HEALTH CARE ANALYTIC SERVICES CONTRACT TRUVEN HEALTH ANALYTICS AND BRANDEIS UNIVERSITY DECEMBER 2013
Healthcare Analytic Services Contract - Status Overview of the contract Current status Preliminary descriptive results for 2008-2012 Special Studies Agenda Key Decisions 1
Contract Overview Duration: May 2013-May 2015 Value: $2,377,290 Goal: Leverage VHCURES and other data assets to support the GMCB. Enhance Vermont s analytic capabilities to understand cost trends, profile subpopulations, measure health status and its relationship to spending, benchmark spending against regional and national trends, model the impact of future policy changes Tasks include: Transform VHCURES into analytic files for research Identify and specify health analysis populations Provide granular estimates for detailed expenditure analyses Provide regional and national benchmarks for key metrics Undertake special studies 2
Current Status We have data from OnPoint for 2007 to partial 2013 Commercial and Medicaid only Medicare data may be available by March 2014 We have created analytic files Preliminary estimates for 2008-2012 Health Analysis Populations Technical Advisory Committee met this morning Health accounts work is well underway Assembling benchmark databases Detailed plans for special studies 3
Focus on Things that Matter Moderately expensive services delivered to a lot of people are a bigger concern than really expensive services delivered to a small number of individuals Spending per member per year is a more useful statistic than spending per user or spending per use Remember Vilfredo Pareto 20% of the people drive 80% of the cost It is useful to focus on high cost claimants are they recipients of many moderately expensive items or a few really big ticket items While it is always important to look at the level of spending, it is also important to identify rapidly growing areas Growth in spending per member per year is a good statistic to evaluate These tabulations focus on spending overall the decomposition special study will look at how much is utilization versus price change 4
Commitment to Transparency Enrollment denominator are those people who had both medical and pharmaceutical coverage Inpatient stays are defined by an acute care billtype and a room and board revenue code; expenditures for claims (professional and facility) are included in allowed amount LTC stays are person/place/svcdate for billtypes 21, 22, 25, 26 in the outpatient services table ER visits summarize unique person/svcdate encounters for facilities reporting an ER revenue code or a provider reporting an ER CPT4/HCPC (excluding ER admits) Facility-based outpatient visits summarize unique persons/provider/dates in outpatient services for facilities with a valid billtype (excluding LTC, ER) Professional visits are all claims in outpatient services not already defined Prescription drugs summarize the data from the pharmaceutical drugs table this does not include drugs administered in acute care hospitals 5
Commercial Spending Per Member Per Year ($), 2008-2012 Commercial 2008 2009 2010 2011 2012 Inpatient Acute-Care Stays 636 703 782 892 925 45.5% 22.9% Facility-based Outpatient Visits 1,041 1,160 1,311 1,528 1,521 46.2% 38.0% ER Visits 162 181 208 247 256 58.2% 7.5% Long Term Care Stays 9 8 8 11 8-9.8% -0.1% Professional Visits 901 944 1,021 1,165 1,110 23.2% 16.5% Prescription Drugs 562 633 710 778 753 34.0% 15.1% Commercial Total 3,310 3,630 4,040 4,621 4,574 38.2% 100.0% Year Change from 2008 % of total PMPY Delta Emergency Room Visits include both the facility fee identified by revenue code and professional fees identified by CPT codes for the same patient on the same day. Prescription Drugs are from outpatient prescription drug claims. Inpatient drug use is not included. 6
Medicaid Spending Per Member Per Year ($), 2008-2012 Medicaid 2008 2009 2010 2011 2012 Inpatient Acute-Care Stays 537 635 698 761 738 37.6% 56.7% Facility-based Outpatient Visits 747 747 731 746 777 4.0% 8.4% ER Visits 130 130 139 170 188 44.5% 16.2% Long Term Care Stays 616 572 544 533 546-11.4% -19.7% Professional Visits 2,825 2,812 2,749 2,768 2,889 2.3% 18.2% Prescription Drugs 652 692 676 687 723 11.0% 20.1% Medicaid Total 5,506 5,589 5,537 5,664 5,861 6.5% 100.0% Year Change from 2008 % of total PMPY Delta Emergency Room Visits include both the facility fee identified by revenue code and professional fees identified by CPT codes for the same patient on the same day. Prescription Drugs are from outpatient prescription drug claims. Inpatient drug use is not included. 7
Key Points on Spending Per Member Per Year Hospital inpatient care is a huge cost driver for both Major Medicaid shift between 2008 and 2009 increased dramatically the importance of hospital inpatient care Looking just at 2009-2012 does not change the basic message Hospital outpatient care is key for commercial but under control for Medicaid ER visits are growing rapidly for both populations but are a smaller part of the overall cost picture Rx costs growing rapidly for commercial but not Medicaid 2008-2011 may involve better capture of commercial Rx in VHCURES. 2011-2012 is much closer to what we observe in other data sources Medicaid LTC spending is going down (!?!) 8
HOW MUCH OF TOTAL SPENDING IS CONCENTRATED IN TOP %? 9
