Page 207 - Big Data Analytics for Intelligent Healthcare Management

200 CHAPTER 8 BLOCKCHAIN IN HEALTHCARE: CHALLENGES AND SOLUTIONS

8.1.1 ROADMAP
The different sections of this chapter are presented as follows: Section 8.2 elaborates on the overview of
healthcare big data and blockchain architecture; Section 8.3 elaborates on the privacy and regulations
associated with healthcare big data; Section 8.4 explores the effectiveness of applying blockchain
based-applications on healthcare big data with adequate case studies; Section 8.5 provides several
blockchain-based challenges and solutions for healthcare big data management; and Section 8.6 con-
cludes the chapter with future research direction followed by necessary references.

8.2 HEALTHCARE BIG DATA AND BLOCKCHAIN OVERVIEW
8.2.1 HEALTHCARE BIG DATA
Laney [13] suggested big data is a large set of complex data with manifold properties that are analyzed
computationally in order to identify patterns associated with it. Big data is a term for a set of data that
are so huge and composite that existing data handling applications are insufficient. 3Vs (volume, ve-
locity, and variety) are used to define the characteristics and dimensions of the huge set of data or big
data. “Volume” relates to the data size and dimensionality. The processing speed of the data is labeled
by “velocity.” Finally, “variety” denotes the combination of a number of different types of data. Due to
the emergence of the modern healthcare sector, most of the components associated with healthcare
industries are producing enormous amounts of healthcare data. Diverse medical records are available
from various sources such as traditional patient data contained in text, clinical images, and sounds
recorded, X-rays and ultrasounds, MRI (magnetic resonance imaging), patient’s conversation with doc-
tors, and several healthcare IoT devices and trackers etc. Lee et al. [14] mentioned two major directions
of EHRs or healthcare big data that are available at this time: sensor data from the different healthcare
devices and electronic medical records (EMRs).
Healthcare data mining and big data are closely linked, as extracting a pattern from EHR can help
doctors in future disease prediction. For example, the relationship between Parkinson’s disease and gait
was clearly analyzed using the text mining approach [15] and the link between Parkinson’s disease and
healthy older adults was analyzed using machine learning techniques based on the pattern generated
using the gait characteristics [16]. Since millions of medical data types exist, a necessity of EHR clas-
sification has been mentioned in several studies [17–19].
EMR data are a list of information gathered by medical institutions from the start of the treatment to
the cure of the disease. It’s a series of time-specific information recorded by hospitals, as shown in
Fig. 8.3.
Widely used IoT equipment is mobile phones, wearables, microphones, CCTV, ambient sensors,
skin-embedded sensors, smart watches etc. The aforementioned devices are collecting information in
order to monitor the status of several body parts or functions of patients. For example, the Parkinson’s
disease (PD) gait data was measured using wearable sensors and those data are used for developing a
diagnostic tool to assess the PD patients [20].
Fig. 8.4 explains the kind of sensor data that can be recorded using different sensors.
Eventually, the aforementioned sources collectively produce the mammoth amount of healthcare
big data. Healthcare data is increasing at an astronomical rate, as mentioned by a survey by Stanford
Medicine [21]: 153 exabytes (one exabyte¼one billion gigabytes) data were generated in 2013 where