The Primacy of Primary Care in Health Services Systems

1
The Primacy of Primary Care in Health Services
Systems
Barbara Starfield, MD, MPH November 2002
2
The Countries, Mid-1990s
Australia Belgium Canada Denmark Finland France Ge
rmany Japan Netherlands Spain Sweden United
Kingdom United States
Starfield
3
Reimbursement of Physicians
Primary Care Specialists

• 1980s 1990s 1980s 1990s
• Belgium F F F F
• France F F F F
• Germany F F F F
• United States F F F F
• Australia F F F F
• Canada F F F F
• Japan F F F F
• Sweden S S S S
• Denmark C/F F/C S/F F
• Finland S/C S/C S S
• Netherlands C/F C/F F S
• Spain S S S S
• United Kingdom C C/F S S

Starfield 2000
4
Are Specialists Limited to Hospital
Practice? Early-Mid 1990s
Belgium France Germany United States No No No No
Australia Canada Japan Sweden No No No No
Denmark Finland Netherlands Spain United Kingdom Yes Yes Yes No Yes
Starfield 2000
5
Primary Care Orientation of Health Systems
Rating Criteria

• Health System Characteristics
• Type of system
• Financing
• Type of primary care practitioner
• Percent active physicians who are specialists
• Professional earnings of primary care physicians
• relative to specialists
• Cost sharing for primary care services
• Patient lists
• Requirements for 24-hour coverage
• Strength of academic departments of family
  medicine

Source Starfield, 1998.
Starfield
6
Primary Care Orientation of Health Systems
Rating Criteria

• Practice Characteristics
• First-Contact
• Longitudinality
• Comprehensiveness
• Coordination
• Family-centeredness
• Community orientation

Source Starfield, 1998.
Starfield
7
Primary Care Scores, 1980s and 1990s
1980s
1990s

• Belgium 0.8 0.4
• France - 0.3
• Germany 0.5 0.4
• United States 0.2 0.4
• Australia 1.1 1.1
• Canada 1.2 1.2
• Japan - 0.8
• Sweden 1.2 0.9
• Denmark 1.5 1.7
• Finland 1.5 1.5
• Netherlands 1.5 1.5
• Spain - 1.4
• United Kingdom 1.7 1.9

Starfield 10/02
8
System and Practice CharacteristicsFacilitating
Primary Care, Early-Mid 1990s
GER
FR
BEL
US
SWE
JAP
CAN
FIN
AUS
SP
NTH
DK
UK
Starfield 11/00
9
Health Care Expenditures per Capita, 1996
Belgium France Germany United States ----- ----- ----- ----- 1693 1978 2222 3708
Australia Canada Japan Sweden ----- ----- ----- ----- 1776 2002 1581 1405
Denmark Finland Netherlands Spain United Kingdom ----- ----- ----- ----- ----- 1430 1389 1756 1131 1304
Starfield 2000
10
Primary Care Score vs. Health Care Expenditures,
1997
UK
DK
NTH
FIN
SP
CAN
AUS
SWE
JAP
GER
US
BEL
FR
11
Primary Care OrientationandPopulation Health
Status
Starfield 10/02
12
Average Rankings for Health Indicators in
Infancy, for Countries Grouped by Primary Care
Orientation
Infant Mortality (1996)
Postneonatal Mortality (1993)
Neonatal Mortality (1993)
Low Birth Weight (1993)
8.8
11.5
7.8
9.5
Lowest (Belgium, France, Germany, US)
6.0
5.5
5.3
7.3
Middle (Australia, Canada, Japan, Sweden)
6.4
4.6
7.8
4.8
Highest (Denmark, Finland, Netherlands, Spain,
UK)
Starfield 04/01
13
Average Rankings for Health Indicators, YPLL
(Total and Suicide) in Countries Grouped by
Primary Care Orientation

• All Except Suicide Suicide All
  Except External
• Female Male Female Male
  Female Male
• Lowest 9.5 10.8 7.3
  8.3 8.8 10.8
• (Belgium,
• France,
• Germany,
• US)
• Middle 3.8 2.8 7.0 7.3
  3.8 3.5
• (Australia,
• Canada, Japan,
• Sweden)
• Highest 7.6 7.4 6.8 5.8
  8.2 7.0
• (Denmark,
• Finland,
• Netherlands,
• Spain, UK)

