See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/315701137

Big Data in Health Care: A Mobile Based Solution

Conference Paper · March 2017

DOI: 10.1109/ICBDACI.2017.8070826

CITATIONS READS
12 2,909

3 authors:

Minerva Panda Syed Mohd Ali

International Institute of Information Technology, Bhubaneswar International Institute of Information Technology, Bhubaneswar
3 PUBLICATIONS   12 CITATIONS    8 PUBLICATIONS   112 CITATIONS   

SEE PROFILE SEE PROFILE

Sanjog Kumar Panda

University of Illinois at Chicago
4 PUBLICATIONS   12 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Fog Computing for GIS & HealthGIS Systems View project

Continuously variable transmission for Indian Bicycles View project

All content following this page was uploaded by Syed Mohd Ali on 30 March 2017.

The user has requested enhancement of the downloaded file.

 Big Data in Health Care: A Mobile Based Solution
Minerva Panda Syed Mohd Ali Sanjog Kumar Panda
Department of Computer Science Department of Computer Science Department of Production Engineering
and Engineering, and Engineering, ,VSSUT, Burla, India
IIIT Bhubaneswar, India IIIT Bhubaneswar, India [email protected]
[email protected] [email protected]

Abstract—In the present Indian scenario, healthcare informa- institutions maintain independent medical databases and thus
tion is independently maintained by hospitals , institutions and electronic health records are poorly interconnected. There is
not readily accessible in a centralized, informed manner. This need of analytics which will provide instant and accurate
greatly limits the health providers’ efforts to improve quality and
efficiency. Through this paper, we address this issue on bringing insight into patient’s medical history — including past clinical
various information from many sources into one place in realtime conditions, diagnoses, treatments, utilization, and outcomes.
which can be truly life saving. Also, low ratio of doctor to patient But unfortunately there is a scarcity of medical datasets upon
and the low per capita income in India hikes the medical expenses which analytics can be carried out, and this has been addressed
thereby increasing the patient’s inaccessibility to receive proper by the real-time data collection feature of our application.So
health care in their reach especially for people in the rural areas.
A means by which the bridge between the patients and doctors there is a major need to shift from volume based to value
can be gapped and how patients can be treated at a lower expense based healthcare system. Taking these facts into consideration,
is the prime concern. This paper focuses on the development of we have proposed to automate healthcare solutions and reduce
a mobile/web application, through which patients sends their unnecessary medical costs.
symptomatic query to the doctors through a server. The mobile A Digitised HealthCare Model for India is proposed which
application will be equipped with first aid instructions, according
to the nature and severity of the symptoms, either the patients has the potential to revolutionize how populations interact with
are directed to respective departments or given emergency help national health services and also strengthen health systems.
for further treatment.Within the time huge amount of data is col- It will play an important role in preventive, promotive and
lected from users and doctors, this big data will be used to train curative health. The major advantage of such a system is to
machines to automate the tasks to some extent. The information detect and predict diseases accurately, easily and faster with
gained from analyzing massive amounts of aggregated health data
can provide useful insight to improve quality and efficiency for the help of machine learning. This is because of the fact that a
providers and insurers alike.This makes the patients reach out set of particular symptoms will not always lead to a particular
for healthcare solutions easily and cheaply and makes healthcare disease and can be causing another set of diseases. So we
a easy reach for the unprivileged also. Thus, this unified model may be able to tap the appropriate set of diseases linked to
can serve as a data collection, delivery as well as an analytic tool the symptoms easily after analysis.
in the healthcare domain.
Index Terms—Health Big Data; Mobile Application; Central-
This model also meets the key challenges posed to us in
ized health database; Health analytics; Remote Healthcare; health sector that are shortage of human resources in the
sector, accessibility of healthcare infrastructure, affordability
of healthcare services especially for the rural population. Some
I. INTRODUCTION
major benefits from the model includes:
In India, millions are deprived from accessing overburdened • Cutting down recurring medical costs
hospitals and inadequate medical facilities. According to In- • Well- maintained medical history
diaSpend, India is short of nearly 500,000 doctors and the • Secured medical records accessible any-time anywhere
claimed doctor-patient population ratio is about 1:1,674 is in • Centralised system with patients having personalised
India. This is below par with the World Health Organization dashboard for self monitoring as well as for surveillance
(WHO) norm of 1:1,000 population. Also, there is an 83% by the doctors.
shortage of specialist medical professionals in community • Socio-demographic factors and locations of patients taped
health centres (CHCs)[1]. The situation is more worse in the and analysed
rural scenario where there would currently be at best 2.7 lakh Thus, this research model can be a great tool in data collection
doctors to serve close to 870 million people, that is one doctor as well as produce real-time data analytics insights[3].
for about 3,200 people. Rural India still needs 6 lakh more
doctors to meet the WHO norm[2]. II. S CENARIO
In addition to the fact that there is a significant shortage People living in rural areas face major crisis due to shortage
of doctors, doctors are also challenged by the lack of access of proper healthcare institutions as well as medical profession-
and insight into the medical history of their patients which can als. The public healthcare centres and community health care
impact the quality of care they can provide to a patient. Health centres are poorly run with less professional human resources

