1. Spending, Quality and Efficiency of Care Addressing the paradox of plenty Elliott S. Fisher, MD, MPH Professor of Medicine Center for the Evaluative Clinical Sciences Dartmouth Medical School Senior Associate VA Outcomes Group White River Junction, Vermont CECS Center for the Evaluative Clinical Sciences

2. Variations in practice and spending The Dartmouth Atlas

3. Variations in practice and spending Two to three fold variations in spending across regions 4,000 6,000 8,000 10,000 12,000 Medicare reimbursements per enrollee Miami, FL$11,352 Los Angeles, CA $9,752 Worcester, MA $8,203 Boston, MA $7,901 Springfield, MA $7,103 San Francisco, CA $6,408 Minneapolis, MN $5,213

4. How can the best medical care in the world cost twice as much as the best medical care in the world? Uwe Reinhardt Cedars-Sinai76,934 UCLA Medical Center72,793 New York-Presbyterian69,962 Johns Hopkins60,653 UCSF Medical Center56,859 Univ. of Washington50,716 Mass. General47,880 Barnes-Jewish44,463 Duke University Hosp.37,765 Mayo Clinic (St. Mary's)37,271 Cleveland Clinic35,455 20,000 40,000 60,000 80,000 100,000 120,000 Inpatient + Part B spending per decedent Spending per Medicare beneficiary with severe chronic disease (Last 2 years of life, 2000-2003) Variations in practice and spending and across American’s Best Hospitals (USN&WR + Cedars)

5. Variations in practice and spending Insights, challenges and opportunities 1. Variations in spending -- implications for health 2. What’s going on? -- some findings and hypotheses 3. What might we do? John Wennberg, MD, MPH Dartmouth Medical School Julie Bynum, MD, MPH Dartmouth Medical School Eric Holmboe, MD American Board of Internal Medicine Rebecca Lipner, PhD American Board of Internal Medicine David Wennberg, MD, MPH Maine Medical Center Lee Lucas, PhD Maine Medical Center Dan Gottlieb, MS Dartmouth Medical School Amber Barnato, MD, MPH University of Pittsburgh Therese Stukel, PhD University of Toronto Brooke Herndon, MD Dartmouth Medical School Jonathon Skinner, PhD Dartmouth Medical School Elliott Fisher, MD, MPH VA Outcomes Group, Dartmouth Denise Anthony, PhD Dartmouth College Brenda Sirovich, MD, MS VA Outcomes Group, Dartmouth Doug Staiger, PhD Dartmouth College Amitabh Chandra, PhD Harvard University Jack Fowler, PhD University of Massachusetts, Boston Patricia Gallagher, PhD University of Massachusetts, Boston Renee Mentnech, PhD Center for Medicare and Medicaid Services Causes and Consequences of Health Care Intensity Dartmouth Atlas of Health Care National Institute on Aging Robert Wood Johnson Foundation California Healthcare Foundation Wellpoint Foundation Aetna Foundation United Health Foundation Commonwealth Fund With support from: Investigators

6. Variations in spending What are the implications for health? Differences in spending largely due to differences in overall quantity of care (intensity) provided to similar populations. Key Question: What does more spending -- greater intensity -- buy? Ann Intern Med: 2003; 138: 273-298 N Engl J Med 2004; 349;17:1665-1667 Health Affairs web exclusives, October 7, 2004 Health Affairs, web exclusives, Nov 16, 2005 Health Affairs web exclusives, Feb 7, 2006 Ann Intern Med: 2006; 144: 641-649

7. Variations in spending What are the implications for health? Study population -- Medicare enrollees Acute myocardial infarctionn = 159,393 Colorectal Cancer n = 195,429 Hip Fracturen = 614,503 Medicare Current Beneficiary Surveyn = 18,190 Study design -- natural experiment: Divided populations into five equal groups according to practice intensity of region of residence Practice intensity measured in different population (other Medicare enrollees in last six months of life)

8. Variations in spending Content of care -- three categories Effective care:Evidence-based services that all patients should receive. No tradeoffs involved. Acute revascularization for AMI Preference-sensitiveTreatment choices that entail tradeoffs among carerisks and benefits. Patients’ values and preferences should determine treatment choice. CABG for stable angina Supply-sensitive Services where utilization is strongly associated serviceswith local supply of health care resources Frequency of MD visits, specialist consultations use of hospital or ICU as a site of care Wennberg, Skinner and Fisher, Geography and the Debate over Medicare Reform Health Affairs, web exclusives, February13, 2002

9. Effective Care: Ratio of Rates in Highest vs Lowest Spending Regions 1.001.52.00.52.53.0 1.001.52.00.52.53.0 Reperfusion in 12 hours for AMI Acute MI Lower in High Spending Regions Higher in High Spending Regions Quintile 1 Quintile 5 55.849.8

