Medicare Payment and Delivery Innovations (with an Emphasis on Accountable Care Organizations)

1. Medicare Payment and Delivery Innovations (with an Emphasis on Accountable Care Organizations) Robert A. Berenson, M.D. Institute Fellow, The Urban Institute Long Term Care Discussion Group Washington, DC 19 February 2013

2. The Basic Problem with Current FFS Payment to Clinicians • The Resource Based Relative Value Scale (RBRVS)-based fee schedule has limitations • The relative values of 7000+ codes approximate resource costs, and not based on services beneficiaries need • What attempts to be an objective process is, despite good intentions, inherently subjective • MedPAC is calling for actual data to inform the CMS-RUC process, e.g., to determine actual time, not estimates, for work values and obtaining real data on practice expenses

3. FFS for Primary Care Has Been Rooted in Face-to-Face Encounters There are various reasons -- – high transaction costs, associated with non-face-to-face, frequent, low dollar transactions; – program integrity concerns – “moral hazard” driving expenditures Yet, increasingly, face-to-face visits do not encompass the work of primary/principal care for patients with chronic conditions and functional limitations (many Medicare beneficiaries and many dual eligibles)

4. FFS Attributes Advantages • Rewards activity, industriousness • Theoretically can target to encourage desired behavior • Implicitly does case-mix adjustment • Commonly used by payers and physicians Disadvantages • Can produce too much activity, physician-induced demand • Maintains fragmented care provided in silos • High administrative and transaction costs • What is not defined as reimbursable is marginalized • Complexity makes it susceptible to gaming and to fraud

5. PPPM – Per Person Per Month – For All Services (Global Payment) • Advantages • Internalizes allocation of activity and costs to meet needs • Direct incentive to restrain spending • Predictable and capped spending • Administratively simple (until address some of the problems) • Low transaction costs • Disadvantages • May lead to stinting on care • Susceptible to cream-skimming • Incentive to cost shift to services outside the PPPM • Can’t specifically promote desired activity • May resist innovation/ new services

6. Shared Savings • Organization or practice receives a bonus if actual spending comes in less than target • Advantages • -- Provider’s own experience determines target – achievable • -- Payer doesn’t spend more if no savings • -- No change in base payment methods -- good transition model • Disadvantages • -- No change in base payment methods – too marginal to work • -- No risk (although can be converted into 2-sided risk) • -- Lack of normative standards penalizes best performers

7. Episode/Condition/Bundle/Case Advantages − internalizes incentives for efficiency within the episode − potentially aligns incentives across siloed providers − arguably, is an intermediate step on the way to real integration Disadvantages − does not fundamentally alter incentive to generate units of service − be careful about what you wish for, e.g. physician-hospital alignment without determination of appropriateness in a FFS environment − currently, political challenges in bundling among providers − technically challenging (esp. for ambulatory care) – vagaries of diagnosis (more episodes in Miami than Minnesota), bias to performance of a procedure in a case rate, sorting out where particular claims are assigned to

8. Public Reporting and Pay-for-Performance (P4P) Advantages • provides a hybrid payment to mitigate disadvantages of pure models; some natural blends – PPPM and under-service measures • can start to actually reward desired performance, instead of rewarding volume of services produced • can include measures of patient experience, which have been overlooked as central to patient centeredness Disadvantages • underdeveloped measure set – especially for physicians • what gets measured gets done? • marginal incentives may be insufficient to counter incentives in the base model it is superimposed over • contributes more administrative complexity • evidence that process measures don’t predict outcomes

9. Examples of Blended or Hybrid Payment Models • PPPM with FFS carve outs or “bill aboves” and public reporting on underuse measures • For Patient-centered medical home, FFS for visits, PPPM for medical home activities and P4P for patient experience • Partial capitation – FFS/PPPM and/or risk corridors and/or particular sector (professional services, but not institutional) • Any of the above with public reporting and/or pay-for-performance − quality measures where they exist, expenditure or utilization targets, patient experience measures

10. Medicare Accountable Care Organizations: Core Characteristics Ability to provide and manage with patients the continuum of care across different institutional settings, at the very least, ambulatory and inpatient Capacity to prospectively set budgets and allocate resources to constituents of the ACO Sufficient size to support comprehensive, valid, and reliable quality measurement In Medicare, no restrictions or incentives to limit full freedom of choice Beneficiary assignment to an ACO based on statistical analysis of where beneficiaries get care – not election

11. Is ACO Just a New Term for PSO (Provider Sponsored Organization)? In BBA 1997, PSOs were created to permit Medicare to engage in financial risk contracting directly with providers They built it and no one came – actually 3 in 10 years Core differences with the ACA’s Shared Savings approach (Sec 3022) are absence of financial risk-taking and no restrictions on beneficiary choice (no gatekeepers).

12. What Else May Be Different? • Perhaps greater flexibility in organizational models • New payment models, no longer full capitation – e.g., FFS w. shared savings, “partial capitation” • Improved risk adjustment (although concern about providers direct incentive for code creep) • Availability of many more performance measures • Prospect of continued pressure on FFS rates, whatever is done with the SGR in Medicare • Prospect for continued reduction of Medicare Advantage overpayments

13. ACA Shared Savings Program Design Features • Real organizations – IPAs, multispecialty group practices, PHOs, joint ventures between hospitals and physician entities • Shared savings payment model – FFS with bonus if come in under a spending target – with a threshold for percentage saved before sharing and savings split to be decided in regulations • Builds on 3-year historic costs associated with patients assigned to ACO on the basis of claims patterns • ACOs need to hit threshold performance on 32 quality measures to be eligible for shared savings

14. Shared Savings Program (cont.) • Also testing two-sided risk (by year 4) and other forms of risk taking • This is a program, not a pilot or demo • Separately, the Center for Medicare and Medicaid Innovation at CMS supports Pioneer ACOs which are more mature organizations that have assumed risk • Altogether now about 150 ACOs serving about 2.5 million in Medicare – both hospital and medical group-based

15. How Would an ACO Work for Purchasers and Commercial Plans? Well-founded concern about Medicare “sanctioned” ACOs developing and using market power in negotiations with health plans to drive prices higher Providers want assurance of protection from FTC/DOJ Some view ACOs as community-wide – a way for competitors to come together to collaborate to reduce spending and increase quality Policy concern is ACOS might reduce costs but not provide the savings to private sector purchasers in reduced premiums

16. Potential ACO Market Power • Some prototypical ACOs obtain 200-250% of Medicare in their commercial insurance contracts • That is a lot to make up in lower utilization. Dartmouth-type utilization variations do not approach these price differentials • BUT the market power issue is not confined to ACOs – exists with non-integrated providers as well – single specialty physician mergers and multi-hospital systems exert market power