IoT in Healthcare: Advantages, Challenges, and Current Methodologies

IoT IN HEALTH CARE DOMAIN

ABSTRACT – Internet of Things or Iot can be defined as the system of physical things which are associated or embedded
with a number of sensors, software’s, connectivity and electronics. This allows a better exchanging of information with other
devices which are connected to the network. Or in simple terms IoT can be defined as the network of physical devices which are
capable exchanging information. It is an emerging technology and has been applied in various fields. One such field is the
healthcare. IoT has greatly helped the process of providing medical care. This report consists of the current and previous
implementations of IoT in the field of healthcare. The report also consists of a literature review which shows the research work
of differ researchers in this fields. Lastly, the report proposes a model for further development. IoT or Internet of Things
generally refers to the ever growing network consisting of physical objects which featured with an IP address for connecting
with the internet. Along with this IoT also refers to the communication, which generally occurs between all this objects and
other devices or systems, which are internet enabled. IoT is used in various domains and one such domain is the healthcare
domain. This is used for various purposes in the healthcare domain like monitoring of health, notification during any kind of
medical emergency and many more.
1. INTRODUCTION (15 MARKS)
IoT is an emerging an emerging technology and has various
advantages as well as disadvantages. Various domains have
already started the use of IoT. One such domain is the
healthcare domain. IoT is generally used in healthcare domain
for monitoring the health remotely and provide notifications
whenever there is an emergency [9]. The IoT devices used in
healthcare for the purpose of monitoring the health might
range from simple devices like blood monitoring devices or
heart rate monitors to highly advanced monitoring devices
which are implanted like the pacemakers, Fitbit electronic
wrist bands or advanced hearing aids. However, along with
the advantages there are various challenges that are faced by
the IoT in the field of healthcare [1]. Healthcare is a complex
sector, there exists different stakeholders who are having
different objectives, and this structure differs in each country,
as there exists different government regulations different
countries [12]. IoT is still an unknown area and healthcare
sector is one of the complex sector having involvement of the
government. The current and the previous state of remote
health monitoring has been presented.
2. PROBLEMS FACED BY IOT
2.1 Lack of Electronic Health Care integration System
Data which are collected from the IoT devices maignt
include certain things like the vital sign’s of the patient,
physical activities or glucose level while the patien is present
at home and many more dose not travel typically to the
electronic health record system [3]. It has also been seen that
is most cases the data is not centralized and can be easily
availale for the providers, which is initially responsible for
limiting the information value as the adat is not presented to
the provider always in context to the clinics. Along with this
there exists certain electronic health recording systems where
the patient is allowed to import the data directly into their
record despite of this there still exists some limitations for few
of the dominant EHR players [7]. Initially, this leaves many of
the providers to remain uncertain about the way in which the
information (outside their record system) is to be handled.
2.2 Security threats
Security threats is a primary concern for the regulatory
bodies present at the healthcare industry. The concern mainly
includes the security of the privacy of the personal healthcare
information, which are stored and conveyed by making use of
the connected devices [2]. Many of the organizations
associated with the healthcare makes sure that the soring of
the sensitive data is done in a secure and encrypted manner.
Along with they are also not having any type of control over
the safety and the security for the data access points, which
are used for the purpose of transmission of the data. Initially
this acts as the significant threat, which increases gradually
depending on the number of devices, which gets connected
network [1].
2.3 Multiple device integration
Integration of multiple devices stands out to be an obstacle
in the path of success of IoT in healthcare. Most of the devices
that are present in the hospitals along with the health devices
need to be connected to the network for collecting data from
the patients [3]. The most prominent challenge that exists is
today’s world is that the manufactures of IoT devices for
healthcare have not agreed upon any set of protocols or
standards. Therefore, whenever multiple number of mobile
devices is connected to the network for collecting the data
then it becomes a very complicated process of grouping all the
information collected. This is because the mobile devices are
having different protocols [5]. Due to lack of homogeneity
between the medical devices or the IoT devices used for
medical process reduces the chances of success while
implementation of the IoT technology in healthcare domain.
2.4 Inferring results from immense data
Different and numerous types of complexities are attached
with the process of aggregation and collection of data. Despite
of the fact that combined results helps in deriving of new
conclusions by inferring the records of the patient, the results
which comes up can be very much challenging and might not
have been checked by any data expert or night not have
undergone any type of analytic program so as to get refined
[4]. Along with this, the identification of valuable as well as
actionable data is one of the critical factor and this is because
medical specialists and physicians find it very difficult to
reach the conclusion regarding the growth of the data. There is
lack of quality due to increased amount of data in the process
of decision-making [6]. Besides this, the concern is becoming
much bigger due to the involvement of new devices, which is
connected to the network that is associated with continuous

