. arXiv
preprint arXiv: 2006. 13262.
[58] Ahsan, M. M., Gupta, K. D., Islam, M. M., Sen, S., Rahman, M. and
Hossain, M. S., 2020. Study of different deep learning approach with
explainable AI for screening patients with COVID-19 symptoms: Using CT
scan and chest x-ray image dataset. arXiv preprint arXiv: 2007. 12525.
[59] Sujath, R., Chatterjee, J. M. and Hassanien, A. E., 2020. A machine learning
forecasting model for COVID-19 pandemic in India. Stochastic
Environmental Research and Risk Assessment, 34, pp. 959–972.
[60] Ambigai, S. D., Manivannan, K. and Shanthi, D., 2018. An efficient virtual
machine migration for smart healthcare using particle swarm optimization
algorithm. International Journal of Pure and Applied Mathematics, 118(20),
pp. 3715–3722.
[61] Demirkan, H., 2013. A smart healthcare systems framework. IT
Professional, 15(5), pp. 38–45.
[62] Abd EL-Latif, A. A., Abd-El-Atty, B., Abou-Nassar, E. M. and Venegas-
Andraca, S. E., 2020. Controlled alternate quantum walks based privacy
preserving healthcare images in internet of things. Optics & Laser
Technology, 124, p. 105942.
Patient-centric smart health-care systems 221
[76] Li, C., Tian, Y., Chen, X. and Li, J., 2021. An efficient anti-quantum lattice-
based blind signature for blockchain-enabled systems. Information Sciences,
546, pp. 253–264.
[77] Akleylek, S., Goi, B. M., Yap, W. S., Wong, D. C. K. and Lee, W. K., 2018.
Fast NTRU encryption in GPU for secure IoP communication in post-
quantum era. In 2018 IEEE SmartWorld, Ubiquitous Intelligence &
Computing, Advanced & Trusted Computing, Scalable Computing &
Communications, Cloud & Big Data Computing, Internet of People and
Smart City Innovation (SmartWorld/SCALCOM/UIC/ATC/CBDCom/IOP/
SCI) (pp. 1923–1928). USA: IEEE.
[78] Bellini, E., Caullery, F., Hasikos, A., Manzano, M. and Mateu, V., 2018,
May. You shall not pass!(once again) an IoT application of post-quantum
stateful signature schemes. In Proceedings of the 5th ACM on ASIA Public-
Key Cryptography Workshop (pp. 1924). USA.
[79] Chen, J., Ling, J., Ning, J., et al., 2020. Post-quantum proxy signature
scheme based on the multivariate public key cryptographic signature.
International Journal of Distributed Sensor Networks, 16(4), p. 15501477
20914775.
[80] Gill, S. S., Kumar, A., Singh, H., et al., 2020. Quantum computing: A tax-
onomy, systematic review and future directions. arXiv preprint arXiv: 2010.
15559. UK: Software Practice and Experience, Wiley.
[81] Kumar, A., Elsersy, M., Darwsih, A. and Hassanien, A. E., 2021. “Drones
combat COVID-19 epidemic: Innovating and monitoring approach,” in
Digital Transformation and Emerging Technologies for Fighting COVID-19
Pandemic: Innovative Approaches, p. 175. Cham: Springer.
[82] Kumar, A. and Sharma, K., 2021. “Digital transformation and emerging
technologies for COVID-19 pandemic: Social, global, and industry per-
spectives,” in Artificial Intelligence and Machine Learning for COVID-19,
p. 73. Cham: Springer.
[83] Kumar, A., Sharma, K., Singh, H., Srikanth, P., Krishnamurthi, R. and
Nayyar, A., 2021. “Drone-based social distancing, sanitization, inspection,
monitoring, and control room for COVID-19,” in Artificial Intelligence and
Machine Learning for COVID-19, p. 153. Cham: Springer.
[84] Krishnamurthi, R., Gopinathan, D. and Kumar, A., 2021. “Wearable devices
and COVID-19: State of the art, framework, and challenges,” in Emerging
Technologies for Battling COVID-19: Applications and Innovations, p. 157.
Cham: Springer.
[85] Kumar, A. and Jain, S., 2021. “Drone-based monitoring and redirecting
system,” in Development and Future of Internet of Drones (IoD): Insights,
Trends and Road Ahead, p. 163. Cham: Springer.
[86] Sharma, K., Singh, H., Sharma, D. K., Kumar, A., Nayyar, A. and
Krishnamurthi, R., 2021. “Dynamic models and control techniques for drone
delivery of medications and other healthcare items in COVID-19,” in
Emerging Technologies for Battling Covid-19: Applications and
Innovations, p. 1. Cham: Springer.
Patient-centric smart health-care systems 223
[87] Lum, M. J., Rosen, J., King, H. H., et al., 2007. Telesurgery via unmanned
aerial vehicle (UAV) with a field deployable surgical robot. In MMVR
(pp. 313–315). Netherlands.
[88] Câmara, D., 2014, November. Cavalry to the rescue: Drones fleet to help
rescuers operations over disasters scenarios. In 2014 IEEE Conference on
Antenna Measurements & Applications (CAMA) (pp. 1–4). Antibes Juan-les-
Pins, France: IEEE.
[89] Kim, S. J., Lim, G. J., Cho, J. and Côté, M. J., 2017. Drone-aided healthcare
services for patients with chronic diseases in rural areas. Journal of
Intelligent & Robotic Systems, 88(1), pp. 163–180.
[90] Graboyes, R. F. and Skorup, B., 2020. Medical Drones in the United States
and a Survey of Technical and Policy Challenges. USA: Mercatus Center
Policy Brief.
[91] Zhang, Y. and Lu, M., 2020. A review of recent advancements in soft and
flexible robots for medical applications. The International Journal of
Medical Robotics and Computer Assisted Surgery, 16(3), p. e2096.
[92] Siegfarth, M., Pusch, T. P., Pfeil, A., Renaud, P. and Stallkamp, J., 2020.
Multi-material 3D printed hydraulic actuator for medical robots. Rapid
Prototyping Journal.
[93] Gharibi, M., Boutaba, R. and Waslander, S. L., 2016. Internet of drones.
IEEE Access, 4, pp. 1148–1162.
[94] Lin, C., He, D., Kumar, N., Choo, K. K. R., Vinel, A. and Huang, X., 2018.
Security and privacy for the internet of drones: Challenges and solutions.
IEEE Communications Magazine, 56(1), pp. 64–69.
[95] Lv, Z., 2019. The security of Internet of drones. Computer Communications,
148, pp. 208–214.
[96] Kumar, A., Gopal, K. and Aggarwal, A., 2014. Design and analysis of
lightweight trust mechanism for accessing data in MANETs. KSII
Transactions on Internet & Information Systems, 8(3).
[97] Kumar, A., Gopal, K. and Aggarwal, A., 2013. Outlier detection and treat-
ment for lightweight mobile ad hoc networks. In International Conference
on Heterogeneous Networking for Quality, Reliability, Security and
Robustness (pp. 750–763). Springer, Berlin, Heidelberg.
[98] Kumar, A., Krishnamurthi, R., Nayyar, A., Sharma, K., Grover, V. and
Hossain, E., 2020. A novel smart healthcare design, simulation, and imple-
mentation using healthcare 4.0 processes. IEEE Access, 8, pp. 118433–118471.
[99] Kumar, A., Srikanth, P., Nayyar, A., Sharma, G., Krishnamurthi, R. and
Alazab, M., 2020. A novel simulated-annealing based electric bus system
design, simulation, and analysis for Dehradun Smart City. IEEE Access, 8,
pp. 89395–89424.
[100] Kumar, A., Aggarwal, A. and Gopal, K., 2018. A novel and efficient
reader-to-reader and tag-to-tag anti-collision protocol. IETE Journal of
Research, pp. 1–12.
[101] Kumar, A., Rajalakshmi, K., Jain, S., Nayyar, A. and Abouhawwash, M.,
2020. A novel heuristic simulation-optimization method for critical
224 The Internet of medical things
11.1 Introduction
1
School of Computer Science, University of Petroleum and Energy Studies, Dehradun, India
226 The Internet of medical things
are just a few of the technologies that use AI. The use of AI presents legal, ethical,
and security problems. Empower new technology to enhance human health and
well-being in areas such as “primary care, service delivery, medical data integra-
tion and analysis, and disease outbreaks and other medical crises.” Projects are
intended to offer 1 billion additional individuals with improved healthcare and
well-being AI applications [2].
One of the driving forces behind the rise of AI is the data economy [3]. It
relates to how much data has grown in recent years and how much more can be
added in the future. The proliferation of data has spawned a new economy, and
there is a continual struggle for data ownership among businesses seeking to
profit from it. The increase in data volume has given rise to big data, which aids
in the management of massive volumes of data. Data science aids in the analysis
of the data. As a result, data science is moving toward a new paradigm in which
robots may be taught to learn from data and provide a range of useful insights,
giving rise to AI. AI refers to artificial intelligence that mimics human and animal
intelligence. It incorporates intelligent agents, which are self-contained creatures
that sense their surroundings and take actions to improve their chances of
achieving a certain objective. AI is a method for computers to imitate human
intellect through reasoning. It is a program with the ability to perceive, reason,
and act. AI is redefining industries by providing greater personalization to users
and automating processes.
ML is at the heart of AI. The first part of ML involves using algorithms to
discover meaning in random and unordered data, and the second part involves using
learning algorithms to identify a link between that knowledge enhance the learning
process. As a result, the ultimate objective of ML is to enhance the computers’
performance on a given job, such as diseases diagnosis, prediction of drugs, and
automation of healthcare services.
The causes for the rapid development may be ascribed to: an ageing popu-
lation need more health care services; an increase in chronic illnesses necessi-
tating more long-term healthcare services; and many older individuals prefer to
live in residential care facilities or in their own homes. These are not insignificant
reasons for the need for additional nurses, but the robotic revolution powered by
AI will be able to address them [4]. To alleviate this deficit in the medical area,
nursing robots [5] are already in use medical treatments that are more effective
and safer.
diagnosis and helpful testing suggestions as a useful tool for identifying various
diseases. Medical providers and patients can benefit more from the huge medical
data of physiology, behavior, laboratory, and medical imaging, which is combined
with the development of ML technology.
Today, hospitals generate abundant data that provides analytics in order to
assist patients in the future which is extremely challenging. AI is a wide phrase that
refers to a large-scale endeavor to develop nonhuman intelligence. AI allows using
data from tens of thousands of patients to anticipate what will happen to a specific
patient and prevent it from occurring. To do so, researchers may utilize data from
tens of thousands of patients to predict who is at danger and intervene to prevent it
from happening. In healthcare, humans are dealing with an enormous quantity of
data, and the only way it can be evaluated is via the use of machines. The ultimate
objective of this type of technology is to reduce the cost, improve efficiency, and
ensure the safety of the service provided. The importance of early therapy in
enhancing outcomes cannot be overstated. Another area where it is beneficial is in
solving the global shortage of medical knowledge. AI’s contribution to healthcare
is that it can aid in the scaling up of some of these complicated and vital jobs. When
extra labor is necessary, which is common in the medical profession, AI models
work ceaselessly, which is a problem.
Another difficult deep learning topic with much more complicated data is
pathology [6]. Being able to visually distinguish between hundreds of diseases is
quite difficult. Additionally, field that has benefited greatly from deep learning
is genomics [7]. Medical data processing is demanding and hard to perform data
modeling. This may be accomplished by correcting bias in the training data as well
as bias in the model architecture and issue formulation. If none of this applies,
researchers may test and guarantee that equitable results and resource allocations
are achieved at the conclusion of AI model deployment. When it comes to public
health, epidemiological models are highly useful, but there are other aspects to
consider, such as climate change, flood forecasts for public health warnings, land
stability, and climate restrictions.
in medical images but now new AI solutions and image detection techniques
have made these efforts much easier. The use of various ML techniques to train
the models from labeled data where inputs and outputs both are known and then
based upon the learning, the model predicts output feature for new test data.
The model keeps on learning and increases its efficiency to give better results.
This automation in the health industry can change the facet of the health industry
by many folds. To provide this automation, one needs to process the health-care
datasets for a well-defined research problem and the system can be trained
for automation.
health-care datasets/databases are small and cannot fulfill the variety of cases that
exist across the huge population in India.
11.5 AI in healthcare
AI is a technology where machines analyze and process the data in way much
similar to human brain mechanism. AI is a big term that includes many other
techniques and technology to simulate the capabilities of human brain. This is a
promising technology that holds the key to the future of humanity [8,9]. It has
huge potential in all domains of life. Healthcare is one of the domains where AI
is most sought after technology due to the global pandemic—COVID-19 the
world is currently facing. Scientists and researchers are decoding AI to answer
the medical research questions that cannot be answered by any other
technologies.
AI can change the landscape of healthcare and is coloring the dream image of
revolutionizing healthcare and help address some of the challenges in the field.
Many countries are investing huge amount in health-care research using AI. India is
also investing in this field since 2012, which has also given rise to many health-care
startups in India [3]. India is a country that provides some of the best hospitals and
highly qualifies medical experts and staff; but still it lacks in proper health-care
infrastructure. There are many initiatives in collaboration with government, big
health-care companies, and hospitals to improve the health-care infrastructure in
India using AI techniques. Currently, there are many firms that are using AI tech-
niques to help in diagnosis and prediction of diseases. One such example is a
startup based in Bangalore which is using ML to diagnose cancer; few examples are
there who are trying to provide primary care and early detection of diseases based
on patients’ previous EMR. Seeing the huge potential of AI in health-care tech-
nology giants Google, Microsoft, IBM, and many others are collaborating with
govt. and health service providers to design software and tools to aid in health-care
services. AI in healthcare around the globe has many success stories and holds
promising results. Figure 11.1 shows the various medical fields where AI can
be used.