10. Effective Care: Ratio of Rates in Highest vs Lowest Spending Regions 1.001.52.00.52.53.0 1.001.52.00.52.53.0 Reperfusion in 12 hours for AMI Aspirin at discharge Aspirin at admission Beta Blocker at discharge Beta Blocker at admission Acute MI Lower in High Spending Regions Higher in High Spending Regions ACE Inhibitor at discharge

11. Effective Care: Ratio of Rates in Highest vs Lowest Spending Regions 1.001.52.00.52.53.0 1.001.52.00.52.53.0 Reperfusion in 12 hours for AMI Aspirin at discharge Aspirin at admission Beta Blocker at discharge Beta Blocker at admission Acute MI Mammogram, Women 65-69 Flu shot during past year Pap Smear, Women 65+ Pneumococcal Immuniztation (ever) General Population Lower in High Spending Regions Higher in High Spending Regions ACE Inhibitor at discharge

12. Effective Care: Association between spending and overall quality rank Baicker and Chandra, Health Affairs, web exclusives

13. Preference-Sensitive Care: Highest vs Lowest Spending Regions 1.001.52.00.52.53.0 1.001.52.00.52.53.0 Coronary Artery Bypass Surgery (CABG) Coronary Angioplasty Procedures after AMI Cholecystectomy Hernia Repair Cataract Extraction Total Hip Replacement Major Surgery (all cohorts combined) Total Knee Replacement Back Surgery Carotid Endarterectomy Lower in High Spending Regions Higher in High Spending Regions Angiography Angiography among appropriate cases

14. Supply-Sensitive Care : Highest vs Lowest Spending Regions 1.001.52.00.52.53.0 1.001.52.00.52.53.0 Office Visits Initial Inpatient Specialist Consultations Inpatient Visits Physician Visits Lower in High Spending Regions Higher in High Spending Regions Electrocardiogram Tests and Procedures CT / MRI Brain Pulmonary Function Test Electroencephelogram (EEG) Discharges Inpatient Days in ICU or CCU Total Inpatient Days Hospital Utilization Feeding Tube Placement Emergency Intubation Procedures -- Last 6 months of life

15. Variations in spending What do higher spending regions -- and systems -- get? Technical quality worse No more elective surgery More hospital stays, visits, specialist use, tests Content / Quality of Care 1,2 Health Outcomes 1,2 (1) Ann Intern Med: 2003; 138: 273-298 (2) Health Affairs web exclusives, October 7, 2004 (3) Health Affairs, web exclusives, Nov 16, 2005 (4) Health Affairs web exclusives, Feb 7, 2006 (5) Ann Intern Med: 2006; 144: 641-649

16. Variations in spending What do higher spending regions -- and systems -- get? Technical quality worse No more elective surgery More hospital stays, visits, specialist use, tests Content / Quality of Care 1,2 Slightly higher mortality No better function Health Outcomes 1,2 Worse communication among physicians Greater difficulty ensuring continuity of care Greater difficulty providing high quality care Greater perception of scarcity Physician’s perceptions 5 Patient-perceived quality 1,3 Lower satisfaction with hospital care Worse access to primary care Trends over time 4 Greater growth in per-capita resource use Lower gains in survival (following AMI) (1) Ann Intern Med: 2003; 138: 273-298 (2) Health Affairs web exclusives, October 7, 2004 (3) Health Affairs, web exclusives, Nov 16, 2005 (4) Health Affairs web exclusives, Feb 7, 2006 (5) Ann Intern Med: 2006; 144: 641-649

17. Major points Higher spending across regions and physician groups is largely due to overuse of supply-sensitive services -- hospital and ICU stays, MD visits, specialist consults; and more is worse.

18. What’s going on? What explains the differences in practice?

19. Patient preferences -- can’t explain the differences observed

20. What’s going on? What explains the differences in practice? Patient preferences -- can’t explain the differences observed Capacity and payment

21. What’s going on? What explains the differences in practice? Patient preferences -- can’t explain the differences observed Capacity and payment 1.0 3.0 4.0 10 40 50 30 20 2.0 Hospital Beds Medical Specialists LowHighLowHigh 32% higher 65% higher Regional Spending

22. What’s going on? What explains the differences in practice? Patient preferences -- can’t explain the differences observed Capacity and payment Whatever capacity is in place will be fully utilized

23. What’s going on? What explains the differences in practice? Patient preferences -- can’t explain the differences observed Capacity and payment -- are important drivers

24. What’s going on? What explains the differences in practice? Patient preferences -- can’t explain the differences observed Capacity and payment -- are important drivers Clinical decision-making

25. What’s going on? What explains the differences in practice? Patient preferences -- can’t explain the differences observed Capacity and payment -- are important drivers Clinical decision-making

26. What’s going on? What explains the differences in practice? Patient preferences -- can’t explain the differences observed Capacity and payment -- are important drivers Clinical decision-making -- in the gray areas -- is critical

