Linda Magno Debbie Peikes Arnold Chen Jennifer Schore Randy Brown

1. Don’t These Demonstrations Ever Work? Mixed Evidence from the Four-Year Medicare Coordinated Care DemonstrationIntroductionAcademyHealth Annual ConferenceJune 9, 2008 Linda Magno Debbie Peikes Arnold Chen Jennifer Schore Randy Brown

2. Roadmap • Background • Impacts on Service Use/Cost • Impacts on Quality of Care • What Distinguishes Effective Programs • Conclusions and Ongoing Work

3. BackgroundTheory Behind CC for Medicare FFS • Problem: Rapidly increasing Medicare costs • Chronically ill account for 75% of expenditures: • Half of beneficiaries have 1+ (of 8) conditions • 12% have 3+ and account for 1/3 of all costs • High rates of inpatient admissions • Many seem preventable • Often preceded by non-adherence, failure to recognize warning signs • Patients see 5+ physicians per year

4. Causes of “Preventable” Costs • Difficulty adhering to drugs/diets/self-care advice • Care not always evidence-based • Some patients lack transportation, support services • Patients and providers communicate poorly: • Patients don’t call soon enough or divulge fully • Providers don’t ensure patient understands • Providers don’t talk to each other (no incentives) • Typical advice if no appointments: “Go to the ER”

5. The Promise of Coordinated Care

6. Why Medicare Investigated CC • Intuitive appeal • Potential to improve lives and reduce costs • Claims of huge effects in other markets • HMOs and employers are buying it: • 1997: $78 million • 2000: $1.2 billion (2008: est. $1.8 billion) • Large, identifiable target population

7. Extension/Expansion • Secretary must extend/expand projects if initial evaluation (first 2 years) found • Savings • Budget neutrality plus improved quality and beneficiary/provider satisfaction • Secretary may, by regulation, incorporate beneficial components of projects into Medicare program on permanent basis

8. Goals of the Demonstration • CMS hoped to learn: • Do the programs improve quality? • Do the programs reduce gross cost? • Are the programs budget-neutral? • What program types/features work best? • What types of patients do they work for?

9. The Demonstration Programs • 15 were selected in January 2002 • Wide variation in negotiated fees: $80 to $444 PMPM (average = $235) • Voluntary enrollment model

10. Program Hosts Represented a Variety of Organizations • 5 commercial CC/ DM providers • 3 academic medical centers • 4 hospitals/ integrated systems • Others: hospice, retirement community, long-term care facility

11. MCD JHH Avera Mercy HQP Carle U of Md Washington University Georgetown QMed Charlestown Hospice CenVaNet CorSolutions Quality Oncology Programs Served 16 States + D.C. Hospice = Hospice of the Valley; HQP = Health Quality Partners; JHH = Jewish Home and Hospital Lifecare System; MCD = Medical Care Development; U of Md = University of Maryland.

12. Nurses as Care Coordinators • Staff were primarily registered nurses; most had cardiac or geriatric experience • Caseloads varied from 36 to 155; half were between 60 and 86 • Program patients did not “graduate” • Most contact was by telephone

13. Programs Varied Widely on Key Dimensions • Few had sophisticated IT or home telemonitoring • 12 programs drew patients from physicians they had experience with • Programs focused on teaching patient about self care and communication • Service arrangement was not a focus • Few had medication lists from providers • Enrollment varied widely • 3 served 95 to 115 • 9 served 415 to 725 • 3 served 1,100 to 1,500