978-1-5090-6399-4/17/$31.00 2017
c IEEE 131
and limited medical facilities. Also, there is a scarcity of spe- consult doctors online, manage digital health records and read
cialized centres. Local medical centre may have medications health tips.
to provide first aid, but may not have the expertise. People
in rural areas have to travel long distances to access proper
health care which can be the difference between life and death
in the case of emergency. Moreover there is lack of ambulance
in the remote areas, even in urban areas to pinpoint the exact
location of the patient is difficult, hence we need some kind
of interface between the patient and the hospital in order to
the save the crucial time. Patient should be given proper first
aid during the time until some sort of expert help arrives, not
everyone is aware of right kind of first aid in every situation.
Also, doctors often have limited insight into all of the pa-
tient’s disease and comorbidity status. This missing knowledge
can significantly have a negative impact on the accuracy of
medication and also lead duplication of medical tests and thus
incurs more cost. Knowing about patient’s medical history will
help doctor to more precisely identify the cause of the patient’s
condition and eventually lower number of inapt practices
during treatment. Fig. 2. Practo Mobile App.

III. R ELATED W ORK

There are various applications used today which try to C. Portea
provide remote healthcare to the patients[4][5]. Few of the Helps patients connect with specialised doctors for consul-
most popular applications are tation and treatment.It also helps delivering medicines online
A. AmWell and get medical opinions with hospital-quality nursing care at
American Well is ranked as the most popular consumer home.
telehealth app in the world by app analytics services company
App Annie. It connects the patients remotely with the doctors.
The patients can log in with the application or website with
their query and can check for the doctor’s profile and schedule
a meeting (via video, audio or text). A web based queue will
be generated and expected waiting time will be provided. After
the meeting, the doctor comes up with remedies and all the
necessary steps to be considered.

Fig. 3. Portea Mobile App.

D. Isabel (Ann Arbor, Mich.)

Isabel is a clinical decision support tool and gives healthcare
professionals access to an online system that helps them arrive
at an accurate diagnosis more quickly.The diagnostic tool is to
Fig. 1. AmWell Mobile App. understand multiple symptoms. It has a database of over 6000
diseases and conditions. The results provided also consider age
B. Practo and gender. It helps to research and understand the symptoms
A complete health app to book doctor appointments at so as to have a better discussion with the doctor about the
clinics and hospitals, order medicines, set medicine reminders, possible diagnoses.

132 2017 International Conference On Big Data Analytics and computational Intelligence (ICBDACI)
 the server side to routinely analyse the input of the patient.
This analysis contains extracting keywords from the text input
received from the patient and storing those keywords into a
database along with the original text. The next step of this
analysis requires medical examiner to allot this patient to the
relevant ward. Moreover this information will also be updated
in the above mentioned database. This whole procedure will
be automated once we collect enough data from users and
achieve desired accuracy.

Fig. 4. Isabel Mobile App.

IV. METHODOLOGY
Our motive is to create an automated data-driven model that
Fig. 6. Data collection feature of our solution.
integrates the process of healthcare data collection, delivery
and its data analysis. We believe that this model will enable
healthcare organizations and institutions to fully understand This learning system, which is trained in the real-time for
and leverage the capabilities of health big data analytics the whole process will improve overtime with more incoming
and achieve the ultimate goal by improving the quality of data points. This method serves as an effective way for
care while lowering costs and enhancing clinician and patient collecting data from users. It is this big data which is used for
satisfaction. analysis for different purposes. This method can be employed
When an individual’s medical history is documented, it for data collection and further real time analytics. It will
helps to assure that health care providers provide the most help us track/predict the exact disease easily and by what
effective treatment and support for the individual’s illnesses parameters it was mistaken to be another disease preliminarily.
and health conditions so that they maintain the best possible The physicians can use the system for decision support by
health. A person’s medical history is made up of many querying the medical literature stored.
different pieces of information that tell the complete account Thereafter, communication is established between the near-
about that individual’s current and past health. The Model est hospital and the patient who may be asked for more
aims to provide a centralised platform where patient’s database inputs and answering questions for further medication. After
is maintained and thus it will serve as a source to learn the treatment is over the discharge summary is mined to find
about their medical history. It will help ensure maintaining the actual disease and is updated again to the database[7][8].
all medical records at a place accessible any time anywhere. We have divided our server into two layers – first layer
It will also help in studying the Socio-demographic factors will handle non-emergency queries as explained above. The
associated with healthcare. second layer will handle the emergency situations[9].In case of
When the user signs up, unique identification number is emergency, we will bypass the process of examining keywords
generated against which records are maintained. The patient and connecting to the ward rather we will directly connect the
through a mobile device inputs information regarding his/her patient to the nearest hospital using GPS and will dispatch
general details – name, age, gender and major symptomatic the ambulance right away to the user exact coordinates[10].
query. They can upload voice and pictures also. Using these Though we will maintain the separate database of emergency
communication means, vital data concerning the health of the situation which can be further analysed and visualised to get
person is sent to the cloud for processing. If the input from the the insights.Prescriptions, medicine lists and other medical
user is in the form of voice, then we will first convert speech reports (like X-Rays, Blood tests etc.) are maintained in the
into text using google speech to text API[6]. If the data is database and easily accessible by the patient, at any point of
already in the form of text, then we will send it directly to time. The same can be accessed by the doctor appointed for the
the server along with the coordinates of the patient obtained patient and the patient need not carry physical reports every
by GPS of his/her mobile phone. time.
A successful and sustainable analytics strategy requires
V. B IG DATA IN P ROPOSED SOLUTION
building foundational elements of the model first in order to
support the upper levels of the model in later years. Thus The volume of worldwide healthcare data in 2012 was
initially, there is a need for physicians / medical examiners on 500 petabytes. That is estimated to grow in 2020 to 25,000

