1. Big Data and the
Healthcare Sector
Jonathan Hannahs
David Hewitt
Chris Groves
2. What is Big Data?
0 Global data is going to grow by 40%
this year. McKinsey & Company
0 Big data is data that exceeds the
processing capacity of conventional
database systems. O’Reilly Media
0 The ability to spot trends and
patterns in large and diverse data
sets has been in the hands of large
corporations for many years.
3. Wannamaker’s Dilemma
0 John Wannamaker was a Department
Store magnet during the 30’s and 40’s.
0 “I know that half of my advertising
doesn’t work. The problem is that I
don’t know which half.” - John
Wannamaker
0 Because of Google AdWords and the use
of Big Data Analytics, the question of
which part of the advertising is effective
has been answered and caused a
revolution in the advertising business.
0 Currently, doctors recommend patient
care centered on the “standards of care”
treatment and on their own intuition
and training
0 Big data promises to answer for
healthcare what it did for advertising,
which treatments work and which
treatments don’t.
4. Currently the U.S. spends $2.6 trillion on healthcare every
year and $600 billion of that spending is on variations in
treatment (treatments that do not work).
- O’Reilly Media
5. Characteristics of Big Data in
Healthcare - Data Silos
0 Pharmaceutical R&D Data
0 Clinical Data
0 Activity (claims) and Cost
Data
0 Patient Behavior and
Sentiment Data
6. 4 Key Attributes of Big Data
0 Volume: The amount of data being
processed is much greater than processes
in the past.
0 Variety: In the healthcare sector,
unstructured data, combined with
traditional structured data, can offer the
sector a wealth of undiscovered
information.
0 Velocity: With data changing by the day,
hour, or minute, this is the speed at which
you can process data and make decisions
based on that data.
0 Veracity: Data uncertainty and planning
for the uncertainty is a unique
characteristic of big data that sets it apart
of traditional structured data.
(Source: Analytics: The Real World Use of Big Data)
7. Characteristics of Big Data in
Healthcare - Top 5
1. A greater scope of
information.
2. New kinds of data analysis.
3. Real time Information.
4. Data influx from new
technologies.
5. Non-traditional forms of
data.
8. Healthcare & Big Data
Drivers of Adoption
0 Implementation of the
Affordable Care Act.
0 Growing costs for new,
revolutionary technologies
combined with shrinking
reimbursements.
0 Ability for New Big Data
Systems to Process
Unstructured and Structured
Data.
9. 5 Ways Big Data Will Enable Healthcare
Quality Improvement and Cost Cutting
1. Right Living: Big data can help patients take an active
role in treating their current ailments but also in
preventing future issues by encouraging them to eat
right, exercise, and adhere to medication.
2. Right Care: Integration and application of big data
tools will promote evidence-based care that will be
personalized to the patient.
3. Right Provider: Big data can help the healthcare
sector match patients with specific providers based on
outcomes and the patients past history.
4. Right Value: Big data tools can help drive down cost
by helping to eliminate fraud, waste, and abuse and
assist with implementing an outcome-based payment
system.
5. Right Innovation: Big data tools will help improve
specific therapies and care and also help in the
innovation of research and development of new care
techniques.
(Source: The Big Data Revolution in Healthcare)
10. Challenges to Big Data Adoption
0 Silos of Information
within the Healthcare
Sector
0 Patient Privacy and
Government Regulation
0 Provider Trust in Data
Analysis Over Instincts
11. Big Data Top Companies
IBM, a leader in the big data field, has
partnered with Memorial Sloan-
Kettering Cancer Center to introduce
their “Watson” technology to the clinical
setting. Watson isn’t only a search
engine but relies on probabilistic
algorithms to analyze millions of pages
of unstructured data in patient records
and the medical literature to make a
diagnosis and answer treatment related
questions.
12. Big Data Top Companies
Humedica is a private company based in
Boston that offers a cloud-based population-
wide analytics platform. Humedica’s system
connects patient information across the
medical setting and time period to enable
providers to get a holistic view of patient
care.
Over 25 Million Patients
13. Big Data Top Companies
Explorys is a privately held
company based in Cleveland, Ohio.
Explorys is a spinoff company of the
Cleveland Clinic.
The Explorys platform enables
providers to do the following:
0 Complete searches across patient
populations and care venues to
help identify disease trends.
0 Coordinate rules-driven patient
registries.
0 View performance matrix.
14. Big Data in Healthcare
Case Study – Success
Sharp Community Medical Group (San Diego, Ca)
0 Partnered with IBM
0 Created a system to capture and utilize data
generated by its customers and staff.
0 Helps doctors see trends in SCMG patient
populations
0 How many people have uncontrolled
diabetes.
0 How many women haven't had their
mammography screening
0 Considering using natural-language processing
to record and store data.
