American Healthcare in Transition: Exploring the Evolution of National Health Expenditures from 1960 to the Present

1
American Healthcare in Transition Exploring the
Evolution of National Health Expenditures from
1960 to the Present
Jack Homer with Gary Hirsch and Bobby
Milstein System Dynamics Winter Camp Austin,
TX January 5-6, 2007
2
Background

• Facts
• U.S. has worlds highest healthcare costs, but
  our health is only so-so
• Most spending is for treating existing disease
  (downstream) rather than for prevention
  (upstream)
• Most spending is for chronic illness, not acute
  infections or injuries
• 15 of Americans have no health insurance
  coverage
• Are these facts connected? Some preliminary
  thinking
• Upstream/downstream loops CDC 2003, AJPH 2006
• Healthcare system loops HPSIG/ISDC 2005
• Now, we want to refine our thinking in light of
  historical data and via model-based
  theory-testing
• Key data source National Health Expenditures
  (NHE) 1960-2004 from Centers for Medicare and
  Medicaid Services (CMS)

Homer JB, Hirsch GB. System dynamics modeling
for public health Background and opportunities.
American J Public Health 2006 96 452-458.
3
Healthcare stakeholder map (presented at ISDC
2005)
Insurers/Payers (Public, Private) - Reimbursement
criteria rates for risk disease mgmt
and urgent care - Number of competitors

• Health Care Costs
• - Risk disease mgmt
• - Urgent care
• Administrative
• Capital investments

Providers (MDs, RNs, Hospitals) - Risk disease
mgmt extent and efficacy - Urgent care extent
and efficacy - Specialty fragmentation -
Lobbying of insurers regulators
Employers - Health coverage
Drug/Device Makers - Developing high-tech
products for urgent care and risk
disease mgmt - Lobbying of insurers
regulators
General Public - Improvement of living
conditions Funds available Citizen
involvement
Regulators Monitors (Public, Private) - Usage
guidelines controls
Patients - Health and risk status
4
Stock-flow view of disease and spending
Urgent care and disease management both prolong
the lives of people with disease. Urgent care
unequivocally raises healthcare spending, whereas
disease mgmt. can be a net cost saver, because it
prevents expensive urgent care.
5
Closing the loops Initial dynamic hypothesis
We expected to see naturally greater reinvestment
in urgent care than in disease/risk mgmt., and
also selective squeezing out of D/R mgmt. as
healthcare coverage declines.
6
Personal healthcare spending per capita
1960-2004, by National Health Expenditures
components (in year 2000 dollars personal
healthcare does not include admin., public
health, research, capital investments)
Personal healthcare consistently accounts for
83-85 of all health spending. The fastest
growth was in hospital care 1960-82, non-hospital
services 1983-94, and prescription drugs
1995-2004.
7
Estimated urgent care vs. disease/risk management
portions of personal healthcare spending (in
year 2000 dollars)
Urgent care includes all hospital services plus
some fraction of non-hospital services. For
2002-04, we roughly estimate that fraction as
30. This makes urgent care about 50 of
spending, the other 50 being for disease/risk
management. We posit D/R mgmt. as proportional
to Rx drug spending, which accounted for 12 of
spending in 2002-04. Thus, the estimated ratio of
D/R mgmt to Rx drugs (50/12) 4.2.
8
Self-reported health status, 1993-2004 (National
Health Interview Survey for G/F/P since
1997, Behavioral Risk Factor Surveillance System
for other)
The fraction of adults with health less than
very good has increased steadily since 1993.
9
Prevalence of cardiovascular risk factors,
1991-1999 (National Health Interview Survey)
The fraction of adults with 1 or more risk
factors has grown from 58 to 62, and the
fraction with 2 or more risk factors from 24 to
28. All risk factors grew except smoking (which
declined only 1 percentage point in the 90s),
with obesity being the largest contributor to the
overall growth in risk factor prevalence.
10
Health insurance coverage, 1987-2004 (US Census)
During the 1987-2003 period, private coverage
fell 7 percentage points, while total coverage
fell only 3 percentage points. Thus, more than
half of those who have lost private insurance
have had government coverage, generally Medicaid,
to fall back on.
11
Revised dynamic hypothesis
Increasingly high costs led to two reactions by
insurers first, overall reimbursement
restrictions, and then, a shifting of priorities
toward D/R mgmt away from urgent care.
12
What-if tests for understanding causal
contributions

• Base
• No invest cut (No cut in investment rate)
• Fixed 35 investment rate (Base declines to
  15 by 2005)
• No DRM shift (No shift toward disease/risk
  management)
• Fixed 79 of investments to Urgent Care
  (Base declines to 39 by 1995 before rebounding
  to 69 by 2005)
• No covg down (No decline in private coverage
  after 1975)
• Private coverage remains at 85 (Base declines
  to 73 by 2005)
• No obese up (No exogenous increases in risk and
  disease onset)
• Onset multipliers remain at 1 (Base risk onset
  multiplier increases to 1.5 during 1980-2005,
  disease onset multiplier increases to 1.33 during
  1990-2005)

