GROWTH OF POPULATION-BASED PAYMENTS IS NOT ASSOCIATED WITH A DECREASE IN MARKET-LEVEL COST GROWTH, YET David Muhlestein, PhD JD Nathan Smith, PhD June 25, 2017
ACKNOWLEDGEMENTS Funding: Commonwealth Fund Research Contributors: Mark McClellan, MD PhD – Duke-Margolis Center for Health Care Policy Jim Landman, PhD JD – Healthcare Financial Management Association Keith Moore – McManis Consulting 2
PRESENTATION OVERVIEW 1. Background 2. Analysis & Findings 3. Conclusions & Implications 3
BACKGROUND 4
COST GROWTH National Health Expenditures $12,000 20% 18% $10,000 16% 14% $8,000 12% GDP $6,000 10% 8% $4,000 6% 4% $2,000 2% $- 0% 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015 Spending Per Capita Spending as % of GDP Source: CMS National Health Expenditure Data 5
COST VARIATION Risk-Adjusted Standardized Per Capita Costs for Medicare Beneficiaries 6
NEED FOR ACCOUNTABLE CARE Under our current payment system, high value care is not rewarded • Payment system does not incentivize coordinated, quality care • Fragmented delivery system hinders coordination and quality Accountable care requires simultaneous reform of the payment and delivery systems ACO Definition : a group of providers responsible for the cost and quality outcomes of a defined population.
ACO GROWTH 32.4 Million Lives 1000 35 923 842 838 831 833 900 767 30 Number of Lives Covered (Millions) 731 800 741 744 644 25 700 621 652 611 Number of ACOs 600 481 20 453 442 500 464 15 400 314 326 300 10 212 167 200 5 75 65 61 100 0 0 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 2011 2011 2011 2012 2012 2012 2012 2013 2013 2013 2013 2014 2014 2014 2014 2015 2015 2015 2015 2016 2016 2016 2016 2017 # of ACOs # of Covered Lives 8 Source: Muhlestein, Saunders, McClellan. “ Growth of ACOs and Alternative Payment Models in 2017”
GEOGRAPHIC VARIATION IN GROWTH 9 Source: Leavitt Partners Center for Accountable Care Intelligence
ACO RESULTS TO DATE • Small reduction in covered-population cost • Wide variation in individual ACO performance • More general improvement in quality • No strong correlation between cost and quality 10
RESEARCH QUESTION Is the growth of population-based payment models associated with a decrease in health care cost growth or quality improvement at the market level? 11
ANALYSIS & FINDINGS 12
STUDY DESIGN Methods • Growth Curve • Fixed Effects • Qualitative case studies in nine markets Data • Medicare Claims Data – per beneficiary biannual cost growth for Medicare • Truven Health Market Scan Commercial Claims Database – commercial biannual cost growth • Torch Insight – commercial database from Leavitt Partners with ACO lives by market over time • Medicare Hospital Compare – quality data Unit of Analysis: Core-Based Statistical Area (metropolitan areas) Time Frame • Commercial – 2012-2014 • Medicare – 2012-2015 13
SEMI-ANNUAL COST AND ACO GROWTH (AVERAGES) $4,887 $4,969 6% $5,000 5.6% $4,500 5% 5.0% $4,000 Semi-Annual Costs VBP Penetration 4% 4.3% $3,500 3.6% 3% $3,000 2.8% 2% $2,500 1.7% 1% $2,289 $2,000 0% $1,500 $1,602 2012S1 S012S2 2013S1 2013S2 2014S1 2014S2 Period National VBP Penetration National Medicare Costs National Commercial Costs 14
GROWTH CURVE Measure Intercept Slope Quadratic Cubic Medicare Costs -16.45 (4.5)*** 1.5 (1.76) -0.38 (0.64) 0.03 (0.06) Commercial Costs -2.04 (3.61) -2.91 (2.56) 0.33 (0.41) Mortality - HF 0.02 (0.01)** 0 (0) Mortality - Pneumonia -0.01 (0.01) 0 (0) Mortality - AMI 0 (0.01) 0 (0) Readmission - HF -0.04 (0.01)*** 0.01 (0) Readmission - Pneumonia -0.02 (0.01)~ 0 (0) Readmission - AMI -0.1 (0.05)* -0.07 (0.04) Readmission - CJR -0.11 (0.03)*** 0 (0.03) *** p < .001, ** p < .01, * p < .05, ~ p < .10 Note: Numbers represent regression coefficient and SE in parentheses Findings Standard errors in parentheses *** p < .001, ** p < .01, * p < .05, ~ p < .10 • Baseline • Markets with larger proportions of the population covered under a VBP model had lower initial per- beneficiary Medicare costs; no difference for commercial costs • Growth • VBP penetration was not associated with the rate of growth for either Medicare or commercial costs. • Quality 15 • No change in quality
FIXED EFFECTS Costs Mortality Readmission Medicare Commercial HF Pneum AMI HF Pneum AMI CJR Overall Costs Costs VBP -0.81 -0.94 0.01* 0.01 0 0 0 0 0 0.01~ Penetration (0.84) (1.59) (0.01) (0.01) (0.0) (0.0) (0.01) (0.0) (0.0) (0.01) Observation 7,675 5,759 3,452 2,769 3,500 3,103 3,497 3,512 2,537 3,459 s R-squared 0.16 0.08 0.05 0.38 0.1 0.5 0.84 0.28 0.54 0.25 CBSAs 962 962 881 737 887 809 887 888 671 880 Standard errors in parentheses Findings *** p < .001, ** p < .01, * p < .05, ~ p < .10 • Growth of population-based models was not associated with a decrease in Medicare or commercial cost growth • Coefficients directionally suggest lower cost growth • Heart Failure mortality grew worse in markets with higher VBP penetration – likely random/spurious association • When limiting the analysis to markets with higher rates of VBP penetration, there was ~$2 of slower growth (p=0.18) 16
CONCLUSIONS & IMPLICATIONS 17
NO EVIDENCE, YET 1. An increased growth in population-based payment models has not been shown to be associated with a decrease in cost growth at the market level, also not associated with an increase in cost growth 2. A minority of providers within most markets accept financial responsibility for a minority of their patients • The “tipping point” for providers to really focus on cost containment has not been reached 3. Still unknown whether a a concerted focus on lowering costs across all payers and patients will lead to lower costs 18
THE STORY BEHIND THE NUMBERS 1. Change is Hard, Change when Times are good is nearly Impossible • Fee-for-service has been very good for most providers 2. The current business model favors fee-for-service 3. Many organizations are preparing to bear risk under population-based models, but they have not fully embraced the model, yet 4. There has not been a precipitating event to cause markets to “tip” 19
POLICY IMPLICATIONS 1. Policy of encouraging the adoption of VPB to lower costs is still based more on theory than evidence 2. Until a critical mass of payments come through these models, providers are not going to make substantive changes 3. Each market will have a different driver that can encourage greater adoption of these payment models 20
KEY TAKEAWAYS • Non-uniform growth of costs and population- based payment models around the country • Evidence to date does not show that increases in population-based models leads to a slow down in cost growth (also doesn’t lead to an increase in costs) • Providers are preparing for population-based payment models, but are content to focus on fee-for-service for now 21
QUESTIONS? 22
Smart on Value www.leavittpartners.com
Recommend
More recommend