The applications mentioned earlier, some are already in use or in their early
phases of commercialization around the globe. But there are few misconceptions
about this high end technology—one of the most prevailing is—it will kill human
jobs or in terms of healthcare it will replace the medical staff or doctors. It is a
misconception, as there are still many areas of AI which have many challenges. It is
very far from replacing doctors or medical staff [4].
Currently, AI algorithms are lacking to fulfill the requirements of processing
any type of health-care data. It needs lot of efforts and big initiatives to build a self-
learning model that should be able to treat and manage diseases and can help in
clinical decision-making, it needs human intervention at many stages [5]. But
collaborative efforts of govt. big firms and medical fraternity can produce the
promising results in coming years.
The promising application of AI in health-care sector in current years is in
imagery detection, for example, cancer detection from the images and retinal
diseases detection where it provides great accuracy and very quick and precise
results.
Application of intelligent techniques in health-care sector 231
Artificial
intelligence
Automation of administration.
Improve patient experience.
have achieved very high accuracies in determining skin cancer as well as other
skin diseases from various contextual, micro- and macro images [9].
3. Designing treatment plans and diagnosis: AI can help in designing pre-
scriptions and treatment plans for patients. AI can analyze data from other
patients and suggest strategies for treating other patients. From medical images
such as X-rays, MRIs, and ultrasounds, AI has higher capability to detect signs
of disease based upon various algorithms that suggest ways to diagnose them
as well. For example, the use of support vector machines (SVM) to detect heart
cancer based upon details of cell and use of neural networks to detect lung
cancer.
Following are ML techniques that are used [13]:
1. Supervised learning: In this kind of learning, model is provided with labeled
data. Inputs and outputs both are known, and the model tries to learn through
the given data. This data used to train the model is known as train data. After
learning the model tries to predict output values and if they are wrong, model
tries to improve the accuracy.
2. Un-supervised learning: In this kind of learning, input to the model is fed
with unlabeled data. The model is not provided with any relation between
inputs and outputs. This learning technique is very complex as model has to
learn itself. Unsupervised learning is mainly used for cluster analysis, anomaly
detection, etc.
The most immediate need for AI in healthcare is disease diagnosis. Early
diagnosis of common diseases such as breast cancer, diabetes, coronary artery
disease, and tumors can help patients control and minimize their chances of dying
from these illnesses. With the advancement in ML and AI, several classifiers
and clustering algorithms, including SVM, logistic regression/linear regression,
K-nearest neighbor, decision trees, Naı̈ve Bayes, random forest, and adaptive
boosting, are used for the prediction of the abovementioned diseases and gain
unprecedented insights into diagnostics [11].
As discussed earlier, health-care dataset comes in the form of image/video
and text/signals which is quite challenging to process and predict patterns. The AI
algorithms need to classify the available dataset into different classes. These
classes serve as training dataset to effectively recognize medical events or pat-
terns in the data. These algorithms need to be trained for any outliers as well for
better efficacy.
beneficial. These challenges may look daunting, but they might be an opportunity
for growth for organizations wanting to embrace exemplary responses to the clin-
ical and ethical problems raised by a tough but promising set of AI-based health-
care practices. Rapid advances in AI research, as well as government and private-
sector funding, make it extremely probable that AI will be widely used in health-
care delivery, with significant cost-cutting and service-quality-improvement
potential. In current scenario, huge amount of work needs to be done in the realm
of AI in health-care. Government needs to lay down guidelines in terms of privacy
and security, law and responsibility, psychological and ethical issues as far as AI in
healthcare is concerned. AI will become a strong tool for saving lives and
improving their quality of life if humans develop on all of these areas.
References
[1] Secinaro, S., Calandra, D., Secinaro, A., et al. The role of artificial intelli-
gence in healthcare: a structured literature review. BMC Medical Informatics
and Decision Making 2021; 21: 125. https://doi.org/10.1186/s12911-021-
01488-9.
[2] Mead, L. Global Summit Focuses on The Role of Artificial Intelligence in
Advancing SDGs. SDG knowledge hub 2018. [accessed 10 August 2019].
Available at: http://sdg.iisd.org/news/global-summit-focuses-on-the-role-of-
artificial-intelligence-inadvancing-sdgs/.
[3] Vijai, C and Wisetsri, W. Rise of artificial intelligence in healthcare startups
in India, Advances in Management, 2021; 14(1): 48–52.
[4] Pepito, J. A. and Locsin, R. Can nurses remain relevant in a technologically
advanced future? International Journal of Nursing Sciences 2019; 6(1):
106–110 [accessed 20 November 2019]. Available at: https://www.science-
direct.com/science/article/pii/S2352013218301765.
[5] Väänänen, A., Haataja, K., Vehviläinen-Julkunen, K. and Toivanen, P. AI in
healthcare: a narrative review [version 1; peer review: 1 not approved].
F1000Research 2021; 10: 6. https://doi.org/10.12688/f1000research.26997.1.
[6] Yu, K.-H. , Beam, A. L. and Kohane, I. S. Artificial intelligence in health-
care. Nature Biomedical Engineering 2018; 2: 719–731. www.nature.com/
natbiomedeng.
[7] Kohli, P. S. and Arora, S. Application of machine learning in disease predic-
tion. 2018 4th International Conference on Computing Communication and
Automation (ICCCA), 2018. 978-1-5386-6947-1/18/$31.00 2018 IEEE.
[8] Claire Muñoz P. and Aneja, U. Artificial Intelligence for Healthcare:
Insights from India, ISBN: 978 1 78413 394 8, 2020. [Accessed 27 July
2021]. Available at: https://www.chathamhouse.org/2020/07/artificial-
intelligence-healthcare-insights-india.
236 The Internet of medical things
Analyzing clinical data is a great challenge in today’s digital data world. This
perspective imposes the need of machine learning (ML) algorithms to extract useful
patterns in clinical data. This chapter improves patient care by diagnosing disease
accurately. It also helps to study the importance of clinical data and managing into
PySpark environment. Various disease datasets are trained to ML techniques
(MLT) to identify the best model. It ensures the collection of clinical data from
different sources, integrating and extracting the useful patterns with less time
consumption. This approach improves the understanding of clinical data and
improves patient care.
12.1 Introduction
The promising approach of ML in health-care data is improving day by day. It is an
important technique that drives to the advancement of artificial intelligence.
Maintaining the history of patient’s data improves the health-care populations and
minimizes cost of healthcare. Nevertheless, the ability to manage huge datasets is
difficult for human knowledge and it is not reliable to convert or analyze medical
data. Converting the data into useful sights and attaching to real-time data is tedious.
Managing clinical data information and achieving a good decision is difficult
in health-care scenario. Extracting useful information to make a decision with
traditional man-made analysis is not sufficient. Medical data analysis is moved to
efficient analysis to promote a useful information and diagnose meaningful infor-
mation. The objective of this chapter is to understand that MLT concepts are
suitable for clinical data analysis. The prediction of disease might be different
based on features. This targets the improvement of individual patient care and
suggests the supporting features. For transforming clinical data [1] to useful data,
this approach is suitable. A novel ML algorithm [2] is developed for diagnosing
disease. The main objective of this chapter is to develop and classify MLT for
1
Department of Computer Science, Periyar University, Salem, India
238 The Internet of medical things
different diseases. The histories of patient records of various diseases such as dia-
betes, kidney and heart disease are diagnosed through MLT.
The objective of this work is to analyze clinical data and make all the models
run in pipeline process for datasets. Section 12.2 discusses the related work about
clinical data analysis in MLT. Section 12.3 comprises the results of clinical datasets
validation and results. Section 12.4 concludes the chapter.
Clinical Records
Diabetes Train
Heart Machine
Data Feature learning
preprocessing selection techniques
Kidney
Test
Other
clinical
records
ML pipeline
Process
LR
Train
Model
PySpark RF validation
Data frame and
source Test selection of
model
Other model
Figure 12.2 represents the flow of clinical data analysis in PySpark framework;
the data source is divided into training and testing. To develop a spark environment
[18] create a Spark Context to connect with clusters. Using the one hot encoding
method converts the categorical data into numerical data, String Indexer transforms
the string features into numerical form, for example, “M” represents that the male
is converted into 0 and similarly “F” represents the female as 1. Certain techniques
are used for preprocessing. The data is transformed to the pipeline to run the
algorithm simultaneously to achieve better results.
12.3.1 Dataset 1
Pima dataset is collected from UCI repository consists of nine features that are
calculated with PySpark approach. Table 12.2 represents the accuracy of different
MLT and their accuracy level: logistic regression (LR), decision tree (DT), random
forest (RF), support vector classifier (SVC) and gradient boosting (GBT). High
accuracy is achieved in RF classifier model with 81 percentage of accuracy and
240 The Internet of medical things
minimum test error of 0.19. The time consumption for running the algorithm is low
in LR model and it is high in GBT model.
Table 12.3 statistics explains the raw prediction and probability of age feature.
The value increases and decreases on the basis of age. It is high at age 25, 22 and
31; only the first 10 rows are listed in the table.
Tables 12.4–12.7 describe the prediction and probability values of different
methods. The value is increasing and decreasing based on the age factor. Based on
Managing clinical data using machine learning techniques 241
age, the probability value cannot be concluded in different MLT. It is purely based
on the prediction and training results of the algorithm.
12.3.2 Dataset 2
Heart disease data is collected from [15] UCI repository consisting of 303 instances
and 14 features. This data is divided into train and testing and given as input to ML
pipeline process. Table 12.8 summarizes the accuracy of different MLT, where RF
attains high accuracy rate.
Table 12.9 shows the age value of heart disease dataset with raw prediction and
probability. The value of raw prediction is high at the age of 54 and minimum at the
age of 57. Similarly, the probability is maximum at the age of 58 and low at the age
of 57. So, the prediction value is obtained from these statistics value. This table
concludes that if raw prediction value increases, the probability value decreases.
Tables 12.10–12.13 describe clinical dataset 2 manipulated using different
MLT raw predictions and probability value. The value increases and decreases on
the basis of the age factor. In DT method, the values are similar for raw prediction.
Considering RF method at age of 52, the prediction value is maximum and at the
age of 54, the probability value is high. In SVC and GBT, the value is high at the
age of 64. The higher the age value, the prediction value is high.
12.3.3 Dataset 3
Diabetes dataset is collected and consists of 100 000 instances. Using feature
selection [19], only 11 features are selected to divide to test and train. This data is
given as input to pipeline MLT observe Table 12.14 results. The accuracy obtained
by RF model is high compared to other models.
Table 12.15 summarizes the raw prediction and probability value based on age.
The prediction value is high at the age of 35 and is minimum at age 65. The
probability value is low at the age of 65.
Tables 12.16–12.19 describe the various raw prediction and probability values
for diabetes dataset. For certain age value, the value is high, which means per-
centage of predicted probability is high.
12.3.4 Dataset 4
Chronic kidney disease dataset is collected from UCI repository [17] and consists
of 400 patient details. Feature consists of a combination of string, integer and
double value. The data is preprocessed using techniques in PySpark. Table 12.20
describes the accuracy and test error values. RF technique obtained high value and
minimum time consumed for classifying the algorithm.
Managing clinical data using machine learning techniques 245
Table 12.21 summarizes the kidney disease raw prediction and probability
value. The raw prediction value is high at the age of 72 and minimum at 63. The
probability value is maximum at the age of 59 and minimum at age 33.
Managing clinical data using machine learning techniques 247
Tables 12.22–12.25 show the raw prediction and probability value of chronic
disease dataset. The probability and raw prediction based on age factor increases
and decreases.
248 The Internet of medical things
1.2
0.8
Dataset 1
0.6
Dataset 2
0.4 Dataset 3
0.2 Dataset 4
0
LR DT RF SVC GBT
MLT
160
140
120
100 Dataset 1
80
Dataset 2
60
40 Dataset 3
20 Dataset 4
0
LR DT RF SVC GBT
MLT
Figure 12.3 explains the different datasets with various MLT. Among the
models, RF model that acquires high accuracy is the best for clinical data analysis.
The second highest accuracy is DT. Figure 12.4 describes the time consumed to train
and test the algorithm. The RF acquires minimum time in seconds for large dataset.
12.4 Conclusion
This chapter discusses three different aspects of clinical data. First step is to ana-
lyze the data, integrate the model using pipeline and validate the results. Among the
different benchmark datasets, RF works better in accuracy and time. It also works
best in raw prediction based on age. This helps us to define the usage of MLT and
their scope of clinical data from small to large dataset. Comparison with other
model diagnosis disease better and improve the patient care. For different features,
managing clinical data using MLT gives better results.
References
[1] A. Colubri, T. Silver, T. Fradet, K. Retzepi, B. Fry, and P. Sabeti,
Transforming clinical data into actionable prognosis models: machine-
learning framework and field-deployable app to predict outcome of Ebola
patients. PLoS Negl Trop Dis. 2016; 10(3): e0004549.
[2] S. Dagdanpurev, S. Abe, G. Sun, et al., A novel machine-learning-based
infection screening system via 2013–2017 seasonal influenza patients’ vital
signs as training datasets. J Inf Secur. 2019; 78(5): 409–421.
[3] C. Verdonk, and F. Verdonk, How machine learning could be used in clin-
ical practice during an epidemic. Critical Care, 24(1). doi:10.1186/s13054-
020-02962-y.
[4] S. J. Patel, D. B. Chamberlain, and J. M. Chamberlain. (2018). A machine
learning approach to predicting need for hospitalization for pediatric asthma
exacerbation at the time of Emergency Department Triage. Academic
Emergency Medicine, 2018; 25(12): 1463–1470. doi:10.1111/acem.13655.
