Slide 1 Bending the cost curve Addressing the problem of “supply-sensitive” care 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

Slide 2 Variations in practice and spending across U.S. Regions

Slide 3 The paradox of plenty: cross sectional evidence What do higher spending regions -- and systems -- get? Technical quality worse No more major elective surgery More hospital stays, visits, specialist use, tests, procedures 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 Physician-reported quality 5 Patient-reported quality 1,3 Lower satisfaction with hospital care Worse access to primary care (1) Ann Intern Med: 2003; 138: (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: More hospital beds per capita (32%) More medical specialists (65%) and internists (75%) Resource levels 1 Supply sensitive care

Slide 4 Trends in spending and quality What do higher spending regions -- and systems -- get? Regions with greatest spending growth had smallest gains in heart attack survival Skinner, Health Affairs, February 2006

Slide 5 Differences in spending What are the underlying causes? Explains less than 10% of state differences in spending Little impact on growth in utilization across states Malpractice environment 3,4 Capacity strongly correlated, but explains less than 50% Payment system ensures all stay busy Capacity / payment system 5 No difference in decisions with strong evidence More likely to intervene in “gray” areas (when to see patient, when to refer, when to admit) Clinical judgment 6,7 (1) Pritchard et al. J Am Geriatric Society; 46: , 199 (2) Anthony et al, under review (3) Kessler et al. Quarterly Journal of Medicine 1996;111(2): (4) Baicker, Chandra, NBER Working Paper W10709 (5) Fisher et al. Ann Intern Med: 2003; 138: (6) Sirovich et al. Archives of Internal Medicine. 165(19): (7) Sirovich et al, J Gen Intern Med. 2006;21(Suppl4):164. Slight preference for specialist care in high spending No difference for tests (if MD says not needed) No difference in preferences for aggressive EOL care Patient preferences? 1,2

Slide 6 Likely diagnosis Local capacity and culture drive practice and spending 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. 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. Payment system ensures that existing (and new capacity) is fully utilized -- and generously rewards growth.

Slide 7 Some examples A payment system that rewards growth and higher intensity care… Management of coronary artery disease -- the case of Elyria, Ohio Percutaneous Coronary Interventions Age-sex-race adjusted rate per 1000 enrollees in 2003

Slide 8 Some examples A payment system that rewards growth and higher intensity care… Management of coronary artery disease -- the case of Elyria, Ohio Percutaneous Coronary Interventions Age-sex-race adjusted rate per 1000 enrollees in 2003

Slide 9 Some examples A payment system that rewards growth and higher intensity care… Management of coronary artery disease -- the case of Elyria, Ohio New York Times, August 18, 2006

Slide 10 Some examples A payment system that rewards growth and higher intensity care… Management of coronary artery disease -- the case of Elyria, Ohio Use of erythropoetin (under current payment system) New York Times, May 9, 2007

Slide 11 Some examples A payment system that rewards growth and higher intensity care… Management of coronary artery disease -- the case of Elyria, Ohio Use of erythropoetin (under current payment system) Differences in use of physician workforce across academic medical centers Dartmouth Atlas of Health Care 2006 MayoDukeUCSFUCLACedars Hospital days (L6M)* Physician visits (L6M)* Medical specialist FTE (L2Y)** Primary care FTE inputs (L2Y)** Total Physician FTE (L2Y)** * Measures are per person / per decedent ** Measures are per 1000 decedents

Slide 12 Some thoughts on moving forward We need to consider underlying causes of rising costs, poor quality Failure to recognize key role of local system (capacity, clinical culture) as driver Assumption that more is better Equating less care with rationing Payment system that rewards more care, increased capacity, high margin treatments, entrepreneurial behavior Foster development of local organizations (delivery systems) accountable for care (with incentives to limit future growth) Balanced information on risks / benefits Comprehensive performance measures Reform of payment system (long term) Shared savings as interim approach Underlying causeGeneral Approach

Slide 13 Organizational accountability and incentives to slow growth Per-beneficiary spending in EHMS (n = 4772) sorted into quintiles by magnitude of per-beneficiary growth ( ) $4000 $3000 $ Average spending* on MD services per beneficiary at EHMS * Using standardized payments, using 2003 RVU ** Percent increase calculated relative to average 1999 per-beneficiary spending Absolute increase per benef. $936 $198 $431 $551 $675 Percent increase ** 46% 10% 21% 27% 33% Average Annual Rate 9.9% 2.4% 4.8% 6.1% 7.3%

Slide 14 Organizational accountability and incentives to slow growth Per-beneficiary spending in EHMS by BETOS category (highest and lowest quintiles of per-beneficiary growth ( ) Lowest Growth Quintile Medicare spending per enrollee Other Major Procedures Minor Procedures Tests Imaging E and M 6% 29% 18% 2% 0% 27% 10% Highest Growth Quintile 27% 80% 65% 38% 19% 116% 46% Percent increase in per-beneficiary spending Each Quintile includes approximately 20% of the Medicare population Differences in growth likely due to: active recruitment of physicians physician location decisions expansion of facilities (imaging) Control of spending will require altering incentives for growth

Slide 15 Payment reform Challenges and opportunities Barriers to comprehensive payment reform are substantial Public opposition to capitation; provider concern about bearing risk Development of other prospective payment approaches years away Might “shared savings” approaches help in the interim? Key notion: establish target growth rate; reward physician groups that achieve per-beneficiary spending growth below the target with portion of savings Theory being tested in the Physician Group Practice demonstration Has important advantages: Preserves fee-for-service payment (a plus for patients and MDs) Provides incentive to avoid increases in capacity (and to reduce capacity where feasible); and to improve care in domains previously ignored: care coordination, end-of-life care Can be done with existing claims data