15. Big Data in Healthcare
Case Study – Success
Camden Coalition of Health Care Providers (Camden, NJ)
0 Collect and analyze data to improve preventative care
for the city.
0 Created a map of medical “hot-spots” that showed that
there were areas of the city where a disproportionately
high level of calls for an ambulance.
0 Data showed that these residents did not receive
effective preventative care.
0 Demonstrated that 80% of the city’s medical costs
could be attributed to 13% of the patients
0 A sample of 36 patients that were given preventative
care brought the group’s monthly average of 62
hospital visits down to 37 and reduced the patients’
hospital costs by 56%
16. Big Data in Healthcare
Case Study – Success
Reasons for Success of Case Studies
0 Buy in- In both cases, the organization
in the SCMG case and the medical
community in the Camden case,
supported fully the implementation of
big data solutions.
0 Measurable results – In both cases,
the data stored and analyzed resulted
in improvement.
0 Adaptive to operational need – Both
organizations tailored their big data
solution to the needs of the people
who need it and are adapting the
solution as circumstances change.
17. Big Data in Healthcare
Case Study – Failure
Canadian research group led by Dr.
Damian Cruse and an American research
group led by Dr. Andrew Goldfine
0 Both research groups were studying
electrical activity in the brains of
patients in a vegetative state.
0 Goldfine found that the Canadian group
did not rely on standard statistical tests
in their research. They used a computer
program based on machine learning to
comb through tens of thousands of data
points.
0 “Big Data” solution recorded all
electrical fluctuations in the EEG and
may have recorded artifacts as brain
impulses.
0 Big data tools very subject to false
positives
Reasons for this failure in this case
include:
0 Using an unproven data
collection method- The method
used to collect the data was a
departure from the norm and was
untested.
0 No data quality oversight- The
researches for the Canadian group
did not double check the computer
operated system to make sure the
data being collected was accurate.
0 Uneducated operators- The
Canadian researches did not
familiarize themselves with the
data collection method before
trusting it implicitly.
18. The Ohio Lottery Commission
0 State’s only legally operating seller of lottery
games
0 Operated more like a business than a state
agency
0 Goal is to deliver a cost effective
entertaining product to the public with all
bottom line revenue being transferred to the
Lottery Profits for Education Fund
0 Since the OLC relies heavily on licensed
retail agents to sell the product, making sure
that product is on hand in the locations and
prominently displayed at key sale points is
essential to the success
0 OLC’s sales reached a record last year of
$2.78 billion
19. Ohio Lottery & Big Data
0 OLC has been using big data
applications since 2007.
0 Currently in place are data warehouses
and reporting tools for human capital
management and financial functions.
0 The State of Ohio uses two main
applications for compiling and
distribution of data.
0 Oracle’s PeopleSoft application -
Enterprise PeopleTools 8.46 is used
for data entry.
0 IBM’s Cognos software is used for
reporting and analytics.
0 OLC could benefit from more specific
big data functions.
20. Potential Applications of Big
Data Tools at the Ohio Lottery
0 3 departments generate usable data on a daily
basis that if properly utilized could build a greater
competitive advantage; the Lottery Call Center,
Sales, and Marketing.
0 Currently, the three departments operate on
separate systems.
0 A single data warehouse to store all the
information from all departments would improve
times for resolution of customer and retailer
requests.
0 Sharing information in real time would be a great
benefit to the organization.
0 Sales representatives in the field would have
better information regarding call center data.
0 Call center agents would have better data of what
is happening in the field which will lead to more
accurate resolutions over the phone.
0 Marketing would have a better picture of where
promotions are needed.
21. The Ohio Lottery & Big Data
Management Challenges
0 Management of Information – The
OLC will need to first build a
management team to design and
manage the new system.
0 Building Infrastructure – To make
this system truly beneficial to all
users, it will need to be accessible by
all users.
0 Building employee trust in data - If
a data driven solutions are to grow
and thrive in the agency, all
employees involved need to be on
board.
22. Big Data and Competitiveness
@ The Ohio Lottery
0 Big data would have an impact on the
ability of the OLC to anticipate customer
desires and react to changes.
0 Competing for discretionary dollar puts
OLC products in direct competition with
a wide variety of products at the same
price point (i.e. Snack foods, beverages,
tobacco products, etc.)
0 OLC specific data shared across
departments will allow the more for
more effective sales and marketing
solutions at the majority of retail
locations.
23. Big Data and Competitiveness
@ The Ohio Lottery
0 Opening up of the state to casino gaming
offers the OLC an opportunity to expand
their reach into a new market segment.
0 The OLC has begun to partner with Ohio
horse racing tracks to create a new entity
in the state’s gaming/gambling
industry, the Racino.
0 The analysis of data from Video Lottery
Terminals at Racinos will assist the OLC
to uncover many ways to improve profits.