Forecasting National Health Expenditures in a CDHC Environment Presentation to Consumer Driven Healthcare Summit, Washington, DC Charles Roehrig Paul Hughes-Cromwick Stephen Parente September 14, 2006

2 Outline  Background  Modeling Framework  Potential Impacts  Current Evidence  Forecasts

3 Background What do we mean by consumer driven healthcare?  High deductibles with savings accounts  Increasing amounts of consumer information Prices Quality Enhanced e-tools Shared decision-making  Incentives for healthy behavior (sometimes) Is there a way to make this work for those with low incomes?

4 Background CDHC impact on national health expenditures  Near term vs. long term  Direct vs. indirect

5 Modeling Framework Healthcare expenditures are determined by:  Need --- which leads to  Use --- which leads to  Payments This includes the impact of technology which affects all three factors

6 Modeling Framework Population Needs 34% 19% 6% 8% 33% 36% 18% 7% 5% 35% Use Payments Privately Insured Under 65 Medicaid Medicare Uninsured 65 and Over 59% 11% 2% 16% 12% 40% 15% 7% 3% 35% 100% Source: Altarum Health Sector Model (AHSM-US 2004)

7 Potential Impacts Why CDHC might reduce need:  Risky behavior since own health care $ at stake  Preventive services if exempt from deductible  HSA contributions tied to healthy behaviors  Cultural shift driven by: Better information Constant media attention to health issues Depends upon benefit design

8 Potential Impacts Why CDHC might increase need:  Preventive services if not exempt from deductible  Reduced adherence to prescribed medications  Postponement of necessary care / delayed Dx Depends upon benefit design

9 Potential Impacts Why CDHC might reduce utilization:  Higher deductible raises price to consumer  Information will increase self-care options  Shared decision-making tends to reduce use Depends upon benefit design

10 Potential Impacts Why CDHC might increase utilization:  More preventive services to avoid future costs  Care is free after exceeding deductible  Better access for previously uninsured Depends upon benefit design

11 Potential Impacts Why CDHC might reduce prices paid:  Individuals will shop for lower prices due to: Higher deductible Better price and quality information An environment that encourages price consciousness  Prices will fall for products/services due to: Increased price elasticity of demand Discounts for cash or HSA debit card payment Long term: shift toward cost reducing innovations Depends upon benefit design

12 Potential Impacts CDHC has the potential to affect long term trends primarily through relentless pressure on prices  Current system rewards expensive innovations  CDHC rewards innovations that improve value Lower cost ways of achieving same benefit Same-cost ways of gaining much greater benefits Will CDHC bargain hunters drive cost-reducing innovation? Will reduced prices simply lead to increased utilization? What about high-cost illness?

13 Potential Impacts Population Category Percent of Population Percent of Spending Per Capita Spending Very Healthy40%2%$200 Somewhat Healthy 52%43%$3,500 Chronically Ill7%30%$17,000 Catastrophic1%25%$100,000 Illustrative Privately Insured Population

14 Potential Impacts Will CDHC impact spending above the deductible? It could conceivably happen this way:  Step 1: Deductible-driven bargain hunting induces and rewards cost-reducing innovations  Step 2: These innovations are incorporated into management of spending above the deductible (tail wags the dog)

15 Potential Impacts Summary CDCH has the potential to reduce personal health expenditures through:  Reducing need  Reducing utilization  Reducing prices Depends upon benefit design

16 Current Evidence: Industry  Aetna Fewer primary care visits More specialist visits Fewer ER visits Fewer hospital admissions Lower expenditure increases  Humana Greater use of primary care and prescriptions Less use of ER and specialists Better adherence to maintenance medications Overall reduction in rate of increase in expenditures

17 Current Evidence: Industry  Lumenos Increased preventive care Reduced outpatient visits Reduced pharmaceutical costs – more generics Reduced cost trend Improvements in diet and exercise  UnitedHealth Group Increased use of preventive care Reduced use of hospital and ER Expenditures actually fell

18 Current Evidence: Researchers  Greene No impact on use of generics Discontinuation of some “essential” chronic illness medications  Parente Some reduction in pharmaceutical costs but no decline in brand name share Increase in hospital costs – free care after deductible Note: plan studied was ‘generous’

19 Current Evidence: Conclusions  Industry and academia differ Academia provides details to support conclusions Industry has not released underlying evidence  Academic research may not be representative Mostly HRAs Limited to a few companies and plans Primarily ‘generous’ plans  Different CDHPs will have different impacts

20 Forecasts  CDHP Enrollment Under Bush Proposal Specifics of proposal Minnesota enrollment estimation model Enrollment estimates  Impact on National Health Expenditures CDHP assumptions Altarum Health Sector Model (AHSM) AHSM expenditure estimates

21 Bush 2006 Proposal President’s 2006 State of the Union (SOTU) speech and explained in detail in the 2006 Treasury Blue Book. As we understand that proposal, it has three related parts:  1. Tax treatment of HDHP premiums : Individuals covered by eligible HDHP would be allowed an “above-the-line” deduction in determining their adjusted gross income. In order to further level the playing field between individual health insurance and ESI, individuals covered by eligible HDHP would receive a refundable tax credit equal to the lesser of: (1) 15.3 % of the HDHP premium or (2) 15.3% of their wages subject to employment taxes.  2. Tax treatment of HSA contributions : The amount that could be contributed before taxes to the HSA would be increased to the out-of-pocket limit for the individual’s HDHP (currently, $5,250 for single coverage and $10,500 for family coverage). In effect, this provision would make all out-of-pocket spending under the HDHP eligible for pre-tax status. In addition, individuals making after-tax contributions to the HSA would be allowed an employment tax credit similar to the premium credit described in #1 above.  3. Low-income tax credit : A refundable tax credit would be offered to low-income individuals and families for the purchase of eligible HDHP. The credit would provide a subsidy of up to 90 % of the health insurance premium, up to a maximum dollar amount, and it would be phased down to zero at higher incomes. Full details of the credit are provided in the 2006 Treasury Blue Book.

22 Minnesota CDHP Enrollment Model Estimate plan offerings using linked data Merge employer data Estimate hedonic premium regression Assign plan choices to full MEPS sample Estimate plan choice regression Use parameter estimates to predict plan choice probabilities for MEPS Re-scale take-up rates Define HSA plan design & premium Simulate impact of proposed policies Model Estimation Choice set Assignment/ Prediction Policy Simulation MEPS Data Sources CDHPseHealthinsurance

23 Enrollment Estimates  Simple table by age and by source

24 CDHP Assumptions Assume generous plan as studied by Parente  Use of Rx falls by 10%  No other effects

25 Altarum Health Sector Model

26 AHSM Expenditure Estimates