250 The Internet of medical things
13.1 Introduction
The Internet of Things (IoT) has proven useful in the field of electronic-health
(E-Health) management as a network of sensors gathering data both locally and
remotely. The collection of patient vitals and the provision of essential track and
trace services for pandemic management have been made possible thanks to a
combination of body area networks and field monitoring devices. Health data such
as blood pressure (BP), temperature, and heart rate can be collected by locally
based E-Health mechanisms. This data can be saved locally and accessed by a
health-care provider. Local systems can also be used to notify patients when they
need to consult with medical staff to take medicine. Remote-based E-Health is
critical for health-care providers because it allows them to access patients and
patient data from afar. Patient vitals and location may be sent to local or faraway
medical facilities at regular intervals for monitoring purposes. In the event of a
global pandemic, such as the 2019 coronavirus (COVID-19), it is important to
follow social distance guidelines and track and trace patients successfully. These
two factors play an important role in limiting the virus’s global spread. IoT ser-
vices’ ability to provide remote data collection and monitoring of patients in
quarantine has made them a key component in the fight against virus pandemics.
To manage a rapidly spreading respiratory pandemic, health workers and autho-
rities need data. In the case of COVID-19, data can be used to begin the diagnosis
of infection as well as track the spread of the virus in the community. Body tem-
perature, location, and travel history are the most important pieces of information.
These parameters can alert officials as to whether or not more investigation and
1
Department of Electronics and Instrumentation Engineering, Erode Sengunthar Engineering College,
Anna University, Erode, India
2
Department of Biomedical Engineering, Erode Sengunthar Engineering College, Anna University,
Erode, India
252 The Internet of medical things
testing are needed. Initially, health workers relied on manual methods such as
infrared thermometers to measure temperatures and verbal interrogation of people
about their backgrounds and locations. Because of the increased contact with
potentially infected subjects, this posed a risk to health workers. As infection rates
reached the millions, it had also become a more difficult approach.
some of the health-care areas where IoT integration appears imminent. The
majority of today’s devices are portable diagnostic devices with traditional
connectivity.
13.4 Benefits
Smartphone auxiliary
healthcare applications
Clinical
communications Voalte One, Clinical reach, Amcom mobile connect, Practice unite,
Emergency medicine program, Tradassan, mVisum, Vocera.
apps
13.5 Challenges
13.5.1 Data security and privacy
Data security and privacy are two of the most serious threats posed by IoT. Data is
captured and transmitted in real time by IoT devices. All of these considerations
make the data extremely vulnerable to cybercriminals who can hack into the system
and compromise both patients’ and physicians’ personal health information (PHI).
Cybercriminals may use patient information to produce fake IDs in order to pur-
chase drugs and medical equipment that they will later sell. Hackers will also file a
false insurance claim in the name of a patient.
13.5.2 Cost
You would be surprised to learn that cost factors are included in the challenge parts. I
know most of you are, but the bottom line is that IoT has not yet made healthcare
more accessible to the average person. The rise in health-care costs is a cause for
concern for everyone, particularly in developed countries. The situation has led to the
development of “medical tourism,” in which patients with critical illnesses use
health-care facilities in developed countries for a fraction of the cost. The principle of
IoT in healthcare is a fascinating and promising one.
hypertension and diabetes. In hospitals, telemetry has been used to track a large
number of patients with limited staff by transmitting biometric measurements
such as heartbeat and BP from wearable, wireless instruments on patients to
central monitoring. IoT should be used to decrease the workload and improve the
productivity of medical personnel while also lowering the risk of infection among
health-care workers.
Doctors
Hospitals
Alternaive and
medicine Nursing
Homes
Patient
Care
Networks,Edge Devices,
CloudComputing
Community
health
self-reporting and
decentralized
testing
National
communication
Public health
Cloud Mobile phone coordination
Edge
server Global
communication
Internet
Epidemiology
database
Real-time data
update
Radio tower
Connectivity
Geo-spatial
Camera contact
tracina
Boimetric
reader On-board
camera and
sensors
a doubt, the goal of clever healthcare. Intelligent healthcare, for example, uses
wearable devices, a flexible web, and IoT to collect data from people, equipment,
and agencies involved in health-care services and then uses that data to oversee and
respond to health-care needs in a thoughtful manner. Take COVID-19, for example.
By gathering, incorporating, and analyzing exact, appropriate, and high-quality data
in real-time, intelligent healthcare can prevent the virus’s transmission and spread.
To track new COVID-19 cases, intelligent healthcare will gather data through
patient-centered health-based apps. In addition, instead of encounter-based care,
wearable technology (bodyworn-sensors) can be used to provide healthcare to
COVID-19 patients via constant linked care. Additionally, through continuous data
Use of IoT and mobile technology in virus outbreak tracking and monitoring 263
stream and growth, potential COVID-19 hotspots can be proactively identified and
tracked. This makes the virus’s prevention and spread easier. In addition, by com-
bining several data sources, intelligent healthcare will increase community safety.
We examine existing business developments and how they are applied to healthcare
to gain a better understanding of health-based IoT. We have divided the IoT-based
health-care evolution into two categories based on our results. To begin, consider the
use of H-IoT prior to COVID-19. Second, how has the IoT in healthcare react to the
COVID-19 pandemic.
care of their own. Doctors have been able to better serve their patients in a timely
fashion thanks to mobile technology. According to Heron (2010), equated monthly
installment (EMI) has aided in the treatment of a variety of disorders, including
obsessive-compulsive disorder, generalized anxiety disorder, panic disorder, and
social phobia. There are a slew of other ailments for which mobile technology has
made a difference in people’s lives.
13.12 Conclusion
We suggest a model for IoT-based health-care systems in this chapter, which can be
used for both general systems and systems that monitor special conditions. Then,
for each component of the proposed model, we submitted a detailed and systematic
overview of the state-of-the-art works. Several nonintrusive, wearable sensors were
demonstrated and evaluated, with a focus on those that monitor vital signs, BP, and
blood oxygen levels. The suitability of short-range and long-range communication
requirements for health-care applications was then compared. For short-range and
long-range communications in healthcare, bilateral lower extremity and Narrow
Band-Internet of Things (NB-IoT) emerged as the most suitable standards. Recent
cloud-based data storage research was introduced, demonstrating that the cloud is
the best option for storing and coordinating big data in healthcare. Several studies
have also found that data processing in the cloud is much better than data proces-
sing on wearable devices with their limited resources. The most important dis-
advantage of using the cloud is that it adds security risks; as a result, we introduced
several works aimed at enhancing cloud security. Access control policies and
encryption were discovered to substantially improve security, but no known stan-
dard is suitable for immediate implementation in a wearable, IoT-based health-care
system. We found several important areas for future study based on our analysis of
state-of-the-art technologies in the fields of wearable sensors, communication
standards, and cloud technology. Machine learning and the development of a secure
but lightweight encryption scheme for cloud storage are the two areas where
researchers looking to make substantial changes in the field of IoT-based healthcare
have the most opportunities.
266 The Internet of medical things
Further reading
[1] M.U. Ahmed, S. Begum, and J.-B. Fasquel (Eds.) 4th International
Conference, HealthyIoT 2017, Angers, France, 2017, pp. 3–9.
[2] K. Govinda Contemporary Applications of Mobile Computing in Healthcare
Settings, VIT University, Vellore, India, 2018, pp. 51–57. Copyright:
2018.
[3] A.K. Singh, R.S. Singh, A.K. Pandey, S. Udmale, and A. Chaudhary ISBN:
9780128214725, 11th November 2020, pp. 36–59.
[4] C. Chakraborty, A. Banerjee, A. Kolekar, H. Maheshkumar, and B.
Chakraborty Internet of Things for Healthcare Technologies, March 2020
Pages: 21–43. ISBN: 978-981-15-4111-7.
[5] A. Gantait, J. Patra, and A. Mukherjee Defining Your IoT Governance
Practices, January 19, 2018. Updated January 20, 2018.
[6] P. Sharma How Edge Computing in Healthcare Is Transforming IoT
Implementation. Apc.com, December 12, 2017.
[7] Edge Computing for IoT in Buildings. Navigant Research, Q4 2018.
[8] Global Smart Healthcare Products Market Will Reach USD 66.7 Billion by
2024, Zion Research, January 2019.
[9] Gartner, IT Glossary, Internet of Things http://www.gartner.com/it-glossary/
internet-of-things/.
[10] Gartner, Press release, 2013, online at http://www.gartner.com/newsroom/id/
2636073.
[11] ITU Internet Reports, The Internet of Things, November 2005 http://www.
itu.int/osg/spu/publications/internetofthings/InternetofThings_summary.pdf.
[12] IERC – European Research Cluster on the Internet of Things, “Internet of
Things – Pan European Research and Innovation Vision”, October, 2011.
[13] L. Adori, A. Iera, and G. Morabito The Internet of Things: A Survey,
Computer Networks, 54 (2010) 2787–2805.
[14] H. Jun-Wei, Y. Shouyi, L. Leibo, Z. Zhen, and W. Shaojun A Crop
Monitoring System Based on Wireless Sensor Network, Procedia
Environmental Sciences, 11 (2011) 558–565.
[15] A.M. Vilamovska, E. Hattziandreu, R. Schindler, C. Van Oranje, H.
DeVries, and J. Krapelse RFID Application in Healthcare – Scoping and
Identifying Areas for RFID Deployment in Healthcare Delivery, RAND,
Europe, 2009.
[16] P. Pande Internet of Things – A Future of Internet: A Survey, International
Journal of Advance Research in Computer Science and Management
Studies, 2(2), 2014.
[17] Internet of Things: From Research and Innovation to Market Deployment –
IERC 2014. http://www.internet-of-things-research.eu/pdf/IERC_Cluster_
Book_2014_Ch.3_SRIA_WEB.pdf.
[18] Proteus, Digital health feedback system, http://www.proteus.com/technol-
ogy/digital-health-feedback-system.
Use of IoT and mobile technology in virus outbreak tracking and monitoring 267
14.1 Introduction
Early diagnosis of neonatal disorders may be crucial for timely intervention and
treatment. As some rare but potentially harmful diseases in newborns that can man-
ifest themselves with clinical symptoms, such as seizures and apneas, affect the
movements of the patient, movement monitoring and analysis may be an effective
diagnostic tool. In particular, some types of seizures can be characterized by jerky
periodic movements of one or more body parts, usually limbs and head. On the other
hand, apneas are associated with the absence of periodic breathing movements [1].
A seizure can be defined as an age-dependent clinical event characterized by a
neurological dysfunction caused by paroxysmal alterations of neurological, beha-
vioral and/or automatic functions [2]. One of the most common outward effects is
an uncontrolled shaking due to involuntary and rapidly contraction and relaxation
1
Department of Engineering and Architecture, University of Parma, Parma, Italy
2
Department of Medicine and Surgery, University of Parma, Parma, Italy
*
Deceased on 16 September 2020
270 The Internet of medical things
of one or more muscle groups. Preterm and at-term newborns are more likely to
suffer from a seizure within 28 days after birth or 44 weeks of conceptional age,
respectively [2]. The estimated incidence is 2:6% for overall newborns, 2:0% for
term neonates, 11:1% for preterm neonates and 13:5% for infants weighing less
than 2 500 g [3]. Hypoxic–ischemic encephalopathy and stroke are not only two of
the more frequent etiologies, but brain malformation and infection can also be
triggers [4]. As reported in [4,5], several classifications have been proposed, but
usually four main types of clinical manifestations are considered indicative of
neonatal seizures: subtle, clonic, tonic and myoclonic [2]. Each clinical type of
seizure is characterized by distinguishable features and requires focused analysis
and diagnostic approach. Clonic seizures, for instance, are associated with rhythmic
and slow movements.
On the other hand, apneas can be defined as sudden interruptions of the
respiratory airflow. In newborns, these episodes are considered to be significant if
lasting longer than 20 s or less if associated with other symptoms, i.e., bradycardia
and cyanosis [6]. As reported in [1], among the main causes of neonatal apneas, we
recall seizures, cerebrovascular events [7] and congenital disorders, such as con-
genital central hypoventilation syndrome (CCHS) [8,9]. CCHS, in particular, is a rare
life-threatening disease caused by a defect in the PHOX2B homeobox gene [9]. It
mainly occurs during sleep and is responsible of alveolar hypoventilation. It is
usually associated with cyanosis, apnea or cardiorespiratory arrest [10]. Finally, three
main categories of apneas can be identified, i.e., central, obstructive and mixed,
according to the presence or the lack of an obstruction of the upper airway [6].
Due to the severity of these neonatal disorders, early treatments are needed to
prevent life-threatening episodes as well as lifelong consequences. To this purpose,
efficient monitoring tools must be deployed. The investigation of modern mon-
itoring systems based on video processing solutions can be considered a promising
and effective alternative to conventional equipment. Motion analysis plays a key
role to detect anomalous movements related to the aforementioned disorders and
will be discussed in the next sections.
airflow and the oxygen saturation of the blood, respectively [11]. Cameras can also
be employed for simultaneous traditional monitoring. A schematic overview of the
polysomnogram test is shown in Figure 14.1(a) and an example of recorded data is
reported in Figure 14.1(b) where the first four traces are the EEG channels, and the
subsequent traces are, from top to bottom, snoring noise, nasal flow, thoracic
movements and oxygen saturation. The abnormal breathing pattern is characterized
by recurrence of central apneas (closed boxes), in the absence of airway obstruction
and snoring. Central apneas determine severe oxygen desaturations.
Besides being expensive and moderately invasive, especially for newborns,
these techniques are almost exclusively deployed in clinical settings and require
trained medical staff who may not be available full time. To make home care more
accessible, various monitoring systems have been developed, e.g., smart bed [12]
and wearable-sensor-based systems [13], but they still require contact with the body
of the patient. Contactless solutions, on the other hand, may be devised for the
automatic detection of anomalous activities potentially related to neonatal dis-
orders. To this purpose, digital cameras can be used to frame the movements of a
patient to be analyzed through proper video processing algorithms. The integration
of these novel approaches allows to enhance both hospital and home constant
monitoring by providing low-cost preliminary alert signals to be possibly further
investigated by conventional diagnostic tools, i.e., the EEG.
