decision camp 2014 - carole-ann berlioz-matignon - preparing for exceptional behavior
DESCRIPTION
Healthcare plans are customized for each client. Insurance companies have to incorporate all changes negotiated for each customer into their automated system. Carole-Ann will present techniques that have been adopted by Healthcare Insurers to reduce the number of business rules into their systems, and therefore reduce the maintenance on their traditionally creeping systems.TRANSCRIPT
Preparing for
Exceptional BehaviorBusiness Rules / Decision Management
In Healthcare
Carole-Ann Berlioz-Matignon
Introductions
Carole-Ann Berlioz-Matignon
Co-founder & CPO
Sparkling Logic
@cmatignon
Case Study
Division of BCBS of TN
Delivers
Wellness programs
to BCBS of TN and
others
Wellness Programs
What are they?
• Offered by Medical Insurance companies
• Complement Insurance policy
• Aim at keeping members healthy, preventing expensive disease down the road
• By promoting / encouraging goodbehavior
Why?
• Healthcare cost skyrocketing
• Few diseases cause major financial burden
• Saves $$$ in claims
• Evolution towards Population Health Management
• Healthier employees are more productive too
• And improves quality of life for member
How it works
Detect At-Risk
Population
Educate / Encourage
Good Behavior
Reward based on Activity / Results
Lots of Decisions
Detect At-Risk
Population
Educate / Encourage
Good Behavior
Reward based on Activity / Results
Who is at risk? Which activity would prevent
disease?
Which incentive would they respond
to?
What reward should they be credited for?
Did they fulfill all the requirements for the reward?
Issue Detection
Victim of our Success
More Members
Æ More Volume
Plan gets updated with every activity reported, lab test received, etc.
Wellness
Recommendation
Activity Monitoring
Reward Accounting
Onlife’s Challenge #1: Velocity
x15
Insurance AInsurance BInsurance CInsurance D
Onlife’s Challenge #2: Diversity
Issue DetectionVictim of our
Success
More Customers
Æ More
Customization
Each program has
its own specificities
Æ Different Rules
Wellness
Recommendation
Activity Monitoring
Reward Accounting
Wellness
Recommendation
Activity Monitoring
Reward Accounting
Issue Detection
Wellness
Recommendation
Activity Monitoring
Reward Accounting
Issue Detection
Wellness
Recommendation
Issue Detection
Activity Monitoring
Reward Accounting
Issue Detection
Wellness
Recommendation
Activity Monitoring
Reward Accounting
Issue Detection
Wellness
Recommendation
Activity Monitoring
Reward Accounting
Insurance EInsurance XYZ
Issue Detection
Wellness
Recommendation
Activity Monitoring
Reward
Accounting
Using Decision Management
Data
DataData
Data
DataData
Data
DataData
Data
Data
DataData
Data
Data
DataData
Data
DataData
Data
DataData
Data
Data
Decision
Service
Issue Detection
Wellness
Recommendation
Activity Monitoring
Reward
Accounting
Cascading Business Rules
Decision
Service
Generic
Rules Insurance A
With a twist
Insurance B
With less rules
Insurance C
With more rules
Issue Detection
Wellness Recommendation
Activity Monitoring
Reward Accounting
Key Benefit
Decision
Service
Flexibility & Agility
New rules can be
managed by
domain experts
• Clinical team
• Product team
Before…
Clinical team
Product team
Spreadsheet
Most of the time spent here…
Would have been okay for short term
Development
Testing
Initial Focus…
Solid Architecture
Now…
Clinical team
Product teamDecision
Service
Minimum time spent here…
Initial Rules
Repository
with full
Traceability,
Lifecycle
Management
& Governance
Deployment
on the Cloud
After…
Clinical team
Product teamDecision
Service
Most of the time spent here…
Quick Enrollment
Repository
with full
Traceability,
Lifecycle
Management
& Governance
Deployment
on the Cloud
Skills
Profile of our Rule Writer
• Business Analyst
• Technically savvy
An anecdote…
• Plans to expand to domain experts
Lesson Learned – Skills
• Not as hard as we thought!
• How the learning curve was reduced
• Having a very short cycle (immediate) from rule writing to testing
• RedPen technique for learning the syntax, especially on collections
• Reporting to see all cases by risk/issue helped clinical team
analyze and approve the rules
• Decision Improvement
• Based on analyzing decision results Clinical team now looking at
offering more proactive issues identification for those nearing an
issue identified status
• Rules writing worked well with agile scrum methodology
Rules Design
• Business changes
• Over time
• Per customer implementation
• For example
• From food pyramid
• To food plate
Lesson Learned – Rules Design
• Cascading decisions allow customization per health plan
• Some customers choose to stay on food pyramid rather than move
to food plate
• Ability to add new fields as rules were added increased
productivity since BA’s did not have to rely on IT
• Easy to change decision flow and reorganize decision
steps rules
• Easily refactoring as needed as the number of rules increased
through the sprints
• IT/Rule Authoring team successfully worked in parallel w/
coordination around the data model
Deployment
Architecture in flux at the start
of the project
Complete redesign of the
infrastructure
Made decision to decouple
completely deployment from
rules implementation
And it worked!
Lesson Learned – Deployment
• IT and Rules Teams can work mostly in parallel
• Needs for coordination on the data model
• Mitigated
• Performance is not an issue, even on the cloud
• Peak loads during enrollment Jan-Mar up to 1 million transactions
per hour, 200 pages per min on web portal
Questions?
Thank You!
Carole-Ann Berlioz-Matignon
Co-Founder & CPO
Sparkling [email protected]
@CMatignon