Integrating Predictive Analytics Into Clinical
Practice For Improved Outcomes & Financial
Performance
June 11
th, 2015
Transforming the HHS Experience Improving the relationship between payers, providers and consumers
Presenters:
Mike Lardieri, AVP Strategic Program Development, North Shore LIJ Health System Ravi Ganesan, President Core Solutions, Inc.
Agenda
• An overview of various predictive analytics tools • The role of EHRs in predictive analytics
• Practical real world examples for improving clinical practice • Using predictive analytics in population health management
Pyramid for Analytical Need
• Level 1 Questions/Curiosity (Physiological)
- More Questions Than Answers
• Level 2 Data (Safety)
- You Need Data to Make Decisions
• Level 3 People (Love/Belonging)
- You Need People who Love Data and are Adaptable to the Changing Nature of Technology & Analytics
• Level 4 Socialization (Esteem)
- Share the Data – Make Sure it Can be Understood by Senior Leaders as Well as the Rank & File
• Level 5 Artificial Intelligence/Smart Systems
- (Self Actualization)
Large Vendors
Challengers
•
Availability of Software as a Service Model
•
Prices are dropping
•
Improved usability for business users
•
Increasing specialization on verticals –
healthcare, customer service etc.
Role of EHRs in Predictive
Analytics
The Health Information Technology Pyramid
Electronic Health Record
Specialized Clinical &
Financial Tools
Apps and Point of
Care Solutions
Devices
Big Data
The Role of The EHR
• Data Collection - Standardized, Validated Data
• Demographics
• Clinical Quality Measures (MU)
• Behavioral Health Quality Measures
• Analytics - Integration
• Seamless integration with Predictive tools through APIs
• Intelligence - Real-time Feedback
• Configurable business rules and workflows to inform and educate users and direct care.
EHR Checklist
Meaningful Use Certified
Stage 2
Plans for Stage 3
Clinical Tools
Integration of standardized behavioral health tools
Single Integrated Database
Gen 1 vs. Gen 2 EHRs
Paradigm Shift
EHR
+
Intelligence
Predictive
Reactive Care Proactive Care
Healthcare Cost Savings Consumer Cost Savings
Real world Examples For
Improving Clinical Practice
Predictive Modeling Process What is the problem we are trying to solve? Gather as much data as available Evaluate Models Run Predictive Analytics
Suicide Predictive Model
Challenge:
• 41,149 suicides reported in 2013 (CDC);
• 10th leading cause of death for Americans;
• After cancer and heart disease, suicide accounts for more years of life lost than any other cause of death;
• Over 19% suicide rate among people 45 to 64 years old; • 77.9% were male and 22.1% were female.
Value Based Purchasing
• CMS program that rewards quality over Quantity. A good indicator of
what the future of reimbursements for behavioral health is going to look like.
Scheduling Efficiency
• Improving access to behavioral health services requires improvements in scheduling efficiencies.
• Same Day Access – Step in the right direction.
• “Third next available” (TNA) appointment and “office visit cycle time” are validated measures, but not widely used. • 2013 study of the Massachusetts private sector reported
wait times of 50 and 39 days for internal medicine and family practice respectively.
• Scheduling has a direct impact on customer experience. http://www.iom.edu/~/media/Files/Perspectives-Files/2015/SchedulingBestPractices.pdf
Revenue Maximization
Opportunities:
•
Impact of ACO on payer mix
•
Identify unit costs and impact of various cost
components
•
Self Pay/Bad Debt Management
•
Transition from fee for service to value based
Using Predictive Analytics in
Population Health
Management
Key Principles for Population Management
1. Population-Based Care: Focus on caring for the whole
population you are serving, not just the individuals actively seeking care.
2. Data-Driven Care: Utilize data and analytics in order to make informed decisions to serve those in your population who most need care.
3. Evidence-Based Care: Make use of the best available evidence to guide treatment decisions and delivery of care.
4. Care Management: Engage in actionable care management for the population you serve.
Source: http://www.integration.samhsa.gov/integrated-care-models/14_Population_Management_v3.pdf
Steps For Implementing
Population Management
Population Management For Blood Pressure Build a List Identify Care Gaps – No Rx Coordinate Care Check
Compliance MonitoringRemote
Identify other gaps
Predicting Sickness – Blue Cross Goal: Identify Patients Likely to be hospitalized in the next 3 months Algorithms based on claims, lab, Rx, height, weight, family history etc. to
score risk.
Assign Health Coach to coordinate care
and reduce readmissions.
Projects At North Shore LIJ
Health System
Opportunities
•Preventable Readmissions •Length of Stay
•Hospital Acquired Conditions •Chronic Care Management
•Predictive Illness / Disease Progressions •Identification of High Cost Cases
•Wellness Program Management •Micro-Segmentation & Plan Design
Q&A
THANK YOU!
Question and Answers