2017 International Conference On Big Data Analytics and computational Intelligence (ICBDACI) 133
 Fig. 5. Flow diagram of the proposed solution.

petabytes - a fiftyfold increase from 2012 to 2020[11]. Thus, [2] Why more doctors are not the answer to India’s health crisis,
Big Data in healthcare domain has great potential to help timesofindia.indiatimes.com/india/Why-more-doctors-are-not-the-
answer-to-Indias-health-crisis/articleshow/54383884.cms
derive meaningful insights after analysis. Big Data in Health- [3] Health IT Outcomes, “The Power of Real Time Analytics at the Point
care can drive clinical decision support, disease surveillance, of Care,” December 2015
epidemic control and population health management[12]. [4] 50 healthcare apps for clinicians and consumers to know,
http://www.beckershospitalreview.com/healthcare-information-
The key function of our application being data collection, technology/50-healthcare-apps-for-clinicians-and-consumers-to-
data collected can be dealt wih big data technologies. In our know.html
proposed system, the patient’s input and discharge sheets form [5] 10 Indian Healthcare Startups You Should Know About,
http://gadgets.ndtv.com/apps/features/roundup-10-indian-healthcare-
the big data. Apache Hadoop provides a technology to process startups-you-should-know-about-792075
these larger volumes of data and also keep the data on the [6] Soltau, H., Liao, H. and Sak, H., 2016. Neural Speech Recognizer:
original data clusters. Per terabyte of storage in Hadoop is, Acoustic-to-Word LSTM Model for Large Vocabulary Speech Recogni-
tion. arXiv preprint arXiv:1610.09975.
on average, 10 times cheaper than a traditional relational data [7] Yang, H., Spasic, I., Keane, J.A. and Nenadic, G., 2009. A text mining
warehouse system[13]. The Hadoop Distributed File System approach to the prediction of disease status from clinical discharge
(HDFS) stores data across multiple data nodes in a simple summaries. Journal of the American Medical Informatics Association,
16(4), pp.596-600.
hierarchical form of directories of files. Conventionally, data [8] P. Samuel Kirubakaran et al, ”An Automated Health Care Computing
is stored in 64MB chunks (files) in the data nodes with a high Model for Continuous Monitoring of Patients for Immediate Medical
degree of compression. Hadoop uses MapReduce on the data Care during Emergency,” International Journal of Computer Science and
Information Technologies, Vol. 6 (2) , 2015, 1307-1311
being processed. [9] Pendyala, V.S., Fang, Y., Holliday, J. and Zalzala, A., 2014, October. A
text mining approach to automated healthcare for the masses. In Global
VI. C ONCLUSION Humanitarian Technology Conference (GHTC), 2014 IEEE (pp. 28-35).
India takes the second place in the world in its population. IEEE.
[10] Pande, V., Ali, S.M., Kumar, S. and Goyal, P., 2014, February. Au-
Increasing population in India over-burdens the health care tomated first aid and medication system for burn victims. In Advance
structure in the country. We see that there is huge incoming Computing Conference (IACC), 2014 IEEE International (pp. 623-627).
data in the health domain. Thus, the proposed model is IEEE.
[11] Care Customization:Applying Big Data to Clinical Analytics and Life,
the solution for efficient data collection, healthcare delivery Intel White Paper, 2013
integrated with analytics. Thus, this system can automate the [12] Andreu-Perez, J., Poon, C.C., Merrifield, R.D., Wong, S.T. and Yang,
health care services for patients as well as doctors. G.Z., 2015. Big data for health. IEEE journal of biomedical and health
informatics, 19(4), pp.1193-1208.
R EFERENCES [13] Augustine, D.P., 2014. Leveraging big data analytics and hadoop in
developing India’s healthcare services. International Journal of Computer
[1] Healthcare crisis, www.hindustantimes.com/india-news/healthcare- Applications, 89(16), pp.44-50.
crisis-short-of-5-lakh-doctors-india-has-just-1-for-1-674-people/story-
SZepTyjJ78WgOVIo93tBVK.html

134 2017 International Conference On Big Data Analytics and computational Intelligence (ICBDACI)

View publication stats