Percentage of Total Spending from Top 10% and 20%, 2012 90% 80% 70% 60% 50% 40% 30% 20% Pareto was right 20% of the people are responsible for 80% of the cost Two thirds of total spending is for the top 10% These patterns are consistent in all years Investigating the drivers of cost for these specialized populations is worthwhile. 10% 0% Top 10% Top 20% Medicaid Commercial 10
Summary We are starting to look at the VHCURES data Next steps will be comparisons with other data sources Our goal will be to set up valuable routine reporting that is: informed by the special studies undertaken by our collaborators at Brandeis highlights the experience of key subpopulations compares the Vermont experience to the rest of the country VHCURES is a valuable asset for understanding health care delivery in Vermont 11
SPECIAL ANALYTIC STUDIES BRANDEIS ANALYTIC TEAM
Overview of Special Studies Specific analyses to support policy questions, drill down from expenditure reports Purpose: Support specific GMCB regulatory and policy decisions Budget review Premium rate review Payment reform and ACO implementation Population health management Most are built around episodes of care linking payment reform, cost of care, quality, efficiency, and population management Combination of long range (e.g. 6 month time frame) and short turnaround as needed, integrated to use shared metrics Transparent, all analytic files and steps shared 13
Study 1: Drivers of Health Spending Growth (Decomposition) Purpose: support budgeting and premium rate reviews by examining payment growth by payer, provider, and service Description Decompose overall health expenditure growth 2008-2012 into price/quantity/intensity (service mix) Compare results by payer, by service, by setting, by geography Focus also on drivers of episode costs for selected diseases Status: Analytic plan being reviewed by Vermont Data request submitted to Truven Health Analytics Timeline: Initial/preliminary results April 2014 Final report June 2014 14
Study 2: Population Health Metrics / Understanding Social Determinants of Health Purpose: Identify potential data sources and metrics for monitoring socioeconomic gaps over time Explore potential opportunities for linkages across different data sets Document SDH indicators in each data set and the methods to produce reports for monitoring SDH and health over time. Approach: Review current health status and health behaviors surveys Describe the types of valid inferences that can be made Recommend opportunities for integrated database to assess population health at individual, community, market and state level Examine associations between socioeconomic factors and health status Status: In design phase, need input from Vermont on specific goals 15
Study 3: Episodes of Care to Support SIM Purpose: Support work group in development of episodes of care Approach: Review of episodes of care and bundled payments currently in use, and of commercial or public groupers and specifications Summary (brief) of epidemiology of episodes under consideration Cost and quality variation across 15-20 episodes, among providers and payers Provide specifications and final documentation around selected episodes Status: proposed, under review Timeline: Reviews and summaries 8 weeks Cost and quality variation 12 weeks Specifications and final documentation around 3 selected episodes 24 weeks 16
Study 4: Care Management for ACO Populations Purpose: Support ACO implementation and population management by assessing cost and quality variation in episodes of care Approach: Choose high cost/high spending diagnoses Document variation across providers in care costs and quality (as available) in ambulatory, acute and post-acute care settings Status: proposed, for discussion Timeline: Analytic plan six weeks after approval of general approach Draft report five months after start Final report six months after start 17
Study 5: Population Hospital and Outpatient Patient Flow Patterns (proposed) Purpose: Document where patients within Vermont are seeking care, for what diagnoses and diseases. Identify high-volume hospitals and markets for various services. Can be basis for comparing costs and quality of services, and modeling changes in services or payments across hospitals Support budget review Approach: Use VHCURES to identify high volume/cost services or episodes Examine and map utilization patterns from patient and hospital perspective Identify high volume and other patterns of hospital inpatient and outpatient care Status: Proposed, for discussion. Need Medicare data for full analysis.
Proposed Special Studies Timeline (estimated) Study Deliverable - Interim or Draft Deliverable Final Report Decomposition April 2014 June 2014 Social determinants of health May 2014 July/August 2014 Episodes of care March 2014 July 2014 ACO implementation support July 2014 September 2014 Patient flow analysis (pending Medicare data Spring 2014) Comparison of cost and quality across hospitals for care episodes What if simulating different payment policies and utilization patterns August 2014 October 2014 January 2015 March 2015 March 2015 May 2015 19
Key Decisions Medicare integration Special Studies agenda Decomposition study Episodes of Care Overall budget and balance with policy goals Update schedule 20