Sensors to
measure brain activity,
eye movements and
muscle activity
Sensors to
measure air flow
and breath
Sensor to
measure oxygen
in blood
(a)
(b)
periodic signals, where the fundamental frequency represents the main unknown
parameter to be estimated.
In the following analysis, we will consider video sequences with sampling
period, T , where frames have dimension W H and are sampled at time instants
iT , i being the frame number.
Where I½x; y; i represents the ½x; y entry of matrix I½i that describes the ith
frame after a proper processing procedure. The motion signal in (14.1) can be
modeled as
¼ c þ Acosð2p f0 iT þ jÞ þ n½i
L½i (14.2)
Where c is a continuous component and n½i are samples of independent
identically distributed zero-mean Gaussian noise. The unknown parameters A, f0
and j represent the amplitude, phase and frequency, respectively, of the periodic
signal and may be collected in a vector q ¼ ½A; f0 ; f. The ML approach can now be
exploited to estimate the vector q. In particular, observing a window of N frames
and following standard methods described in [20], an estimator of the fundamental
frequency can be obtained and expressed as
2
XN 1
b j2pfiT
f 0 ¼ argmaxf L½ie : (14.3)
i¼0
b >h
NA
2
(14.5)
Where the value of the threshold h may be determined by trial and error.
that may be occluded for a single camera. Considering S sensors, a set of motion
signals is defined as in (14.1):
XW X H
s ½i ¼ 1
L Is ½x; y; i; s ¼ 1; 2; . . .; S: (14.6)
WH x¼1 y¼1
Where the processed ith frame for the sth sensor is described by matrix
Is ½x; y; i. The model in (14.2) can be generalized as
s ½i ¼ cs þ As cosð2p f0 iT þ js Þ þ ns ½i
L (14.7)
Where the sampling period T and the fundamental frequency f0 are assumed to
be identical for each capturing device; where present, the subscript s refers to the
sth sensor. Following the same procedure of the single-sensor analysis and
exploiting now data fusion techniques to combine data acquired by different sen-
sors, an estimator of the fundamental frequency can now be formulated as
2
X S NX1
b s ½iej2pfiT :
f 0 ¼ argmaxf L (14.8)
s¼1 i¼0
NX S
b 2 > h:
A (14.10)
S s¼1
Video DoF
input
z –1
X[i] D[i] Binarization
Spatial
L[i] I[i] B[i]
average
Erosion
Figure 14.3 Results of each processing step: (a) gray-scale, (b) difference
filtering, (c) binarization and (d) erosion
basic image filtering operation and the result is threshold to obtain a binary mask,
where white pixels correspond to foreground regions. Finally, the erosion morpholo-
gical operation [20] is implemented to reduce noise as discussed in [16]. Examples of
frames at each processing step are shown in Figure 14.3. Eventually, an example of
periodic movements induced by a clonic seizure and extracted according the procedure
illustrated in Figure 14.2 is shown in Figure 14.4, where the extracted average motion
signal is plotted against the frame number along with a corresponding EEG signal [16].
The two signals exhibit a comparable periodicity.
Average
motion
signal
i [Frame number]
1 1
D = 0.14
0.9 Camera 1
0.8 0.9 Camera 2
Camera 3
0.7
Camera 1 and 2
0.6 0.8 Camera 1 and 3
AUC = 0.95
0.5 Camera 2 and 3
α
Figure 14.5 Performance analysis for (a) single RGB camera (ROC curve); (b)
different RGB camera configurations [1]
results obtained using a single RGB sensor are illustrated in Figure 14.5(a), where
the receiver operating characteristic (ROC) [21] curve is plotted as a function of a
and 1 b for various values of the threshold h. We recall that an optimal predictor
is characterized by a ¼ 1 and b ¼ 1, i.e., all seizures are correctly detected when
present, and an area under the curve (AUC) is equal to 1. In the presented example,
the minimum Euclidean distance D from the ideal configuration is 0.14 and the
AUC is 0.95, which indicates high reliability [21]. In Figure 14.5(b), sensitivity and
specificity values are plotted for optimal values of h and different RGB cameras
configurations. The best performance is achieved when all three sensors are
employed, i.e., S ¼ 3 in (14.6)–(14.10). Depth sensors could also be employed to
better distinguish pathological movements from background noise or random
movements [1].
Video-based solutions for newborn monitoring 277
P0
P0 ϒ0 α0
Temporal BP-IIR filter
(pixel-by-pixel)
decomposition
reconstruction
P1
Spatial
Spatial
P1 ϒ1 α1
X[i]
Video Video
input input
1 0.03
Average luminance signal
0.025
0.8
Periodogram of L
0.02 ^
0.6 f : estimated frequency
0.015
0.4
0.01
0.2 0.005
0 0
0 0.5 1 1.5 2 2.5 3
0 100 200 300 400 500 600 700
Frame (b) Frequency [Hz]
(a)
Figure 14.7 (a) Periodic motion signal example and (b) periodogram
278 The Internet of medical things
0.8
Average luminance signal
0.6
0.4
0.2
0
0 2 4 6 8 10 12 14 16 18
Time [s]
×10−3
400 4
200 2
100
0
0
–100
–200 Pneumogram
Motion signal –2
–300
0 5 10 15 20
Time [s]
fundamental frequency is highlighted at the peak of the function. On the other hand,
an example of an anomalous motion signal is illustrated in Figure 14.18, where the
sudden interruption of the respiration caused by an apnea episode is visible in the
flat central part of the plot.
Finally, for the sake of comparison, an instance of an extracted motion signal is
shown in Figure 14.9 along with the equivalent signal obtained from a pneumograph,
where every period of the pneumographic signal corresponds to a complete respira-
tory act of the patient. Considering that a respiratory act is composed by two phases,
i.e., inhalation and exhalation, a good match of the two signals can be observed.
A number of improvements are possible with reference to the method descri-
bed in Figure 14.6. Among them, we mention [23] where the computationally
intensive reconstruction of the video stream is avoided and the sublevel signals are
directly combined in a multidimensional effective estimator. Direct application of
Video-based solutions for newborn monitoring 279
×10–3
1
6 TD = 30 s
TD = 20 s
TD = 10 s
4 TD = 0 s
2
20 s
0
0 10 20 30 40 50 60
Apnea onset Time [s] Observation window
Negative test
Positive test
14.6 Conclusion
In this chapter, novel techniques for newborn monitoring based on video processing
solutions have been proposed. Considering that disorders such as seizures and apneas
280 The Internet of medical things
References
[1] Cattani L, Alinovi D, Ferrari G, et al. Monitoring infants by automatic video
processing: A unified approach to motion analysis. Computers in Biology
and Medicine. 2017; 80: 158–165.
[2] Volpe JJ. Neurology of the Newborn. 5th ed. Philadelphia, PA, USA:
Saunders Elsevier; 2008.
[3] Ronen GM, Penney S, and Andrews W. The epidemiology of clinical neo-
natal seizures in Newfoundland: A population-based study. The Journal of
Pediatrics. 1999; 134(1): 71–75.
[4] Pellegrin S, Munoz FM, Padula M, et al. Neonatal seizures: Case definition
and guidelines for data collection, analysis, and presentation of immuniza-
tion safety data. Vaccine. 2019; 37(52): 7596–7609.
[5] Riviello JJ. Classification of seizures and epilepsy. Current Neurology and
Neuroscience Reports. 2003; 3: 325–331.
[6] Mishra S, Agarwal R, Jeevasankar M, et al. Apnea in the newborn. Indian
Journal of Pediatrics. 2008; 75: 57–71.
[7] Tramonte J and Goodkin H. Temporal lobe hemorrhage in the full-term neonate
presenting as apneic seizures. Journal of Perinatology. 2004; 24: 726–729.
[8] Rand CM, Patwari PP, Carroll MS, et al. Congenital central hypoventilation
syndrome and sudden infant death syndrome: disorders of autonomic reg-
ulation. Seminars in Pediatric Neurology. 2013; 20(1): 44–55. Pediatric
Autonomic Disorders.
[9] Healy F and Marcus CL. Congenital central hypoventilation syndrome in
children. Paediatric Respiratory Reviews. 2011; 12(4): 253–263.
[10] Cielo CM and Marcus CL. Central hypoventilation syndromes. Sleep
Medicine Clinics. 2014; 9(1): 105–118.
[11] Martin RJ, Block AJ, Cohn MA, et al. Indications and standards for cardi-
opulmonary sleep studies. Sleep. 1985; 8(4): 371–379.
[12] Spillman Jr WB, Mayer M, Bennett J, et al. A smart bed for non-intrusive
monitoring of patient physiological factors. Measurement Science and
Technology. 2004; 15(8): 1614–1620.
[13] Pantelopoulos A and Bourbakis NG. A survey on wearable sensor-based
systems for health monitoring and prognosis. IEEE Transactions on Systems,
Video-based solutions for newborn monitoring 281
Over the past few years, there have been numerous advances in the health-care
industry. Incorporation of Internet of Things (IoT) in the health-care applications
has unlocked numerous possibilities in the way healthcare operates now. This
chapter concentrates on IoT sensor networks in healthcare. The suitability of sensor
networks in healthcare is presented. The applications in which IoT sensor networks
play a crucial role, and thus related sensors are presented. Key sensors, and their
applications in healthcare, are presented. The most common and popular use cases
of IoT sensor networks in healthcare are discussed. The wireless communication
technologies used in IoT sensor networks for healthcare are described along with
their characteristics. Even though health-care IoT (H-IoT) networks have been in
use for some time, and their incorporation in health-care services is only expected
to increase further over the years, their implementation and adoption is still a
challenge. This chapter discusses those challenges in detail. The potential con-
temporary technologies that help in overcoming these challenges are explained.
List of abbreviations
3D three dimensional
EHR electronic health record
GPS global positioning system
H-IoT health-care Internet of Things
IoT Internet of Things
IR infrared
IrDA Infrared Data Association
LAN local area network
MEMS microelectromechanical systems
NFC near-field communication
NFV network function virtualization
1
Department of Computer Science and Engineering, Ramaiah University of Applied Sciences,
Bengaluru, India
284 The Internet of medical things
15.1 Introduction
In the past decades, the interaction of doctors with their patients remained limited to
the patients physically visiting the doctors/nurses and vice versa. With advances in
communication technologies and ease of communication, this interaction was
advanced to the form of tele- and text communications. The health-care technology
evolved further with the introduction of electronic health records (EHRs) that made it
possible for the hospitals to store and access the patient-centric information. The portal
technology allowed patients to log in to the health-care service provider’s website
taking a more active role by keeping a check on their health, accessing their medical
records and tracking their appointments. The cloud computing and big data technol-
ogies are driving the innovation into this field. However, the solution to attain constant
remote monitoring of the patient’s health and vital parameters was made possible
through the Internet of Things (IoT) technology that is powered by the wireless sensor
networks (WSN). Sensors convert physical parameters into signals that can be pro-
cessed further by other devices. When used in the form of wearable devices, these
sensors can be used to check and monitor the physiological parameters of the patient.
The values of these parameters include body temperature, heartbeat, pulse rate, blood
pressure, oxygen saturation and blood sugar. With the help of these sensor-based IoT
devices, real-time patient health monitoring and clinical feedback have become a
possibility. The rising adoption of wearable devices, emergence of connected care
solutions and implementation of digital technologies for healthcare are the main fac-
tors causing the surge in the demand for digital health-care solution and the expansion
of the digital health-care market. The IoT-based health-care market is segmented into
medical devices, system and software and medical services. The medical devices can
be wearable external devices or implanted medical devices, as well as stationary or
mobile medical devices used in the hospitals or health-care centers.
The WSN and IoT-based devices are majorly used for remote patient mon-
itoring, telemedicine, telesurgery, medical imaging, health-care apparatus and
medication management, smart health-care equipment, connected devices and
clinical operations. According to the survey carried out by Global Market Insights,
the global digital health market is increasing exponentially and is predicted to be a
500 billion dollar industry by the year 2025 [1]. According to the report published
by the “Data-Driven Investor” [2], the estimates show that the inclusion of WSN
and IoT-based devices in the health-care industry has saved the health-care industry
IoT sensor networks in healthcare 285
Low-power consumption
devices can be embedded in a variety of medical equipment not only at hospitals and
clinics but also at homes. Sensors meant for medical applications are used to monitor
physical and physiological parameters of the patient for detection, diagnosis, treat-
ment and posttreatment monitoring and management of patient’s health [3,4]. Some
of the popularly used sensors for medical applications are presented in Table 15.1.
Apart from the sensors, encoders are used in magnetic resonance imaging
machines, surgical robots, medical imaging, tomography, X-ray machines and
other critical/noncritical medical devices.
While these sensing devices have become popular for use in medical equip-
ment, they are constrained in resources such as computation and processing power,
battery power, storage/memory and supporting data rate/bandwidth [5]. The dis-
tinct features that have led to the rise of wireless sensing and communication
technology in healthcare are as follows:
1. Microelectromechanical systems (MEMS):
Advances in the MEMS technology have made it possible to implement and
deliver inexpensive remote patient monitoring and diagnostic capabilities. This
has made the early detection and curing of critical medical conditions. This has
1. Remote patient monitoring: Wearable devices and sensors are used to monitor
the vital parameters of the patient. This information can be transmitted wire-
lessly to the doctor’s/nurse’s device or to the cloud from where it can be read by
the medical staff involved in taking care of the patient. Such a facility allows the
patient to move around instead of sitting at one place and allows doctors to
monitor patient’s vitals without requiring to visit the patient often [8].
2. Fitness and activity trackers: People wearing fitness and activity trackers
mounted on wrist, ankle or belt is a common sight these days. These trackers
check the physical activity of the user such as steps taken, stairs climbed and
vital health parameters such as pulse, heartbeat, body temperature and other
basic physiological parameters. This data is stored on the cloud and users can
study it over time to track their fitness and activity levels [9].