collection of the data along with the generation of big data as
well.
2.5 Constant changes in the Hardware and Connectivity
More than one device is required by the patients for the
collection of data that the provider needs. For this purpose,
there might exist a need of more than one sensors and in most
cases, it has been seen that this sensors is used along with a
hub where the data gets pushed [8]. These pubs are designed
for processing the information. It has also been seen that these
hubs are not compatible with the different types of sensors,
which are available, and lacks in common hardware or
wireless connectivity. This will initially lead the patients to
have an extensive hardware with them, which would be
overwhelming as well as costly.
2.6 Interoperability challenges
The patients prefer collection of different sets of data by
making use of different types of medical devices. The usage
of such medical devices depends on the purpose of each
device or according to the instructions of the physicians. In
many cases, it has been seen that the data, which is captured
by the IoT device, stays within the boundaries of each system
and the IoT vendors [8]. This collected information is not
visible to any other systems. However, it is unfortunate that
lack of wider adaptation of the adequate interoperability has
led to the lock down of the data from different IoT devices in
each individual system. This initially leads to the loss of
potential values for the rest of the team associated with
patients care. Current and previous methodologies
3. Current and previous methodologies
The evolution of the medical instruments is evolving at a
slow pace. There exists the need of regulatory approval as
well as training for the medical personnel so as to use to new
equipment’s and the measuring devices. This initially results
in limiting the rate of growth of the new innovations.
According to Moore’s law, the rate of development of the
electronic is growing at a much faster rate and is generally
dictated by the economic considerations. The wearable
sensors generally represents a much more dynamically
evolving set of measurements devices than the conventional
medical instruments. Along with the addition of new sensor
modules updated sensors and obsoleted ones a heterogeneous
mix is to be deployed at any point of time. There is a need of
further development in the machine learning process so as to
deal with the heterogeneous sensory inputs which are
continuously developing. There is also a need of coping up
with the streaming data data of varying dimensionality and
semantics as sensor designs change over time and inevitably
missing values of the data by the analytics which are done on
the data gathered from the wearable sensors. Operating in this
type of environment makes the learning task face significant
challenges despite of the advances made in this area along
with the emergence of big data applications. Big Data consists
of massive volumes of high-dimensional observations, which
are often available at the modes of streaming. The
development of sequential algorithms have taken place in both
domains that is in the primal and dual domains. This are
generally associated with targeting the online supporting
vector machines. This type of algorithms are not designed for
the purpose of dealing with various feature dimensionalities
which varies according to time, the incomplete vectors due to
the absence of the features or failure of the acquisition and in
case if this is not treated properly then it might lead to serious
impairment of the classified performance. It is possible to
input the missing values so as to cope up with the missing data
by making use of the linear or non-linear functions of the
features which are available. This is followed by proceeding
with the clairvoyant learning scheme which is based on the
full data.
Second of all as the data of the sensors are plentiful and
they are completely untagged. So there exists the need of
getting this data associated with the “ground truth
schematics”(diagnosis of the physician) so as to make them
usable in the process of training for machine learning
algorithms. However it is infeasible while requiting for the
additional inputs from the overloaded physicians. So the need
of new creative method arises which would be acting as an
alternative for this. One of the attractive possibility is the
ability to leverage the clinical records and this is becoming
readily accessible by the deployment of the Electronic Health
Record System.
The figure provided above shows the framework of the
current data analytics. The advantage of creating the link with
the clinical record is that all the ongoing clinical process
would be helping in providing data for the training related to
machine learning.
4. Literature Review
This section of the report generally consists of the literature
review of the various issues addressed by different
researchers. The main issues includes the issues related to
chronic diseases, Artificial intelligence in the field of
healthcare, IoT in healthcare and many more.
Islam et al., in the year of 2015 discussed about the use of
IoT for the purpose of remote monitoring of the patients
having a chronic disease [11]. The patients generally requires
a regular follow-up about their conditions. This reduce face-
to-face visits with the doctors.
According to Shima Okada et al. who mainly focused on
the body movements during sleep as they considered that
movements of the body is generally responsible for sleep wke
cycle [10]. In their work, they proposed a model for the
purpose of measuring the body movements of an individual
while sleeping by making use of different image processing.
For the purpose of validating their research they compared the