Source OECD, 1998.
Starfield 2000
14
Average Rankings for Life Expectancy at Ages 40,
65, and 80, for Countries Grouped by Primary Care
Orientation
Age 80
Age 65
Age 40
Male
Female
Male
Male
Female
Female
7.4
8.0
7.8
6.9
8.0
9.5
Lowest (Belgium, France, Germany, US)
3.6
3.8
4.0
3.5
2.5
Middle (Australia, Canada, Japan, Sweden)
4.3
9.5
8.8
8.8
9.3
9.0
8.6
Highest (Denmark, Finland, Netherlands, Spain,
UK)
Starfield 04/01
15
Average Rankings for World Health
Organization Health Indicators for Countries
Grouped by Primary Care Orientation
DALEs Child Survival Equity Overall Health
Lowest (Belgium, France, Germany, US) 16.3 22.5 36.3
Middle (Australia, Canada, Sweden, Japan) 4.8 16.5 26.0
Highest (Denmark, Finland, Netherlands, Spain, UK) 16.0 15.2 31.6
Source Calculated from WHO, 2000.
DALE Disability adjusted life expectancy (life
lived in good health) Child survival survival
to age 2, with a disparities component Overall
health DALE minus DALE in absence of a health
system Maximum
DALE for health expenditures
minus same in absence of a health system
Starfield 10/02
16
Percentage of Individuals Who Smoke per Capita at
Ages 15 and Older, Early-Mid 1990s

Female Male

• Belgium 21.0 31.0
• France 20.0 38.0
• Germany 21.5 36.8
• United States 24.6 28.6
• Australia 23.8 28.2
• Canada 26.0 26.0
• Japan 13.3 60.4
• Sweden 26.6 25.2
• Denmark 40.1 45.9
• Finland 20.0 33.0
• Netherlands 30.5 42.9
• Spain 21.0 44.0
• United Kingdom 28.0 29.0

All countries 1992, except Canada (1991), Spain
(1993)
17
Ranking of Countries by Income Inequality
Earned Income Disposable Income Country
(90/20 ratio) (Gini) Belgium
5 3 France 10
8 Germany 7 6 United
States 11 13 Australia 12
10 Canada 9 7 Japan
1
11 Sweden 2 2 Denmark
8 4 Finland 6
1 Netherlands 4 5 Spain
3 9 United Kingdom 13
12
Starfield 2000
18
Primary Care Features Consistently Associated
with Good/Excellent Primary Care

• System features
• Regulated resource distribution
• Government-provided health insurance
• No/low cost-sharing for primary care
• Practice features
• Comprehensiveness
• Family orientation

Starfield 10/01
19
Primary Care Score and Health Outcomes
Source Macinko et al., 2002.
Starfield 06/02
20
Primary Care Score and Premature Mortality in 18
OECD Countries
10000
PYLL
Low PC Countries
5000
All Countries
High PC Countries
0
1970
1980
1990
2000
Year
Source Macinko et al., 2002.
Starfield 06/02
21
Within-Country Studies

• Ecological analyses Effect of primary care
  doctor to population ratios (US, UK)
• Case control studies (US)
• Hospitalizations for avoidable conditions or
  complications (US, Spain)
• Survey data on impact of affiliation with a
  primary care doctor (US, Spain)
• Path analyses at state and local levels (US)

Starfield 2000
22
Factors Related to In-hospital Standardized
Mortality, England (NHS Hospitals), 1991-2 to
1994-5
Regression Coefficient of
cases admitted as emergency 0.58 hospital
doctors/100 hospital beds -0.47 GPs/100,000
population -0.67 Standardized admission
ratio -0.15 live discharges to home
1.61 patients with co-morbidity 1.51 NHS
facilities/100,000 population -1.12
Source Jarman et al., 1999.
Starfield 2000
23
Rates of Avoidable Adult Hospitalizationfor 6
Conditions and Family Physiciansper 10,000
Population
Starfield 10/02
Source Parchman Culler, 1994.
24
Rates of Avoidable Pediatric Hospitalization for
Diabetes Mellitus and Pneumonia and Family
Physiciansper 10,000 Population
Source Parchman Culler, 1994.
Starfield 10/02
25
Physician Supply and Odds Ratiosof Diagnosis of
Late-StageColorectal Cancer
Urban Non-urban
Primary care physicians 0.92 0.96
Specialty physicians 1.12 1.02
change in odds of late stage diagnosis with each
10 percentile increase in supply of
physicians significant at plt.01
Source Roetzheim et al., 1999.
Starfield 10/02
26
Adjusted Odds Ratios for Severe, Uncontrolled
Hypertension According to Various Risk Factors

• No. of Patients Adjusted
• with Complete Odds Ratio
• Risk Factor Data (95 CI) P Value
• No primary care 204 4.4 (2.2-8.9) lt0.001
• No medical insurance 204 2.2(1.0-4.6) 0.04
• Noncompliance with antihypertensive 199 2.0
  (1.5-2.7) lt0.001
• regimen
• One or more alcohol-related problems 204 2.2
  (0.8-6.3) 0.14
• Illicit drug use 204 1.3 (0.5-3.6) 0.60
• Categorized on a five-point scale.
• In the past year.

