Policy Implications of Mapping Healthcare Outcomes
John D Rockefeller JD MPHAssociate Dean and Lecturer
Geisel School of MedicineDartmouth College
4TH LATIN AMERICAN MEETING ON THE RIGHT TO HEALTH AND HEALTH SYSTEMS, Bogotá, Colombia. April 2 to 4, 2014
1973 – Measuring Health Care in Vermont
Wennberg J, Gittelsohn A. Small area variations in health care delivery. Science 1973;182:1102-8.
1993: From the ashes of the Clinton Health Care Reform was born the Dartmouth Atlas of Health Care
6 billion Medicare claims a year x many many years of data= lots of terabytes of data
www.dartmouthatlas.org
The Dartmouth Atlas of Health Care provides national public reporting of health system performance over time through the lens of variation in utilization, cost,
quality, and patient experience.
The Atlas highlights variation, its causes, and its consequences in order to provide target audiences with compelling data to effect positive changes in the
health care system.
2014: The Dartmouth Atlas of Health Care
Robert Wood Johnson FoundationCalifornia HealthCare Foundation
Charles H. Hood Foundation
Current Funders
Medicare is terrifically important, but so too are other populations:(U.S. Population by Insurance Type – 2010)
Total Population = 305 million
< 19 19-25 26-34 35-44 45-54 55-64 ≥ 65
Age
79m 30m 37m 40m 44m 37m 39mPopulation:
Medicare fee-for-service
Medicare “HMO”
Medicaid
Private/commercial
Uninsured
The Dartmouth Atlas of Health Care
David Goodman, MD MS (Co-PI)Elliott Fisher, MD MPH (Co-PI)Jonathan Skinner, PhDJohn Wennberg, MD MPH (Founder)Kristen Bronner, MA (Managing Editor)Scott Chasan-Taber, PhD (Director of Atlas Analytics)
Julie Bynum, MD MPHNancy Morden, MD MPHShannon Brownlee, MSChiang-hua Chang, PhDTherese Stukel, PhDJeff Munson, MDJohn Erik-Bell, MD MSDouglas Staiger, PhDJames Weinstein, MD MSPhil Goodney, MD MS
Leadership Group Faculty (a dynamic cohort)
6
Elisabeth Bryan, MSThomas Bubolz, PhDDonald Carmichael, MDivJulie DohertyJennifer Dong, MSDaniel Gottlieb, MSJia Lan, MSMartha Lane, MAStephanie Raymond, MA
Nancy Marth, MSSally Sharp, SMJeremy Smith, MPHYunjie Song, PhDDean Stanley, RHCEAndrew Toler, MSStephanie Tomlin, MPARebecca Zaha, MPHWeiping Zhou, MS
The Amazing Staff
Price-adjusted Medicare spending per beneficiaryamong hospital referral regions (2010)
$10,420 to 13,830 (61)9,770 to < 10,420 (62)8,920 to < 9,770 (60)8,100 to < 8,920 (61)6,910 to < 8,100 (62)
Not populated
What causes the variation in spending?
What is the right rate?
How can we make fair comparisons?
Are there different causes of variation in utilization?
How can we improve the value of health care?
Percent of Medicare diabetics with eye exams hospital service areas (2010)
Effective care
October 15, 2013
tdi.dartmouth.edu
Use of drugs to treat osteoporosis following fragility fracture among hospital referral regions (2006-10)
17.4 to 28 .1% (48)15.2 to < 17 .4% (50)13.8 to < 15 .2% (46)12.3 to < 13 .8% (50)
6.8 to < 12 .3% (48)Insufficient data (64)Not populated
Effective care
Use of beta-blockers 7-12 months following discharge for AMI (2008-10)
92 % or More (0)84 % to < 92 % (42)76 % to < 84 % (164)68 % to < 76 % (86)Less than 68% (13)Insufficient data (1)Not populated
Effective care
Quality Dartboards for large Northern New England hospital service areas
Children < 18 yrs – all payer claims data, 2007-10 average
Quality Dartboards for large Northern New England hospital service areas
Children < 18 yrs – all payer claims data, 2007-10 average
Variation in Effective Care
• The choice of service is dictated by strong evidence of effectiveness for almost all targeted patients.
• The benefits almost always outweigh any adverse effects.• Risk adjustment is often not necessary.• The right rate is usually obvious.
17
TURP for BPH discharges per 1,000 male Medicare enrollees (2007)
age-sex-race adjusted
Red dots indicate highest 3, lowest, and HRRs with at least 300,000 FFS Medicare beneficiaries
Idaho Falls, ID 2.79Panama City, FL 2.56Gulfport, MS 2.46Boston, MA 1.19St. Louis, MO 1.07Camden, NJ 1.07Houston, TX 1.05Los Angeles, CA 1.04Manhattan, NY 1.03Philadelphia, PA 1.01Indianapolis, IN 1.00East Long Island, NY 0.82Orlando, FL 0.77Atlanta, GA 0.71Dallas, TX 0.62Salinas, CA 0.24
0.0
0.5
1.0
1.5
2.0
2.5
3.0
Rat
io o
f TU
RP
for
BP
H r
ate
to U
.S.
aver
age
18
Preference-Sensitive Care
• Involves tradeoffs.• Scientific uncertainty often substantial.• The effect of supply (e.g. physicians) is variable.• Patient and provider values are often different.• Decisions that should be based on the patient’s own
preferences.• Decision quality is improved through shared decision-
making and decision aids.
19
Percent of cancer patients dying in hospital among academic medical centers and NCI Cancer Centers (2010)adj. for age-sex-race, cancer type, non cancer conditions
10
15
20
25
30
35
40
45
50
55
Perc
ent d
ying
in h
ospi
tal
Allegheny General Hospital (Pittsburgh, PA) 16.6Univ Hospitals of Cleveland (Cleveland, OH) 16.5Univ of Kentucky Hospital (Lexington, KY) 16.5Akron General Medical Center (Akron, OH) 12.2St. Luke's Hospital (Bethlehem, PA) 11.5
Lenox Hill Hospital (New York, NY) 49.8Maimonides Medical Center (Brooklyn, NY) 48.1New York Methodist Hospital (Brooklyn, NY) 47.3Mount Sinai Hospital (New York, NY) 46.9Beth Israel Medical Center (New York, NY) 44.9
2.0
4.0
6.0
8.0
10.0
12.0
3.0 6.0 9.0 12.0 15.0 18.0
Acute Myocardial InfarctionRate per 1,000 Medicare Enrollees
age-sex-race adjusted
Car
diol
ogis
ts p
er 1
00K
Source: Wennberg, et al. Dartmouth Cardiovascular Atlas
There is virtually no relationship between regional physician supply and health
needs.
Capacity (i.e. supply) is often located without respect for need
22
R2 = 0.49
Nu
mb
er
of V
isits
per
ben
efic
iary
0.0
0.5
1.0
1.5
2.0
2.5
0.0 2.5 5.0 7.5 10.0 12.5 15.0
Number of Cardiologists per 100,000
Physician Supply and Physician Visitsage-sex-race adj.
Cardiologists
Head CT Scans per 1,000 Children(2007-10, age-sex-payer adj.)
14.7 to 19.7 (13)12.3 to < 14.7 (14)10.5 to < 12.3 (14)
8.9 to < 10.5 (14)4.2 to < 8.9 (13)
Insufficient data (1)Not populated
3
5
7
9
11
13
15
17
19
21
Hea
d CT
sca
ns p
er 1
,000
chi
ldre
n
Bangor, ME 11.1Portland, ME 9.7Lebanon, NH 8.9Burlington, VT 8.4