IoT sensor networks in healthcare 289
Remote medication
H-IoT use cases
Robotic surgery
Drug management
Body scanning
By using Wi-Fi technologies, the medical and IoT devices all over the campus
can be connected for reliable communication by virtue of their communication
range and supported data rate. Wi-Fi supports device mobility over the network,
thus providing a dynamic network environment.
Zigbee is a popular WSN technology that uses unlicensed wireless spectrum.
Various medical sensors make use of Zigbee for communication. In H-IoT, the
most common use of Zigbee is in remote patient monitoring. Body sensor network
is used for patient health monitoring, and fitness and activity trackers make the use
of Zigbee sensors. Zigbee provides an appropriate balance of parameters such as
data rate, communication range and transmission power to make it the chosen
technology for these applications.
Bluetooth technology supports both voice and data, provides 1 Mbps of data
rate and uses unlicensed spectrum for communication. These features of Bluetooth
make it suitable choice for H-IoT. In a hospital campus, Bluetooth devices can be
identified and visitors can be informed about the floor plan, doctors on duty and
appointment scheduling through Bluetooth beacons. Bluetooth can be used for
automated patient check-in and checkout, optimized patient flow, scheduling
appointments with doctors, compliance tracking and recording, asset tracking and
wayfinding.
RFID is a short-distance, noncontact communication technology, which does
not require a direct line of sight communication. It is an economical and reliable
technology used for location and identification of objects. In H-IoT applications,
RFID is extensively used to identify, locate and track medical equipment.
Infrared Data Association (IrDA) is a very short-distance, point-to-point,
line of sight, which involves very low bit rate wireless optical communication
technology. This is used to carry out physically secured data transmission. IR
technology is used in remote control of appliances such as television, projector, air
conditioner and such other devices. In healthcare, IrDA can be used for remotely
controlling different medical devices.
Ultra-wideband (UWB) is a short-distance communication technology that
provides data rates of up to 450 Mbps. This uses low-power pulses for commu-
nication and generates low electromagnetic radiation, thus making it suitable for
medical applications. UWB radar is used for the monitoring of patient’s motion
over short distances, real-time exchange of medical images (such as X-rays, car-
diology-, pneumology-, obstetrics-, ear, nose, throat-imaging) and data over a dis-
tance of up to 10 m.
section discusses the challenges involved in the integration of H-IoT sensor net-
work in healthcare.
the information, which can lead to life critical situations. An entire hospital
network can be hacked through the infamous ransomware attack. Authors in
[30] have developed a framework to test the security of IoT devices. An
intrusion detection system for H-IoT networks is presented in [31].
10. Cost: While going digital and using technology in healthcare has brought down
the cost of physical visits for both patients and health-care professionals, as
well as providing ease of interaction/consultation, and patient health observa-
tion, the IoT sensor network technology is still out of reach for many. While
IoT-based healthcare is fascinating and promising, its implementation is yet to
overcome the cost considerations. The successful development, implementation
and optimization of these technologies are still an issue and require setting up a
costly infrastructure to ensure reliable connectivity and communication. Some
of the low-cost smart health-care solutions are presented in [32,33].
With the evolution of IoT and other mobile and wireless communication technol-
ogies, traditional computer networking technologies also need to be evolved, in
order to cater to the demand of contemporary applications such as IoT sensor
networks for healthcare. This section presents some of those contemporary tech-
nologies being used extensively in IoT sensor networks for healthcare.
• Data warehousing
• Remote data access
Cloud data centers • Big data processing and
(in thousands) analytics
• Al and ML
• Centralized services
• Computational nodes
• Gateway nodes
• On local networks
Fog nodes • Data analysis and
(in millions) reduction
• Virtualization
• Correlation and control
Figure 15.3 Edge, fog and cloud for H-IoT sensor networks
devices in the fog network have high compute power compared to the edge network
devices. These are installed as a backbone to the cloud networks. These devices are
located five to six hops away from the sensor nodes and support high-speed con-
nectivity and guaranteed connectivity to the edge devices. Some of the advantages
of fog computing are as follows:
1. faster data computing and processing
2. reduced latency
3. enhanced data security
4. setter network and data control
5. support for online and offline data access
6. decentralized data storage and processing
By adding layers of edge and fog nodes, the network load is partitioned to run
at the optimal network level.
Cloud computing provides enhanced scalability and performance to the IoT
networks. As the end H-IoT devices generate huge amounts of data every second, it is
difficult for the enterprises and health-care organizations to store such huge data for
future access. Due to the issues such as node mobility, node’s geographic location,
latency, network bandwidth and reliability, it is neither efficient nor advisable to run
all the applications in the cloud. That is where edge and fog computing play an
important role. Cloud comprises data centers that are capable of data warehousing, big
data processing and making the data available for future remote access. Cloud pro-
vides Infrastructure as a Service for better H-IoT scalability, storage and processing
power. Some of the benefits of cloud computing for H-IoT networks are as follows:
1. data storage,
2. access to data remotely for future use,
3. remote data processing,
300 The Internet of medical things
4. data security,
5. reduced infrastructure cost at the organization,
6. present-day cloud solutions capable of applying artificial intelligence, machine
learning and big data techniques on the IoT data.
Data plane
Data generating IoT devices and
network devices such as switches
Control plane
controler and cloud
Application plane
IoT sensor network healthcare
applications
the data is acquired from sensors and other IoT devices through the wireless/wired
network infrastructure. The received data is routed and forwarded to the network
controller (in the control plane) through the data plane.
Control plane: Control plane is the decision-making plane. The main entity of
this plane is the network controller that comprises the logic written for traffic control
such as routing, load balancing and traffic engineering. The controller fetches,
maintains different network information, such as network state, topology and statis-
tics, and makes network infrastructure control decisions based on this information.
The network controller receives application requirements from the application layer
and makes decision regarding QoS and dynamic bandwidth allocation. It has an
abstract view of the network, the events in the network and network statistics.
The SDN can be combined with the cloud for better network management and
network data access. This removes the barrier associated with network hardware
access and builds the network implementation cost-effective and agile. As seen
from Figure 15.4, the data from the network controller is uploaded onto the cloud.
Similarly, the network controller also fetches the data from the cloud and transmits
it to the end devices.
Application plane: Application plane presented in Figure 15.4 comprises all
the healthcare-related applications that work with the H-IoT data. The application
plane comprises the receiving applications that obtain the data originated at the
patient. This data is used for diagnosis, analytics, generate statistics and makes
appropriate decisions. As seen from Figure 15.4, the decisions made or messages
can be conveyed back to the patient through the cloud and network controller.
15.7 Conclusion
This chapter presents the role, significance, applications, supporting technologies,
challenges faced and contemporary technologies to overcome those challenges in IoT
sensor network for healthcare. IoT sensor networks play a significant role in
healthcare and have become popular in the health-care industry over the past few
years. According to different surveys and estimates, the use of IoT sensor networks in
healthcare will grow exponentially in the future. This chapter discusses the sensors
used in different healthcare-related applications. The communication technologies
facilitating the H-IoT networks and their characteristics are discussed. Despite
growing popularity of IoT sensor networks in healthcare, there are numerous chal-
lenges in the implementation and realization of H-IoT. These challenges are dis-
cussed in detail and the role of contemporary technologies such as edge/fog/cloud
computing and SDN to overcome some of these challenges is highlighted.
References
[1] https://www.businessinsider.in/science/news/iot-healthcare-in-2020-compa-
nies-devices-use-cases-and-market-stats/articleshow/74126142.cms: acces-
sed on 7 January 2021.
302 The Internet of medical things
[2] https://medium.com/datadriveninvestor/7-staggering-stats-on-healthcare-iot-
innovation-fe6b92774a5c: accessed on 7 January 2021.
[3] Sharma R., Gupta S.K., Suhas K.K. and Kashyap G.S., 2014, April.
Performance analysis of Zigbee based wireless sensor network for remote
patient monitoring. In 2014 Fourth International Conference on
Communication Systems and Network Technologies (pp. 58–62). IEEE.
[4] Al-Khafajiy M., Baker T., Chalmers C., et al., 2019. Remote health mon-
itoring of elderly through wearable sensors. Multimedia Tools and
Applications, 78(17), pp. 24681–24706.
[5] Sharama R., Shankar J.U. and Rajan S.T., 2014. Effect of number of active
nodes and inter-node distance on the performance of wireless sensor net-
works. In 2014 Fourth International Conference on Communication Systems
and Network Technologies (pp. 69–73). IEEE.
[6] Juneja S., Juneja A. and Anand R., 2020. Healthcare 4.0-digitizing health-
care using big data for performance improvisation. Journal of
Computational and Theoretical Nanoscience, 17(9 10), pp. 4408–4410.
[7] Tarouco L.M.R., Bertholdo L.M., Granville L.Z., et al., 2012. Internet of
Things in healthcare: interoperability and security issues. In 2012 IEEE
International Conference on Communications (ICC) (pp. 6121–6125). IEEE.
[8] Marakhimov A. and Joo J., 2017. Consumer adaptation and infusion of
wearable devices for healthcare. Computers in Human Behavior, 76,
pp. 135–148.
[9] Philip V., Suman V.K., Menon V.G. and Dhanya K.A., 2017. A review on
latest internet of things based healthcare applications. International Journal
of Computer Science and Information Security, 15(1), p. 248.
[10] Lomotey R.K., Pry J. and Sriramoju S., 2017. Wearable IoT data stream
traceability in a distributed health information system. Pervasive and Mobile
Computing, 40, pp. 692–707.
[11] Shabana N. and Velmathi G., 2018. Advanced tele-surgery with IoT
approach. In Intelligent Embedded Systems (pp. 17–24). Springer,
Singapore.
[12] Liu L. and Jia W., 2010, September. Business model for drug supply chain
based on the internet of things. In 2010 2nd IEEE International Conference
on Network Infrastructure and Digital Content (pp. 982–986). IEEE.
[13] Wilden J., Chandrakar A., Ashok A. and Prasad N., 2017. IoT based wear-
able smart insole. In 2017 Global Wireless Summit (GWS) (pp. 186–192).
IEEE.
[14] Muneer A., Fati S.M. and Fuddah S., 2020. Smart health monitoring system
using IoT based smart fitness mirror. Telkomnika, 18(1), pp. 317–331.
[15] Raundale P., Gadagi S. and Acharya C., 2017. IoT based biomedical waste
classification, quantification and management. In 2017 International
Conference on Computing Methodologies and Communication (ICCMC)
(pp. 487–490). IEEE.
[16] Patel A.R., Patel R.S., Singh N.M. and Kazi F.S., 2017. Vitality of robotics
in healthcare industry: an Internet of Things (IoT) perspective. In Internet of
IoT sensor networks in healthcare 303
Things and Big Data Technologies for Next Generation Healthcare (pp.
91–109). Springer, Cham.
[17] Chen M.Y., Lughofer E.D. and Polikar R., 2018. Big data and situation-
aware technology for smarter healthcare. Journal of Medical and Biological
Engineering, 38, pp. 845–846.
[18] Manogaran G., Lopez D., Thota C., Abbas K.M., Pyne S. and Sundarasekar
R., 2017. Big data analytics in healthcare Internet of Things. In Innovative
Healthcare Systems for the 21st Century (pp. 263–284). Springer, Cham.
[19] Salahuddin M.A., Al-Fuqaha A., Guizani M., Shuaib K. and Sallabi F., 2017.
Softwarization of internet of things infrastructure for secure and smart
Healthcare. Computer, 50(7), pp. 74–79. doi:10.1109/MC.2017.195.
[20] Rinki S., 2014. Simulation studies on effects of dual polarisation and
directivity of antennas on the performance of MANETs (Doctoral disserta-
tion, Ph.D. thesis, Coventry University, UK).
[21] Sharma R., Kadambi G.R., Vershinin Y.A. and Mukundan K.N., 2015.
Multipath routing protocol to support dual polarised directional communication
for performance enhancement of MANETs. In 2015 Fifth International
Conference on Communication Systems and Network Technologies (pp.
258–262). IEEE.
[22] Sharma R., Kadambi G.R., Vershinin Y.A. and Mukundan K.N., 2015. Dual
polarised directional communication based medium access control protocol
for performance enhancement of MANETs. In 2015 Fifth International
Conference on Communication Systems and Network Technologies
(pp. 185–189). IEEE.
[23] Jha D.N., Michalák P., Wen Z., Ranjan R. and Watson P., 2019.
Multiobjective deployment of data analysis operations in heterogeneous IoT
infrastructure. IEEE Transactions on Industrial Informatics, 16(11),
pp. 7014–7024.
[24] Ullah F., Habib M.A., Farhan M., Khalid S., Durrani M.Y. and Jabbar S.,
2017. Semantic interoperability for big-data in heterogeneous IoT infra-
structure for healthcare. Sustainable Cities and Society, 34, pp. 90–96.
[25] Sodhro A.H., Malokani A.S., Sodhro G.H., Muzammal M. and Zongwei
L., 2020. An adaptive QoS computation for medical data processing in
intelligent healthcare applications. Neural Computing and Applications,
32(3), pp. 723–734.
[26] Gia T.N., Rahmani A.M., Westerlund T., Liljeberg P. and Tenhunen H.,
2015. Fault tolerant and scalable IoT-based architecture for health monitor-
ing. In 2015 IEEE Sensors Applications Symposium (SAS) (pp. 1–6). IEEE.
[27] Piyare R. and Lee S.R., 2013. Towards Internet of Things (IoTs): integration
of wireless sensor network to cloud services for data collection and sharing.
International Journal of Computer Networks & Communications, 5(5),
pp. 59–72.
[28] Hathaliya J.J. and Tanwar S., 2020. An exhaustive survey on security and
privacy issues in Healthcare 4.0. Computer Communications, 153, pp. 311–335.
304 The Internet of medical things
[29] Shakeel P.M., Baskar S., Dhulipala V.S., Mishra S. and Jaber M.M., 2018.