Source Shea et al., 1992.
Starfield 1999
27
Health Care Expenditures and Mortality 5 Year
FollowupUnited States, 1987-92

• Adults (age 25 and older) with a primary care
  physician rather than a specialist as their
  personal physician
• had 33 lower cost of care
• were 19 less likely to die (after controlling
  for age, gender, income, insurance, smoking,
  perceived health (SF-36) and 11 major health
  conditions)

Source Franks Fiscella, 1998.
Starfield 1999
28
Primary Care Reform, 1984-90 to 1994-96
Percent Decline in Mortality - Various Causes,

Barcelona, Spain
45
E 40
M 38
40
M 35
L 35
35
30
Decline
E 23
25
20
15
10
L 6
5
0
Perinatal
Hypertension
E Early Implementation
M Later Implementation
L Late Implementation
Source Villalbi et al., 1999.
Starfield 2000
29
Major Determinants of Outcomes50 US States

• Specialty Physicians More all outcomes worse
• Primary Care Physicians Fewer all outcomes
  worse
• Hospital Beds More higher
  total, heart disease,
• and
  neonatal mortality
• Education No
  relationship
• Income Lower higher heart and
  cancer mortality
• Unemployment Higher higher total
  mortality, lower life span,
• 
  more low birthweight
• Urban Lower mortality (all), longer
  life span
• Pollution Higher total mortality
• Life Style Worse
  higher total and cancer mortality,
• 
  lower life span
• Minority Higher total
  mortality, neonatal mortality, low
  birthweight, lower life span

30
State Level AnalysisPrimary Care and Life
Expectancy
PC physicians/population positively associated
with longer life expectancy.
Source Shi et al., 1999.
Starfield 03/02
31
Path Coefficients for the Effects of Income
Inequality and Primary Care on Health Outcome
50 US States, 1990
Total Mortality
Infant Mortality
.42
.35
-.36
Income Inequality (Robin Hood Index)
-.29
-.33
Primary Care Physicians
.58
-.37
Life Expectancy
Low Birthweight
.41
-.17
plt.05 plt.01.
Source Shi et al., 1999.
Starfield
32
Path Coefficients for the Effects of Income
Inequality and Primary Care on Health Outcome
50 US States, 1990
Life Expectancy
Total Mortality
Neonatal Mortality
.39
.40
-.35
-.38
Income Inequality (GINI COEFFICIENT)
-.18
-.33
Primary Care Physicians
.18
.16
.42
Stroke Mortality
Postneonatal Mortality
-.38
-.33
Life Expectancy
plt.05 plt.01.
Source Shi et al., 1999.
Starfield
33
Does Primary Care Reduce Disparities in Health
across Population Subgroups?
Starfield 10/02
34
Reductions in Inequality in Health by Primary
Care Postneonatal Mortality,50 US States, 1990
Areas with low income inequality High primary
care resources 0.8 decrease in
mortality Low primary care resources 1.9
increase in mortality Areas with high income
inequality High primary care resources
17.1 decrease in mortality Low primary care
resources 6.9 increase in mortality
compared with population mean
Based on data in Shi Starfield, 2000.
Starfield 2000
35
Reductions in Inequality in Health by Primary
Care Stroke Mortality,50 US States, 1990
Areas with low income inequality High primary
care resources 1.3 decrease in
mortality Low primary care resources 2.3
increase in mortality Areas with high income
inequality High primary care resources
2.3 decrease in mortality Low primary care
resources 1.1 increase in mortality
compared with population mean
Based on data in Shi Starfield, 2000.
Starfield 2000
36
Reductions in Inequality in Health by Primary
Care Self-Reported Health,60 US Communities,
1996

• Percent reporting fair or poor health
• Areas with low income inequality
• No effect of primary care resources
• Areas with moderate income inequality
• 16 increase in areas with low primary care
  resources
• Areas with high income inequality
• 33 increase in areas with low primary care
  resources
• compared with median of primary care
  physicians to population ratios

Based on data in Shi Starfield, 2000.
Starfield 2000
37
Primary Care Practice Characteristics
Evidence-Based Summary

• Countries with strong primary care
• have lower overall costs
• generally have healthier populations
• Within countries
• areas with higher primary care physician
  availability (but NOT specialist availability)
  have healthier populations
• more primary care physician availability reduces
  the adverse effects of social inequality

Starfield 1999
38
Primary Care Practice Characteristics
Evidence-Based Summary

• Countries with strong primary care
• have lower overall costs
• generally have healthier populations
• Within countries
• areas with higher primary care physician
  availability (but NOT specialist availability)
  have healthier populations
• more primary care physician availability reduces
  the adverse effects of social inequality

Starfield 1999
39
Conclusions
Both international comparisons and studies within
countries document the beneficial impact of
primary care on effectiveness (health outcomes),
on efficiency (lower costs), and on equity of
health outcomes (reducing disparities across
population subgroups). Health policy should be
directed toward strengthening the primary care
orientation of health systems.