27. What’s going on? What explains the differences in practice? Patient preferences -- can’t explain the differences observed Capacity and payment -- are important drivers Clinical decision-making -- in the gray areas -- is critical Training environment -- preliminary findings underscore influence of local systems

28. Putting together a story… Physician - Patient Encounter Clinical Evidence Professionalism Clinical evidence (e.g. RCTs, guidelines) and principles of professionalism are a critically important -- but limited -- influence on clinical decision-making. Current payment system fosters growth and ensures that existing (and new) capacity is fully utilized. Consequence: reasonable individual clinical and local decisions lead, in aggregate, to higher utilization rates, greater costs -- and inadvertently -- worse outcomes Local Organizational Context (e.g. capacity - culture) Policy Environment (e.g. payment system) Physicians practice within a local organizational context and policy environment that profoundly influences their decision-making.

29. Major points Higher spending across regions and physician groups is largely due to overuse of supply-sensitive services -- hospital and ICU stays, MD visits, specialist consults; and more is worse. Overuse is largely a consequence of differences in clinical judgment (not outright errors) that arise in response to local organizational attributes (capacity, clinical culture) and state / national policies promoting growth and more care.

30. What can be done? Levels of decision-making -- and potential strategic levers Physician - Patient Encounter Local Organizational Context (e.g. capacity - culture) Policy Environment (e.g. payment system) Research priorities (disease biology - clinical practice) Coverage policy Performance measurement / Public reporting Payment system reform Recruitment / practice location decisions Capital investment (hospital, outpatient) Organizational structure (hospital, MD group) Process management (QI, IT adoption) Specialty certification Graduate Medical Education Continuing Medical Education HIT for care and decision-support Informed consumers / shared decision-making

31. What can be done? Levels of decision-making -- and potential strategic levers Physician - Patient Encounter Local Organizational Context (e.g. capacity - culture) Policy Environment (e.g. payment system) Research priorities (disease biology - clinical practice) Coverage policy Performance measurement / Public reporting Payment system reform Recruitment / practice location decisions Capital investment (hospital, outpatient) Organizational structure (hospital, MD group) Process management (QI, IT adoption) Specialty certification Graduate Medical Education Continuing Medical Education HIT for care and decision-support Informed consumers / shared decision-making

32. Organizational accountability Foster accountability for quality and costs Policy initiatives should focus on fostering organizational and professional accountability for longitudinal quality and costs. Formal: Prepaid / multi-specialty group practices (e.g Kaiser) Virtual:Hospitals and their affiliated physicians Hospitals / Medical Staff Majority of physicians work in or admit to only one hospital Chronic disease patients are highly loyal -- allowing comparisons of longitudinal costs and quality Performance measurement -- and payment reform -- would create incentives for hospital and staff to collaborate to improve quality Provides organizational context for capacity management -- and for implementation of information technology, QI, shared decision- making

33. www.dartmouthatlas.org Dartmouth Atlas of Health Care The care of patients with severe chronic illness Goal -- provide hospital specific measures of relative intensity of resource use Approach -- measure resource use in severely ill patients Assign Medicare beneficiaries to hospitals based upon predominant site of care during last 2 years of life (with chronic illness) Adjust for differences in underlying illness Measures include: Medicare reimbursements, utilization rates. Importance Measures reflect relative intensity and costs for other populations Provide insight into volume of supply-sensitive services (a reflection of capacity and culture)

34. Spending and utilization among severely ill patients in selected U.S. and Massachusetts hospitals *last two years of life, all other measures are during last six months of life www.dartmouthatlas.org Cleveland Clinic Cedars Sinai Part B spending* $19,427$6,490 Hospital days 23.114.6 Primary care visits 16.112.9 Medical specialist visits 51.015.0 Physician visits 71.332.1 Percent hospice16.834.2

35. Spending and utilization among severely ill patients in selected U.S. and Massachusetts hospitals *last two years of life, all other measures are during last six months of life www.dartmouthatlas.org Saint Elizabeth’s BaystateMGH Cleveland Clinic Cedars Sinai Cooley Dickinson Part B spending* $12,292$9,519$10,316$8,840$19,427$6,490 Hospital days 18.8 11.7 17.712.123.114.6 Primary care visits 25.4 16.1 19.315.016.112.9 Medical specialist visits 14.5 12.3 19.58.051.015.0 Physician visits 43.3 32.4 42.026.271.332.1 Percent hospice16.520.119.621.816.834.2

36. Major points Higher spending across regions and physician groups is largely due to overuse of supply-sensitive services -- hospital and ICU stays, MD visits, specialist consults; and more is worse. Overuse is largely a consequence of reasonable differences in clinical judgment (not errors) that arise in response to local organizational attributes (capacity, clinical culture) and state / national policies promoting growth and more care. Improving efficiency will require fostering local organizational accountability for the longitudinal costs and quality of care. Performance measurement, public reporting, payment reform and technical assistance should be aligned toward this goal.