Maintaining security and privacy in health care system using learning based
deep-Q-networks. Journal of Medical Systems, 42(10), p. 186.
[30] Lally G. and Sgandurra D., 2018. Towards a framework for testing the
security of IoT devices consistently. In International Workshop on Emerging
Technologies for Authorization and Authentication (pp. 88–102). Springer,
Cham.
[31] Thamilarasu G., Odesile A. and Hoang A., 2020. An intrusion detection
system for internet of medical things. IEEE Access, 8, pp. 181560–181576.
[32] Gia T.N., Jiang M., Sarker V.K., et al., 2017. Low-cost fog-assisted health-
care IoT system with energy-efficient sensor nodes. In 2017 13th
International Wireless Communications and Mobile Computing Conference
(IWCMC) (pp. 1765–1770). IEEE.
[33] Ganesh G.R.D., Jaidurgamohan K., Srinu V., Kancharla C.R. and Suresh S.
V., 2016, December. Design of a low cost smart chair for telemedicine and
IoT based health monitoring: an open source technology to facilitate better
healthcare. In 2016 11th International Conference on Industrial and
Information Systems (ICIIS) (pp. 89–94). IEEE.
[34] Dang L.M., Piran M., Han D., Min K. and Moon H., 2019. A survey on
internet of things and cloud computing for healthcare. Electronics, 8(7),
p. 768.
[35] Ananth M.D. and Sharma R., 2017. Cost and performance analysis of net-
work function virtualization based cloud systems. In 2017 IEEE 7th
International Advance Computing Conference (IACC) (pp. 70–74). IEEE.
[36] Ananth M.D. and Sharma R., 2016. Cloud management using network function
virtualization to reduce CAPEX and OPEX. In 2016 8th International
Conference on Computational Intelligence and Communication Networks
(CICN) (pp. 43–47). IEEE.
[37] Ray P.P., Dash D. and De D., 2019. Edge computing for Internet of Things: a
survey, e-healthcare case study and future direction. Journal of Network and
Computer Applications, 140, pp. 1–22.
[38] Paul A., Pinjari H., Hong W.H., Seo H.C. and Rho S., 2018. Fog computing-
based IoT for health monitoring system. Journal of Sensors, 2018.
[39] Sharma R. and Reddy H., 2019. Effect of load balancer on software-defined
networking (SDN) based Cloud. In 2019 IEEE 16th India Council
International Conference (INDICON) (pp. 1–4). IEEE.
[40] Sharma R. 2021. 11 Software-defined networking for healthcare Internet of
Things. In Applications of Machine Learning in Big-Data Analytics and
Cloud Computing (pp. 231–247). River Publishers.
Chapter 16
Machine learning for Healthcare 4.0:
technologies, algorithms, vulnerabilities, and
proposed solutions
Saumya1 and Bharat Bhushan1
Healthcare 4.0 is motivated from Industrie 4.0. It is a boon to the present health-
care system due to its widespread applications that have boosted its efficiency and
enhanced its services. Healthcare 4.0 is a vision that integrates all the leading
technologies together given that each technology has different benefits to offer to
the system. Various technologies such as big data, Health Cloud (HC), Health Fog
(HF), Internet of Things (IoT), blockchain, and machine learning (ML) are incor-
porated in Healthcare 4.0. There are many applications of Healthcare 4.0 for
patients, health-care professionals, resource management, etc. This chapter aims to
study ML with respect to Healthcare 4.0 and highlights different ML algorithms
and its applications for healthcare in different phases such as prognosis, diagnosis,
treatment, and clinical workflow. ML provides many solutions for Healthcare 4.0
which have applicability for patients, health-care professionals, as well as health-
care facilities. However, there is some vulnerability for ML in healthcare that needs
to be checked. This chapter highlights these vulnerabilities and presents the
recently proposed solutions in this regard.
16.1 Introduction
Healthcare today is covered with many challenges and difficulties such as the high
cost of health-care services, lack of skilled professionals, demands of high-quality
services, lack of collaboration among different health-care service providers, and
high competition among them [1]. These things are raising questions on the relia-
bility of the system for the patients and their satisfaction. Therefore, there is a need to
look for new solutions to overcome these challenges [2]. The need of the hour is to
evolve the health-care system, one such sector evolving is Industry Sector 4.0 which
is developed from Industries 1.0, 2.0, and 3.0. The goal of Industry 4.0 is to build on
automated smart machines that make the system cost-effective, enhance operation,
1
School of Engineering and Technology (SET), Sharda University, Greater Noida, India
306 The Internet of medical things
and increase quality of services. These principles taken from Industry 4.0 are applied
to the health-care system and gave birth to Healthcare 4.0, without any existence of
Healthcare 1.0, 2.0, and 3.0. It aims to enhance the efficiency of health-care services,
including new technologies to increase the quality of services and still being cost-
effective for the patients. The benefits of Healthcare 4.0 are numerous such as
improve flexibility, reliability, scalability, and cost-effectiveness [3]. Apart from this,
it will also help in better management and response to health pandemics like COVID-
19 [4]. Despite its benefits, building such a versatile system for healthcare is chal-
lenging and complex within itself. Many factors are needed to be considered. But,
once built, it will boost its efficacy. Healthcare 4.0 is built on the integration of
human resources, expertise with technologies like blockchain, ML, HC, big data,
IoT, deep learning (DL), and artificial intelligence (AI).
The chapter discusses ML, its algorithms, and its applicability to Healthcare
4.0. ML is a vast field, which is growing every day. It has already transformed
many systems such as governance, transportation, and manufacturing. The use of
these technologies is carried in people’s day-to-day life either knowingly or
unknowingly. ML is now being incorporated into the field of healthcare, which is
influencing this system to provide various advantages [5]. Using ML models in
healthcare has many benefits like it will make the system less dependent on human
resources, fewer chances of errors and thus, increase in efficiency without manual
support. It can assist medical professionals through AI in learning information from
medical books, texts, etc. Medical tasks like monitoring, analyzing, and managing
reports of patient become efficient when embedding of ML with IoT devices is
done [6]. ML can also lead to the discovery of novel therapeutics through analyzing
of large biological databases and finding patterns within them. In near future, the
ML model will also support the radiologists as well as physicians and will boost
medical research and practice [7]. Kumari et al. [8] presented the transition of ML
toward cloud toward the fog and IoT-based healthcare. Multan et al. [9] also dis-
cussed growth in fog computing in the healthcare IoT domain. Mohanta et al. [10]
discussed shifting of paradigm toward healthcare using AI, IoT, and 5G technolo-
gies. Rong et al. [11] reviewed some of the applications using epileptic seizures and
filling of a dysfunctional urinary bladder by the use of AI. With increasing avail-
ability of clinically relevant databases, ML is applied to Healthcare 4.0 for diag-
nostic/identification tasks and prediction tasks [12,13]. In recent years, ML is
gaining importance across several disciplines and led to emerging breakthroughs in
Healthcare 4.0 and machine translation [14,15]. In summary, the major contribu-
tions of this chapter are enumerated as follows.
● This work provides an overview of Healthcare 4.0 and the major technologies
used in it.
● This work categorizes and discusses different applications of Healthcare 4.0.
● This work presents an in-depth analysis of ML algorithms with the main focus
on their applications for healthcare.
● This work presents security and other vulnerabilities involved in adopting ML
for healthcare.
Machine learning for Healthcare 4.0 307
● Finally, this work presents different solutions that can be used for ensuring the
security and efficacy of ML systems in Healthcare 4.0.
The remainder of the chapter is organized as follows. Section 16.2 presents
Healthcare 4.0 and the technologies involved in it, Section 16.3 presents an over-
view of different algorithms involved in ML that are used in healthcare,
Section 16.4 presents ML applicability in healthcare and different functions.
Section 16.5 presents the challenges faced during the deployment of ML in
Healthcare 4.0. Section 16.6 concerns with the solutions to enhance the security for
ML-based systems in Healthcare 4.0. Finally, Section 16.7 concludes the work.
● Blockchain: The concept is used to denote record in the form of blocks and is
very sensitive to any change in the whole chain that is formed using combi-
nation of blocks. Blockchain can facilitate various applications in Healthcare
4.0 such as protecting medical data, sensitive personal patients’ symptoms
stored in the form of data, EHRs data management, etc. Blockchain also pro-
vides security in sharing of data and protection of the patient’s privacy [39,42].
There are numerous applications of blockchain in healthcare. Wang et al. [43]
discussed a framework of parallel health-care systems based on the artificial
systems, including computational experiments together with parallel execution
(ACP) approach. Zarour et al. [44] discussed the impact of blockchain models
for Secure and Trustworthy Electronic Healthcare Records.
● Better accuracy and efficiency of the services: Employing all the technolo-
gies may also help in better maintenance of resources and on-time servicing of
the resources. This will increase the predictability and efficiency of the whole
system.
● Efficiently analyzing patients’ data (medical and nonmedical): Getting a
detailed view of patients’ data that can check other factors responsible for the
disease and can help the professionals in better prediction and creating more
personalized treatments.
● Efficient data analysis: Healthcare 4.0 can help efficient data collection and
transmission on a large scale, these data when analyzed for longer duration
could help in better generalizations as well as results like data of infectious
disease will tell the hotspots of that diseases, age groups affected, moreover it
will help in controlling the rate of infection at large scale.
● Smart decision-making supported by MCPS: Using MCPS in Healthcare
4.0, people can develop smart equipment and automation which can be utilized
to give better prediction and enhanced decision-making. This will also require
less staff and error could be minimized at certain levels, furthermore providing
the people with cost friendly schemes.
● Making a cost-effective system for all: There are many ways to achieve this
in Healthcare 4.0. For example, if people allow proper sharing of advanced,
costly resources and equipment across all health-care providers, the opera-
tional cost will become less and the resources will be utilized in a better way.
classify both linear and nonlinear data. It works by plotting each data item as a
point in n-dimensional space (where n is the number of features) with the value of
each feature being the value of a particular coordinate. Then, it performs classifi-
cation by finding the hyperplane that differentiates the two classes very well. Data
points are decided on the basis of boundaries which are overlaid by hyperplanes.
Data points that fall on any side of the given plane are attributed to various classes.
Moreover, the dimension highly depends upon the total number of features. For
example, the output would be only one line if in case there are two inputs to
machine. Support vectors are set of data that directs toward the hyperplane and
these will help users to maximize the margin of the classifier. For nonlinear data,
SVM uses kernel functions that take input as low dimension feature space and
outputs high-dimension feature space due to which classification becomes easy.
SVM classifiers like one versus all and one versus one are used for multiclass
classification.
16.3.7 K means
It is an unsupervised algorithm that comes under type of clustering technique.
Clustering in UL is an important approach as it uses patterns in a collection of
uncategorized data and finds natural clusters or groups within the data. K Means is
a subtype of types of clustering and comes under partitioned clustering. It is basi-
cally a centroid-based clustering. K Here represents a number of groups and the
algorithm starts from identifying the k centroids and then iterates between two steps
of assigning the data points and updating these centroids. Euclidean distance
measures the next step for updating centroids by taking the mean of all data points
assigned to a particular cluster.
Use case
Data annotation
ML model
Algorithms audit
Real world
survellience
User acceptance
Clinical integration
Regulatory
Multisite validation
practice, NLP techniques are used with other medical related software for analyzing
information from unstructured medical data [69]. ML model is also facilitating the
clinicians with applications for the processing of clinical speech as well as audio data,
which helps in saving their time and reducing workload. It also helps in the diagnosis of
disease related to speech such as dementia [70] and Alzheimer’s disease [71].
Data
collection
Data
annotation
Different
vulnerabilities Model
trainings
Model
depolyment
Testing
16.7 Conclusion
This chapter worked on a way to demonstrate the role of different technologies,
keeping the main focus on ML for Healthcare 4.0. Initially, the chapter defines
Healthcare 4.0 and its applications. Then, various types of ML algorithms are
highlighted. Further, it explores ML applications in healthcare, and their way of
implementation. The chapter also discussed the challenges faced while deploying
ML models in the health-care system. The chapter provides some solutions to
Machine learning for Healthcare 4.0 323
ensure the security, privacy, and robustness of ML models when implemented over
health-care system. ML is a vast field within itself and when this field is integrated
with other technologies, it produces great results. When used in the health-care
system, this field will boost efficiency and enhance its service to many subfields of
healthcare which consequently benefits the human beings.
References
[1] C. Kanchanachitra, M. Lindelow, T. Johnston et al., “Human resources for
health in southeast Asia: Shortages, distributional challenges, and interna-
tional trade in health services,” Lancet, vol. 377, no. 9767, pp. 769–781,
2011.
[2] M. Akay and T. Tamura, “Global healthcare: Advances and challenge
[Scanning the Issue],” Proc. IEEE, vol. 103, no. 2, pp. 147–149, 2015.
[3] G. Alfian, M. Syafrudin, M. Ijaz, M. Syaekhoni, N. Fitriyani, and J. Rhee,
“A personalized HEALTHCARE monitoring system for diabetic patients by
UTILIZING BLE-based sensors and real-time data processing,” Sensors,
vol. 18, no. 7, p. 2183, 2018. doi:10.3390/s18072183.
[4] M. B. Sampa, M. R. Hoque, R. Islam et al., “Redesigning portable health
clinic platform as a remote healthcare system to tackle COVID-19 pandemic
situation in unreached communities,” Int. J. Environ. Res. Public Health,
vol. 17, no. 13, p. 4709, 2020.
[5] M. Jindal, J. Gupta, and B. Bhushan, “Machine learning methods for IoT and
their future applications,” in 2019 International Conference on Computing,
Communication, and Intelligent Systems (ICCCIS), 2019. doi:10.1109/
icccis48478.2019.8974551.
[6] R. Tayal and A. Shankar, “Learning and predicting diabetes data sets using
semi-supervised learning,” in 2020 10th International Conference on Cloud
Computing, Data Science & Engineering (Confluence), Noida, India, 2020,
pp. 385–389. doi:10.1109/Confluence47617.2020.9058276.
