How Hadoop and a Modern Data Platform Can Enable Transformation in Healthcare
-
Upload
dataworks-summithadoop-summit -
Category
Technology
-
view
484 -
download
0
Transcript of How Hadoop and a Modern Data Platform Can Enable Transformation in Healthcare
How Hadoop and a Modern Data Platform Can Enable Transformation in Healthcare
Briefing for 2016 Hadoop Summit
Beata Puncevic
https://www.linkedin.com/in/beatapuncevic
2
Big Data and it’s Value Potential in Healthcare
Reduce the cost of healthcare by $450 billion
Improve the quality of care of through precision medicine and targeted programs
Prevent disease and improve quality of life through preventative medicine and mobile health
Improve effectiveness of medicine through improved R&D productivity
Make the world a better place by curing disease, predicting epidemics and improving access to care
3
Big Data and Reality in Healthcare
With a few exceptions, established healthcare players are still just experimenting with big data through:• POCs• Pilots• Simple workloads• Ad hoc analytical use cases
4
Plan for Today
1. Industry landscape and challenges healthcare companies face when adopting a modern data platform
2. Strategies for overcoming these challenges
3. What we did and what to prepare for
5
• Composition of customer base
• Expectations and preferences
• Distribution channels
Customer Expectations
Integrated Health Management
Regulatory Impacts
Evolving Competitive Landscape
• Patient centered, value based coordinated care
• Outcomes based or alternative reimbursements
• Risk optimization across the health care supply chain
• Federal minimum value and affordability standards
• Evolution of new products and networks
• Margins driven by admin costs instead of MLR
• New entrants from other industries are changing the face of care practices
• Industry consolidation
Industry Context
6
Challenges Healthcare Organizations Face
A complex and often outdated current state
One size fits all processes and methodologies
Conservative culture
Low margins
Resources and skill sets
Immature enterprise architecture disciplines
These conditions make it especially difficult to implement new capabilities while keeping operations running in what often times is a real time business
7
Building a Business Case
Tier 3 – Data Drives the BusinessGenerate revenue
Enable a competitive edge
Tier 2 – Advanced Use CasesPredictive models
Prescriptive analytics
Tier 1 – Essential Use Cases Data FeedsReporting
Running the business
Tier 0 –Foundation• Quality, trusted data• Agile and scalable architecture• Data governance• Efficient data management• Data usage capabilities
Approach
Tactics
Focus on the solving the problems most prominently recognized as a priority:1. Driving cost out of essential data
solutions2. Enabling pursuit of analytics insights3. Building for agility and adaptability
• Obtain centralized funding for initial and subsequent implementations
• Form a team focused on developing an ROI framework
• Measure value using before/after comparisons
• Secure senior business sponsorship
8
Planning for Success
Start with Data Governance
An effective data governance discipline is critical for a big data platform. Plan on building support from senior level data owners, business glossary definitions, frameworks and processes for decision making
Architect for the Long Term
Focus not on meeting 1 or 2 use cases, but on enabling use cases through a platform that is agile, scalable, reliable and can flex with business needs. This requires a well formed perspective on your target state and architectural patterns
Adoption includes Transition
Value cannot be fully realized until the current state is migrated and retired to the new platform. Plan on developing a transitional architecture across all domains, partnering with portfolio management, measuring sunk costs and ROI, and focusing on change management
Plan to Pay Up for the Sins of the Past
Are your data quality, master data management, metadata management capabilities mature? If not, those gaps will become problematic when building a big data platform and should be addressed
It is Not Just About Technology
Work with an experienced partner, but establish a training approach to build the skills you need. Reach out to HR early on to adjust policies, grades etc. Hire the right people to drive changeProcesses and roles will need to adjust. Plan on developing a new operating model and processes