13
What-if test results Healthcare spending,
coverage, disease prevalence, and deaths
No covg down
No invest cut
Base
No obese up
No DRM shift
No DRM shift
No invest cut
No covg down
No obese up
Base
No DRM shift
Base
Shift to D/R mgmt improved health with little
increase in cost
No DRM shift
No covg down
No covg down
Coverage cutback restrained costs a bit but also
slowed health gains somewhat
No invest cut
No obese up
Base
No obese up
No invest cut
Investment cutback restrained costs but also
slowed health gains
Obesity epidemic has been a major driver of
disease and cost
14
Feedback policies for spurring non-medical
upstream efforts at health protection
15
ADDITIONAL SLIDES
16
Anti-hypertensive anti-cholesterol drug
spending as a fraction of all Rx drug spending,
1980-2004 (numerators based on pharmaceutical
industry reports, denominator from NHE)
The combined fraction serves as our estimate of
risk management as a fraction of D/R
management. Why were the anti-hypertensives so
fast to emerge relative to the rest of D/R
management? Perhaps because of their broad
applicability, in both symptomatic and
asymptomatic cases of hypertension and heart
failure.
17
Simulated history (1) Healthcare spending,
Assets, and Coverage
Total assets
Urgent care
Total spending-sim
Total-data
D/R mgmt
Urgent-data
Urgent-sim
D/R mgmt-sim
D/R mgmt-data
Total coverage-sim
Total-data
Private-sim
Private-data
Govt only-data
Govt only-sim
18
Simulated history (2) Disease risk
prevalence, Death rate, and Effects of medical
care/mgmt.
Death rate per total popn-sim
Death rate-data
Per hospitalization-data
Disease or at risk prevalence-sim
Per urgent episode-sim
Any cardiovascular risk-data
Disease-sim
Less than very good health-data
Eff of risk mgmt on disease onset
Eff of urgent care on fatality
Eff of dis mgmt on urgent episodes
19
Effectiveness of care is determined by assets
(equipment, skills) and insurance coverage
20
(No Transcript)
21
(No Transcript)
22
Rapid growth in health spending
The health sector now employs more people than
any other sector of the US economy and tripled
its share of GDP from 1960 to 2000.
Heirich M. Rethinking health care innovation and
change in America. Boulder CO Westview Press,
1999. Pear R. Health spending rises to record 15
of economy. The New York Times 2004 January 9.
23
And the trend is projected to continue.
Centers for Medicaid and Medicare Services.
Health accounts. Centers for Medicaid and
Medicare Services, 2004. http//www.cms.hhs.gov/st
atistics/nhe/projections-2003/t1.asp
24
Downstream efforts have led to major achievements
Population Death Rate from Coronary Heart
Disease, 19501998
700
600
500
400
Age-adjusted Death Rate per 100,000 Population
300
200
100
50
1950
1960
1970
1980
1990
1995
1955
1965
1975
1985
Year
Marks JS. The burden of chronic disease and the
future of public health. CDC Information Sharing
Meeting. Atlanta, GA National Center for Chronic
Disease Prevention and Health Promotion 2003.
25
But health-related quality of life has worsened
14 increase
Source Centers for Disease Control and
Prevention. Health-related quality of life
prevalence data. National Center for Chronic
Disease Prevention and Health Promotion, 2003.
Accessed March 21 at lthttp//apps.nccd.cdc.gov/HRQ
OL/gt.
26
Upstream work is a very small fraction of health
spending
Upstream Prevention and Protection --------------
--------------------- Total ? 3
Downstream Care and Management ------------------
-------------- Total ? 97
Brown R, Elixhauser A, Corea J, Luce B, Sheingod
S. National expenditures for health promotion and
disease prevention activities in the United
States. Washington, DC Battelle Medical
Technology Assessment and Policy Research Center
1991. Report No. BHARC-013/91-019.
27
Why so little upstream work?

• Health professionals focus on disease
  management and care, where their expertise, the
  weight of scientific evidence, and the urgency
  lie.
• Upstream work requires public concern and
  citizen organizing. But public concern is
  diffuse and not necessarily focused on health
  issues.

As a result, the health system naturally tends
toward managing affliction rather than preventing
incidence and protecting against vulnerability.
28
The U.S. health system is resistant to change
At least six times since the Depression, the
United States has tried and failed to enact a
national health insurance program.
-- Lee Paxman
Lee P, Paxman D. Reinventing public health.
Annual Reviews of Public Health 1997181-35.
29
Types of healthcare reform initiatives (ISDC 2005)

• Expanding access
• Improving coverage to employees, the poor,
  children
• Providing health care resources to inner cities
  and rural areas
• Containing cost
• Government limits on capacity, service provision,
  or reimbursement
• Employer shift to managed care plans
• Improving quality of care
• State regulation of facilities, professional
  licensure, Medicaid quality monitoring
• JCAHO setting of standards, NCQA evaluation of
  managed care orgs
• Protecting health
• Risk management, promotion of healthy lifestyles,
  family planning
• Safer workplaces, better housing, safer
  neighborhoods