[7] L. Xing, E. A. Krupinski, and J. Cai, “Artificial intelligence will soon change
the landscape of medical physics research and practice,” Med. Phys., vol. 45,
no. 5, pp. 1791–1793, 2018.
[8] A. Kumari, S. Tanwar, S. Tyagi, and N. Kumar, “Fog computing for
Healthcare 4.0 environment: Opportunities and challenges,” Comput. Electr.
Eng., vol. 72, pp. 1–13, 2018.
[9] A. A. Mutlag, M. K. A. Ghani, N. Arunkumar, M. A. Mohammed, and
O. Mohd, “Enabling technologies for fog computing in healthcare IoT systems,”
Future Gener. Comput. Syst., vol. 90, pp. 62–78, 2019, [online] Available:
http://www.sciencedirect.com/science/article/pii/S0167739X18314006.
[10] B. Mohanta, P. Das, and S. Patnaik, “Healthcare 5.0: A paradigm shift in
digital healthcare system using artificial intelligence, IOT and 5G commu-
nication,” in 2019 International Conference on Applied Machine Learning
(ICAML), 2019, pp. 191–196. doi:10.1109/ICAML48257.2019.00044.
324 The Internet of medical things
[38] H. Qiu, M. Qiu, M. Liu, and G. Memmi, “Secure health data sharing for
medical cyber-physical systems for Healthcare 4.0,” IEEE J. Biomed. Health
Inf., vol. 24, no. 9, pp. 2499–2505, 2020. doi:10.1109/JBHI.2020.2973467.
[39] J. Al-Jaroodi and N. Mohamed, “Blockchain in industries: A survey,” IEEE
Access, vol. 7, pp. 36500–36515, 2019.
[40] P. Verma and S. K. Sood, “Fog assisted-IoT enabled patient health mon-
itoring in smart homes,” IEEE Internet Things J., vol. 5, no. 3, pp. 1789–
1796, 2018. doi:10.1109/JIOT.2018.2803201.
[41] D. M. West, How 5G Technology Enables the Health Internet of Things,
Brookings Center for Technol. Innovation, vol. 3, pp. 1–20, 2016.
[42] N. Mohamed and J. Al-Jaroodi, “Applying blockchain in industry 4.0
applications,” in Proc. IEEE 9th Annu. Comput. Commun. Workshop Conf.
(CCWC), 2019, pp. 0852–0858.
[43] S. Wang, J. Wang, X. Wang, et al., “Blockchain-powered parallel healthcare
systems based on the ACP approach,” IEEE Trans. Comput. Soc. Syst.,
vol. 5, no. 4, pp. 942–950, 2018. doi:10.1109/TCSS.2018.2865526.
[44] M. Zarour, M. T. Ansari, M. Alenezi, et al., “Evaluating the impact of
Blockchain Models for secure and trustworthy electronic healthcare
records,” IEEE Access, vol. 8, pp. 157959–157973, 2020. doi:10.1109/
ACCESS.2020.3019829.
[45] V. Chandola, A. Banerjee, and V. Kumar, “Anomaly detection: A survey,”
ACM Comput. Surv. (CSUR), vol. 41, no. 3, p. 15, 2009.
[46] K. Polat and S. Gunes, “Prediction of hepatitis disease based on principal
component analysis and artificial immune recognition system,” Appl. Math.
Comput., vol. 189, no. 2, pp. 1282–1291, 2007.
[47] A. K. Pandey, P. Pandey, K. Jaiswal, and A. K. Sen, “Data Mining clustering
techniques in the prediction of heart disease using attribute selection
method,” Heart Disease, vol. 14, pp. 16–17, 2013.
[48] D. Mahapatra, “Semi-supervised learning and graph cuts for consensus based
medical image segmentation,” Pattern Recognit., vol. 63, pp. 700–709,
2017.
[49] W. Shen, M. Zhou, F. Yang, C. Yang, and J. Tian, “Multi-scale convolutional
neural networks for lung nodule classification,” in International Conference
on Information Processing in Medical Imaging. Springer, 2015, pp. 588–599.
[50] M. N. Sohail, J. Ren, and M. Uba Muhammad, “A Euclidean group assess-
ment on semi-supervised clustering for healthcare clinical implications
based on real-life data,” Int. J. Environ. Public Health, vol. 16, no. 9,
p. 1581, 2019.
[51] H.-C. Kao, K.-F. Tang, and E. Y. Chang, “Context-aware symptom checking
for disease diagnosis using hierarchical reinforcement learning,” in Thirty-
Second AAAI Conference on Artificial Intelligence, 2018.
[52] W. Bai, O. Oktay, M. Sinclair et al., “Semi- supervised learning for network-
based cardiac MR image segmentation,” in International Conference on
Medical Image Computing and Computer-Assisted Intervention. Springer,
2017, pp. 253–260.
Machine learning for Healthcare 4.0 327
The unprecedented growth of population in urban areas has been causing a chal-
lenge for the citizens in their day-to-day lives such as road congestion, public
security, environmental pollution, electricity shortage and water shortage. To
control and resolve all these issues, new technologies have been developed for
smart cities. Intelligent services and better applications are deployed in smart cities,
by combining the Internet of Things (IoT) with the technologies like data mining
(DM) and deep learning (DL). Many sectors like healthcare, governance, agri-
culture and public safety can increase their efficiency with the help of these new
technologies and can convert these into smart applications for smart cities.
Different kinds of computing like edge computing, fog computing and cloud
computing support to provide better insights into analytics with the help of big data
in smart cities. All these technologies are transforming or raising healthcare eco-
systems, leading them in the direction of smart healthcare. This permits surgeons to
get real-time data of their patients distantly with the help of wireless communica-
tion. Smart healthcare is established on new technologies to convey enriched and
valued healthcare facilities for patients. This chapter explores the current chal-
lenges that are faced during the indigenous development of the smart cities.
Furthermore, the chapter discusses the theoretical background of smart cities with
the explanation of their components. Moreover, the chapter describes the necessity
of computational infrastructure for smart cities in a framework of big data and DM.
The chapter highlights some mining methods for extracting important information
from huge and mixed data. Additionally, the chapter examines the advancement of
healthcare sector in smart cities in context of big data and DM.
17.1 Introduction
Today almost 55% of the population are living in the urban cities, and it is calcu-
lated that in the coming 30 years about 68% of the population will be living in
1
School of Engineering and Technology (SET), Sharda University, Greater Noida, India
332 The Internet of medical things
smart cities [1]. According to an estimation by the United Nations, it is said that the
population will increase by 2.5 billion by the year 2050. This huge amount of
population will pose a great challenge, as it is highly difficult for the development
and sustainable management of the urban cities areas [2]. It will also be a great
challenge for the people to provide a very good quality of life (QoL) for the citizens
of smart cities [3]. Therefore, it is very important to understand the need for urgent
and effective solutions to support various demands of the population that grow at a
very fast speed [4]. On the other side, the evolutions of the big data analytics
(BDA) and the IoT [5] are the major factors that are responsible for the highly
efficient implementation of the smart cities’ services [6]. The BDA and IoT have
gained immense popularity and attention as shown in Figure 17.1.
Across various domains, various intelligent systems have been developed with the
emerging technologies [7], which help us to make different domains that are highly
efficient. The benefit of smart cities is the processing and collection of data by a
different technology of computing, to provide better interconnection and communica-
tion [8]. This leads to a better service in the smart services by uplifting the security,
privacy, availability, sink and transportation. After looking at these issues, to support
the development of the smart cities, authorities have shown a great interest toward the
smart cities and invested billions of dollars [9]. IoT can be referred to as making the
interconnection between the infinite numbers of the smart devices [10]. This technol-
ogy helps the physical objects to enact as human and does their behaviors such as
decision-making, thinking, hearing, sensing and seeing in order to coordinate actions,
share information and communicate with other [11]. By these abilities, one can trans-
form these normally simple objects or physical devices into intelligent devices that can
operate on their own and adjust to the real-time circumstances and work peacefully
without any disturbance of a third member like humans [12].
BDA and DM concepts are specific to smart cities as a whole and especially to
smart healthcare systems in many ways [13]. Social media is reconstructing the health-
related communications and share medical information, by varying the way of keeping
the health records that evaluate by specialists. The availability of web resources on
social media feeds and online forums delivers very important facts [14]. Indeed, BDA
100 Internet
of things
80 Big data
analytics
60
40
20
0
1 9 5 1 9 5 1
0–0 1–0 3–0 5–0 6–0 8–0 0–0
201 201 201 201 201 201 202
Lazaroiu et al. [20] investigated the different aspects of energy consumption and energy
efficacy problems and were aware of the fact that cities are consuming 75% of the
world’s energy production capacities, and these cities are also responsible for generating
80% of carbon dioxide emissions. For a balanced use of the world’s precious energy
resources, a model is developed that helped us to resolve the issues with energy effi-
ciency. Some components of smart cities are discussed in the succeeding subsection [21].
constant monitoring of the trees in the smart cities that will help one to build a
better smart environment. It is further noticed that trees in the smart cities could
also damage the network cables, and it also occurs disruption in power connections,
and some researchers suggested an efficient way, which is by developing a dynamic
laser-based scanning system, results in locating well-organized trees in the smart
cities [44]. Castelli et al. [45] discussed that the pollution in the air can cause severe
damage to the environment, which many urban cities are already facing. A report
published by World Health Organization says many people are dying because of air
pollution every year. However, making it very necessary to keep an eye on the air
quality index management, this could be done through the various traffic mon-
itoring sensors installed at the certain location in the smart cities [46]. The water
management is a challenging issue if not managed properly, as it has been seen in
the smart cities and few factors remain as a challenge such as maintenance cost,
new contaminants, and aging water infrastructure. Some experts suggest that ICT
systems improved the quality of drinking water across the smart cities. Water
management is directly responsible for the QoL that people are spending, as it is
also important for the governance to keep the contaminants as low as possible [47].
Smart people [22] Smart economy [28] Smart governance [33] Smart mobility Smart environment Smart living [48]
[38] [43]
● Healthcare study ● Employment ● IT connectivity ● Transport ● Pollution ● Sewerage and
● Telecommunication opportunities ● Online citizen services ● Walkability ● Green building sanitation
● Entertainment ● Smart manufacturing ● Digital media ● Traffic ● Renewable energy ● Water supply
● Cultural activities ● Smart data protection ● Wireless ● Smart grids ● Electricities
● Smart and advanced networks ● Air quality ● Building and home
security ● Trees ● Public open space
● Smart agriculture
sustainability
338 The Internet of medical things
platforms. Moreover, platforms like fog, edge and cloud computing are used. Edge
computing and fog computing are specially designed as extensions for the enabled data
analytics and cloud network for the source of the new data creation [52]. These three
computing techniques are explained in detail in the subsections.
computing and cloud computing share the mutually similar service such as
deployment, networking and storage. Furthermore, fog computing technology is
only available in few geographical locations in the world, whereas cloud computing
is available everywhere without any complication of the geographical location.
Moreover, fog computing is specially designed for the use of real-time IoT appli-
cations. These types of computing help us with low latency, minimal issues with
privacy and security, and limited network bandwidth.
17.4.1 Classification
In classification, a process is followed through which an object is assigned to the
predefined category, classification techniques can predict the destination class of
Big data analytics and data mining 341
each object of the data. For supervised learning, the targeted labels are assumed
before processing. Before the classification of unknown data/objects or unlabeled
data, a classifier (prediction function) classifies the required training as most
algorithms are classified into two types. The first one is computing the probability
of an item whether it belongs to a particular class or not and the second is com-
paring it with the cut-off value [63]. The evaluation of the performance of classi-
fication model can be done as follows:
number of correct prediction Tp þ Fp
Accuracy ¼ (17.1)
total number of predictions Tp þ Fp þ Tn þ Fn
17.4.2 Clustering
A cluster can be defined as the objects present in a special identical group. With the
help of clustering algorithm, the clusters are segregated into a certain fixed number
of subclusters on the basis of their features. Clustering follows unsupervised
learning technique, for example, in few medical centers, certain patients have an
unknown disease but researchers of the medical field are able to uncover and
provide cure of those diseases as they are having little data related to the symptom
of disease and improvement made by the patients based on some treatments that are
provided by the doctors [65]. This clustering will help in a way to segregate the
patients on the basis of their recognized symptoms and progress, which will help
the doctors to perform a better and improved treatment.
17.6 Conclusion
Technologies like big data, ML, AI, DL and DM create a deep impact on people’s
lifestyle. The development of the smart cities is highly proportional to such tech-
nologies. All these technologies can be clubbed together with the IoT devices to
provide more efficiency. Many issues present in the real-world cities can be solved
with these technologies in the new upcoming smart cities. For example, the issues
like traffic management can be resolved with the fully automated robot system and
by designing it with a better infrastructure, which will increase the efficiency in
traffic management. Moreover, sustainable concepts of smart cities should carry
keywords like environment, land use, transportation and relations with each other.
The designer of the smart cities should have few aspects in his mind before
346 The Internet of medical things
designing the smart cities’ layout such as the population of the smart cities, size of
the smart cities, how much it is accessible to the service center of the prominent
company and how much accessible it is to the other cities that exist nearby. The
main aim of our study through this chapter was to spread the awareness among the
people that what are the key future trends of the smart cities and how we can build a
smart healthcare to provide better treatment with ease. The chapter has described a
theoretical background of smart cities in which some important components are
discussed. Furthermore, the chapter highlights the need of computational infra-
structure for smart cities in a correlation with BDA. Moreover, the chapter high-
lights some mining methods for mining of big as well as heterogeneous data.
Additionally, the chapter has described the advancement of healthcare sector in
smart cities in context of BD and DM techniques. In future, we enhance this study
in a way of providing more security and ease.
References
[1] Agbali M., Trillo C., Fernando T., Oyedele L., Ibrahim I. A., and Olatunji V. O.