Consider The Whole Platform
A big data platform alone is not enough, plan on rounding out the architecture with data management, data integration and data access capabilities. Be aware of current state limitations and any dependencies in the bottom part of the stack
9
Case Study: Program Goals
Establish a vision for data as an enterprise asset and guiding principles
Operationalize a governance model for prioritization and data-related decisions
Build a modern, agile data architecture that enables necessary data capabilities
Implement by choosing strategic business use cases to quickly enable value
Outcome• The organization is unified around a
common perspective and provides a set of guardrails for data decisions
• The organization obtains the right skills and adopts the most optimal organizational structure
• Data decisions are made consistently enabling reduction of complexity, re-use and cost reduction
• Technical platforms, processes and architecture are nimble, cost effective and enable advanced analytical capabilities
• Transformation of our information infrastructure begins right away, rapidly creating business value
Recommendations
Refresh the operational model to align skills and clarify roles
10
Case Study: Foundational Capabilities
Foundational Capabilities
1. Rapid data ingestion and integration
2. Proactive data quality management
3. Common definitions in a business glossary
4. Person and Member 360 data foundation
5. Trusted source of enterprise data6. Data as a service data exchange7. Event-driven and real-time data
management 8. Search and collaboration9. Self service reporting and
dashboards10. Advanced analytics
Desired Business Outcomes
• Administrative cost reduction resulting from improved data quality and re-use of enterprise data
• Faster and cheaper project implementations
• Rapid data integration
• Fast and cost effective data exchange with third parties
• Agile and adaptive solutions and capabilities
• Enablement of business analysts to significantly reduce their data preparation efforts
• Analytical data insights that enable new business opportunities and processes
• Reduced complexity of our data and technologies
11
Case Study: Reference Architecture
Data Ingestion For Purpose Data Zone
Data is processed and published with applied
joins, predefined relationships, and
derivations for purpose
Refined Data ZoneData is cleansed, standardized, and
mastered into books of record
Metadata Layer
ELT/ETLOrchestrate the
transfer of data, and capture lineage
Transactional
Data Analysis
Reporting and BI
Data MasteringRaw Data ZoneData is in the native file format in which it
was received
Secure File
Messaging
SOA ClientsExternal Sources Data Access Layer
HL7
Internal Sources
File Storage
Messages
Data Ingestion Layer
Message StorageStream
Ingestion
SQL Result Ingestion
Mainframe Copybook
File Copy
Relational DBSource
Entity ManagementRelationship Mgmt.
Data Enrichment Layer
Entity Relationship Matching
Entity ResolutionRelation Linking
ODS
Data Services
Data Feeds
Inquiry MicroServices
Scheduler
Advanced Analytics
Analytical and Distribution
For Purpose Data Files
Data Marts
Application Extract
Application API
System/ Event Logs
External Database3rd Party Extract
Device Data/ Data Stream
Social Media
Hosted Application
Technical Metadata
Metadata Management
Business Glossary
Data Catalog
Business
Metadata
Data Profiling
Reference Data
Table Storage
Data Job Scheduling
Books of Record
API Gateway
Event Publish MicroServices
Message Queue
SecurityAPI Catalog
RoutingLoad Balancing
Monitoring
Secure FTP
Analytic/Data Science
User
InternalApplications
Portals
ExternalService/Event Consumers
ExternalVendors
(Batch Files)
Standard Data Feeds
Custom Data Feeds
Governance
Consumers
AnalyticWorkspaces
Cubes In-Memory ReportConsumer
MobileConsumer
Business/Data Analysts
Data Quality
Data Search
Data Sharing and Annotation
Data Governance
Web Logs
Data Validation
Audit, Balance and ControlData Quality
1
2
3
4
5
6
7
8
Data Access Layer
9
10
Capabilities1. Rapid data
ingestion and integration
2. Proactive data quality management
3. Common definitions in a business glossary
4. Person and Member 360 data foundation
5. Trusted source of enterprise data
6. Data as a service data exchange
7. Event-driven and real-time data management
8. Search and collaboration
9. Self service reporting and dashboards
10. Advanced analytics
Q&A
Beata Puncevic
https://www.linkedin.com/in/beatapuncevic