Big Data Challenge

1
Big Data Challenge From EMRs to Translational
ResearchMiroslav Koncar, PhDOracle Healthcare
Business Development DirectorHL7 Croatia Chair,
Associate Professor for Biomedical Informatics
2

• What is Big Data
• Why it is important for Healthcare Providers Now

3
MIT Big Data DefinitionWhat kind of big are we
talking about here?

• U.S. Healthcare data is growing by 15 Petabytes a
  day currently
• Up to 80 of healthcare data is currently
  unstructured.

Big Data The management and use of ultra-large
amounts of information Management and use
efficient storage, search, analysis,
visualization Ultra-large gt1Petabyte
5 million bytes (megabytes) Complete printed works of Shakespeare
20 billion bytes (gigabytes) Complete recorded works of Beethoven
10 trillion bytes (terabytes) Printed works in the Library of Congress
2 quadrillion bytes (petabytes) Google Earth information managed today
Typical persons EHR 1MB healthy young, no images 40MB middle-aged w/health issues, no images 3-5GB with several health issues w/images
600 petabytes-10 exabytes (exabyte billion gigabytes) Estimated size of current US digitized patient data
1.5 million Estimated different medical devices (WHO)
4
Where Are Your Big Data Opportunities? Lower care
cost and raise care quality

• Capacity Management
• Staffing Optimization
• Protocol Optimization
• Process Efficiency
• Supply Standardization
• Cost Accounting

• Personal Preferences
• Individualized Treatment
• Biometrics
• Data Management
• Apply Research
• New Clinical Protocols

• Chronic Disease Mgt.
• Care Patterns
• Population Health
• Pay for Performance
• Telemedicine
• Patient Advocacy

• Customer Satisfaction
• Disease Mgt. / Prevention
• Social Media Interactions
• Care / Process Improvements
• Marketing / Branding

5

• Why Oracle for Big Data

6
Big Data Oracles DefinitionLarge volumes, high
change velocity, complex variety, unknown value
per byte

• Emerging sources of data, ie. sensor, machine
  generated, social, images, unstructured, etc.,
  that exhibit any or all of the four Vs,
  potentially integrated with traditional
  structured data
• Healthcare providers can acquire, organize,
  process, analyze massive amounts of data,
  primarily through the use of scale-out on
  low-cost hardware, non-relational data models,
  APIs

7
Oracles Leadership Across the Entire Healthcare
Ecosystem Provides Us Unique Design Insights

• Leading provider in
• BioPharma
• Clinical Trials
• Translational Research
• Healthcare Insurers
• CMS
• HIE Insurance Exchanges
• Population Health Mgt.
• Healthcare Data Models
• Healthcare Analytics
• Interoperability
• Identity Security Mgt.

8
Our Translational Research Center (TRC) Can Be
Added For Those Involved in The Research Domain
9

• Customer Success Story
• Erasmus MC, The Netherlands

10
Erasmus MC

• Erasmus Overview
• The biggest Dutch University Medical Centre
  located in the City of Rotterdam
• A 1.200-bed, 1.4BUS operational budget, 12.000
  employees
• Erasmus MC is one of the top 40 best research
  institutes of the world
• Long years an Oracle customer
• Full ERP (financials, procurement, HR, learning
  management), Performance Management
• Technology Database, Fusion Middleware

11
Translational Research in Erasmus

• Today - Clinical genetic testing is typically
  done for patients who are considered at risk of
  carrying genetic variations that are linked to a
  particular disease or disease predisposition.
• Next step - all variants in coding and
  non-coding DNA is mapped at once.
• The vast amount of knowledge that is offered by
  whole-genome sequencing can provide the scientics
  with new oppprtunities for disease management
• Barrier how to manage vast amounts of data in
  real time, and manage all risks associated
  (security, perfromance, etc)

12
HuVariome Database Business Problem

• Genome sequencing is time and resources demanding
  process
• Complete human variants per sample generate 1.5TB
  of data
• Need to provide overview of all variants in
  genetically defined populations
• Variations are displayed in the context of public
  knowledge
• Statistical quality measure for variants
• Disease association and functional impact
• Provide easy access to translational research
  scientists
• Simple search web page to access variation data
• Summary reports for disease association
• Cytogenetic visual representation of variations

13
Why Oracle Exadata
Data storage is 10x smaller, Scanning process is
2000x faster
14
Conclusion

• Complete genomics offer new opportunities for
  disease management
• Identifying novel DNA variants relevant for Dx
  and therapy
• Companion diagnostics for targeted therapy
• Key areas of research cancer conditions
• Visit Erasmus and HuVariome project at
• http//www.erasmusmc.nl/
• http//www.youtube.com/watch?vKVvVs2SVfHU

15
miroslav.koncar_at_oracle.com