(2019). Towards a refined conceptual framework model for a smart and sus-
tainable cities assessment. IEEE International Smart Cities Conference (ISC2),
pp. 658–664, doi:10.1109/ISC246665.2019.9071697.
[2] Suryanegara M., Prasetyo D. A., Andriyanto F., and Hayati N. (2019). A
5-step framework for measuring the quality of experience (QoE) of Internet
of Things (IoT) services. IEEE Access, 7, 175779–175792, doi:10.1109/
ACCESS.2019.2957341.
[3] Mishra K. N. and Chakraborty C. (2020). A novel approach towards using
big data and IoT for improving the efficiency of m-health systems. In: Gupta
D., Hassanien A., Khanna A. (eds), Advanced Computational Intelligence
Techniques for Virtual Reality in Healthcare. Studies in Computational
Intelligence, vol. 875. Cham: Springer. https://doi.org/10.1007/978-3-030-
35252-3_7.
[4] Gupta A., Chakraborty C., and Gupta B. (2019). Medical information pro-
cessing using smartphone under IoT framework. In: Mittal M., Tanwar S.,
Agarwal B., Goyal L. (eds), Energy Conservation for IoT Devices. Studies in
Systems, Decision and Control, vol. 206. Singapore: Springer. https://doi.
org/10.1007/978-981-13-7399-2_12.
[5] Mishra K. N. and Chakraborty C. (2020). A novel approach toward enhan-
cing the quality of life in smart cities using clouds and IoT-based technolo-
gies. In: Farsi M., Daneshkhah A., Hosseinian-Far A., Jahankhani H. (eds),
Digital Twin Technologies and Smart Cities. Internet of Things (Technology,
Communications and Computing). Cham: Springer. https://doi.org/10.1007/
978-3-030-18732-3_2.
[6] Gupta A., Chakraborty C., and Gupta B. (2019). Monitoring of epileptical
patients using cloud-enabled health-IoT system. Traitement Du Signal, 36(5),
425–431. doi:10.18280/ts.360507.
Big data analytics and data mining 347
[18] Bhushan B., Khamparia A., Sagayam K. M., Sharma S. K., Ahad M. A., and
Debnath N. C. (2020). Blockchain for smart cities: A review of architectures,
integration trends and future research directions. Sustainable Cities and
Society, 61, 102360. doi:10.1016/j.scs.2020.102360.
[19] Khamparia A., Singh P. K., Rani P., Samanta D., Khanna A., and Bhushan B.
(2020). An internet of health things-driven deep learning framework for
detection and classification of skin cancer using transfer learning.
Transactions on Emerging Telecommunications Technologies. doi:10.1002/
ett.3963.
[20] Lazaroiu G. C. and Roscia M. (2012). Definition methodology for the smart
cities model. Energy, 47(1), 326–332. doi:10.1016/j.energy.2012.09.028.
[21] Sethi R., Bhushan B., Sharma N., Kumar R., and Kaushik I. (2021).
Applicability of industrial IoT in diversified sectors: Evolution, applications
and challenges. In: Kumar R., Sharma R., Pattnaik P.K. (eds), Multimedia
Technologies in the Internet of Things Environment, Studies in Big Data,
vol. 79. SIngapore: Springer. https://doi.org/10.1007/978-981-15-7965-3_4.
[22] Malik A., Gautam S., Abidin S., and Bhushan B. (2019). Blockchain
technology-future of IoT: Including structure, limitations and various pos-
sible attacks. 2nd International Conference on Intelligent Computing,
Instrumentation and Control Technologies (ICICICT), Kannur, India, 2019,
pp. 1100–1104, doi:10.1109/ICICICT46008.2019.8993144.
[23] Yeh H. (2017). The effects of successful ICT-based smart cities services:
From citizens’ perspectives. Government Information Quarterly, 34(3), 556–
565. https://doi.org/10.1016/j.giq.2017.05.001.
[24] Belanche-Gracia D., Casaló-Ariño L. V., and Pérez-Rueda A. (2015).
Determinants of multi-service smartcard success for smart cities develop-
ment: A study based on citizens’ privacy and security perceptions.
Government Information Quarterly, 32(2), 154–163. https://doi.org/10.1016/
j.giq.2014.12.004.
[25] Chatterjee S. and Kar A. K. (2017). Effects of successful adoption of
information technology enabled services in proposed smart cities of India.
Journal of Science and Technology Policy Management, 9(2), 189–209.
https://doi.org/10.1108/jstpm-03-2017-0008.
[26] Niforatos E., Vourvopoulos A., and Langheinrich M. (2017). Understanding
the potential of human–machine crowdsourcing for weather data.
International Journal of Human-Computer Studies, 102, 54–68. https://doi.
org/10.1016/j.ijhcs.2016.10.002.
[27] Mone G. (2015). The new smart cities. Communications of the ACM, 58(7),
20–21. https://doi.org/10.1145/2771297.
[28] Gautam S., Malik A., Singh N., and Kumar S. (2019). Recent advances and
countermeasures against various attacks in IoT environment. 2019 2nd
International Conference on Signal Processing and Communication
(ICSPC). https://doi.org/10.1109/icspc46172.2019.8976527.
[29] Johnson P., Iacob M. E., Välja M., van Sinderen M., Magnusson C., and
Ladhe T. (2014). A method for predicting the probability of business
Big data analytics and data mining 349
[42] Lee G., Mallipeddi R., and Lee M. (2017). Trajectory-based vehicle tracking
at low frame rates. Expert Systems with Applications, 80, 46–57. https://doi.
org/10.1016/j.eswa.2017.03.023.
[43] Anagnostopoulos T., Kolomvatsos K., Anagnostopoulos C., Zaslavsky A.,
and Hadjiefthymiades S. (2015). Assessing dynamic models for high priority
waste collection in smart cities. Journal of Systems and Software, 110, 178–
192. https://doi.org/10.1016/j.jss.2015.08.049.
[44] Corbett J. and Mellouli S. (2017). Winning the SDG battle in cities: How an
integrated information ecosystem can contribute to the achievement of the
2030 sustainable development goals. Information Systems Journal, 27(4),
427–461. https://doi.org/10.1111/isj.12138.
[45] Castelli M., Gonçalves I., Trujillo L., and Popovic A. (2016). An evolu-
tionary system for ozone concentration forecasting. Information Systems
Frontiers, 19(5), 1123–1132. https://doi.org/10.1007/s10796-016-9706-2.
[46] Polenghi-Gross I., Sabol S. A., Ritchie S. R., and Norton M. R. (2014).
Water storage and gravity for urban sustainability and climate readiness.
Journal – American Water Works Association, 106(12). https://doi.org/10.
5942/jawwa.2014.106.0151.
[47] Breetzke T. and Flowerday S. V. (2016). The usability of IVRs for smart
cities crowdsourcing in developing cities. The Electronic Journal of
Information Systems in Developing Countries, 73(1), 1–14. https://doi.org/
10.1002/j.1681-4835.2016.tb00527.x.
[48] Cilliers L. and Flowerday S. (2017). Factors that influence the usability of a
participatory IVR crowdsourcing system in a smart cities. South African
Computer Journal, 29(3). https://doi.org/10.18489/sacj.v29i3.422.
[49] Hussain A., Wenbi R., da Silva A. L., Nadher M., and Mudhish M. (2015).
Health and emergency-care platform for the elderly and disabled people in
the smart cities. Journal of Systems and Software, 110, 253–263. https://doi.
org/10.1016/j.jss.2015.08.041.
[50] Pramanik M. I., Lau, R. Y. K., Demirkan H., and Azad M. A. (2017). Smart
health: Big data enabled health paradigm within smart cities. Expert Systems
with Applications, 87, 370–383. https://doi.org/10.1016/j.eswa.2017.06.027.
[51] Vincent J. M. (2006). Public schools as public infrastructure. Journal of
Planning Education and Research, 25(4), 433–437. https://doi.org/10.1177/
0739456x06288092.
[52] Mukherjee M., Shu L., and Wang D. (2018). Survey of fog computing:
Fundamental, network applications, and research challenges. IEEE
Communications Surveys & Tutorials, 20(3), 1826–1857. https://doi.org/10.
1109/comst.2018.2814571.
[53] Bonomi F., Milito R., Zhu J., and Addepalli, S. (2012). Fog computing and
its role in the internet of things. Proceedings of the First Edition of the MCC
Workshop on Mobile Cloud Computing – MCC ‘12. https://doi.org/10.1145/
2342509.2342513.
[54] Rashid M. and Wani U. I. (2020). Role of fog computing platform in ana-
lytics of internet of things- issues, challenges and opportunities. Fog, Edge,
Big data analytics and data mining 351
[68] Li R., Zhao Z., Yang C., Wu C., and Zhang H. (2018). Wireless big data in
cellular networks: The cornerstone of smart cities. IET Communications, 12
(13), 1517–1523. https://doi.org/10.1049/iet-com.2017.1278.
[69] Luo F., Cao G., Mulligan K., and Li X. (2016). Explore spatiotemporal and
demographic characteristics of human mobility via Twitter: A case study of
Chicago. Applied Geography, 70, 11–25. https://doi.org/10.1016/j.apgeog.
2016.03.001.
[70] O’Sullivan F. and Holzinger S. A. (2021). AI and big data in health-
care: Towards a more comprehensive research framework for multi-
morbidity. Journal of Clinical Medicine, 10(4), 766. doi:10.3390/
jcm10040766.
[71] Viceconti M., Hunter P., and Hose R. (2015). Big data, big knowledge: Big
data for personalized healthcare. IEEE Journal of Biomedical and Health
Informatics, 19(4), 1209–1215. doi:10.1109/JBHI.2015.2406883.
[72] Zhang Y., Qiu M., Tsai C. W., Hassan M. M., and Alamri A. (2017).
Health-CPS: Healthcare cyber-physical system assisted by cloud and big
data. IEEE Systems Journal, 11(1), 88–95. doi:10.1109/JSYST.2015.
2460747.
[73] Hossain M. S. and Muhammad G. (2018). Emotion-aware connected
healthcare big data towards 5G. IEEE Internet of Things Journal, 5(4),
2399–2406. doi:10.1109/JIOT.2017.2772959.
[74] Sahoo P. K., Mohapatra S. K., and Wu S. (2018). SLA based
HEALTHCARE big data analysis and computing in cloud network. Journal
of Parallel and Distributed Computing, 119, 121–135. doi:10.1016/j.
jpdc.2018.04.006.
[75] Hussain S., Hussain M., Afzal M., Hussain J., Bang J., Seung H., and Lee S.
(2019). Semantic preservation of standardized healthcare documents in big
data. International Journal of Medical Informatics, 129, 133–145.
doi:10.1016/j.ijmedinf.2019.05.024.
[76] Syed L., Jabeen S. S. M., and Alsaeedi A. (2019). Smart healthcare frame-
work for AMBIENT assisted living using IoMT and big data analytics
techniques. Future Generation Computer Systems, 101, 136–151.
doi:10.1016/j.future.2019.06.004.
[77] Zhou S., He J., Yang H., Chen D., and Zhang R. (2020). Big data-driven
abnormal behavior detection in healthcare based on association rules. IEEE
Access, 8, 129002–129011. doi: 10.1109/ACCESS.2020.3009006.
[78] Nazir S., Khan S., Khan H. U., et al. (2020). A comprehensive analysis of
healthcare big data management, analytics and scientific programming.
IEEE Access, 8, 95714–95733. doi:10.1109/access.2020.2995572.
[79] Kim J. and Chung K. (2020). Multi-modal stacked denoising autoencoder for
handling missing data in healthcare big data. IEEE Access, 8, 104933–
104943. doi: 10.1109/ACCESS.2020.2997255.
[80] Jothi N., Rashid N. A., and Husain W. (2015). Data mining in healthcare – A
review. Procedia Computer Science, 72, 306–313. doi:10.1016/j.procs.
2015.12.145.
Big data analytics and data mining 353
[81] Akay A., Dragomir A., and Erlandsson B. (2015). Network-based modeling
and intelligent data mining of social media for improving care. IEEE Journal
of Biomedical and Health Informatics, 19(1), 210–218. doi:10.1109/
JBHI.2014.2336251.
[82] Yassine A., Singh S., and Alamri A. (2017). Mining human activity patterns
from smart home big data for health care applications. IEEE Access, 5,
13131–13141. doi: 10.1109/ACCESS.2017.2719921.
[83] Kovalchuk S. V., Funkner A. A., Metsker O. G., and Yakovlev A. N. (2018).
Simulation of patient flow in multiple healthcare units using process and
data mining techniques for model identification. Journal of Biomedical
Informatics, 82, 128–142. doi:10.1016/j.jbi.2018.05.004.
[84] Amin M. S., Chiam Y. K., and Varathan K. D. (2019). Identification of
significant features and data mining techniques in predicting heart disease.
Telematics and Informatics, 36, 82–93. doi:10.1016/j.tele.2018.11.007.
[85] Sun C., Li Q., Cui L., Li H., and Shi Y. (2019). Heterogeneous network-
based chronic disease progression mining. Big Data Mining and Analytics,
2(1), 25–34. doi:10.26599/BDMA.2018.9020009.
[86] Gonçalves C., Ferreira D., Neto C., Abelha A., and Machado J. (2020).
Prediction of mental illness associated with unemployment using data
mining. Procedia Computer Science, 177, 556–561. doi:10.1016/j.procs.
2020.10.078.
[87] Kaur I., Doja M., and Ahmad T. (2020). Time-range based sequential mining
for survival prediction in prostate cancer. Journal of Biomedical Informatics,
110, 103550. doi:10.1016/j.jbi.2020.103550.
[88] Zhang Q., Lian B., Cao P., Sang Y., Huang W., and Qi L. (2020). Multi-
source medical data integration and mining for healthcare services. IEEE
Access, 8, 165010–165017. doi: 10.1109/ACCESS.2020.3023332.
Index