This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike License. Your use of this
material constitutes acceptance of that license and the conditions of use of materials on this site.
Copyright 2011, The Johns Hopkins University and Barbara Starfield. All rights reserved. Use of these materials
permitted only in accordance with license rights granted. Materials provided “AS IS”; no representations or
warranties provided. User assumes all responsibility for use, and all liability related thereto, and must independently
review all materials for accuracy and efficacy. May contain materials owned by others. User is responsible for
obtaining permissions for use from third parties as needed.
Population Health: New
Paradigms and Implications
for Health Statistics
Barbara Starfield, MD, MPH
National Committee on Vital and Health Statistics
(Washington, DC, 2002)
Challenges of 21st Century
Health Systems
• The poor and worsening position of US
population health indicators
• Quality of health services
– Responsiveness to needs
– Comorbidity
– Coordination of care
– Overuse
• Quality of health systems
– Population outcomes
– Disparities (inequities)
Starfield 02/04
H 6176
Is population health
the sum of individual
health?
Starfield 10/02
H 5943
Distinction between focus on
individuals and populations or
subpopulations is at the heart
of distinctions among the
branches of medicine.
Starfield 10/02
H 5945
Clinical Medicine
What disease might this patient have,
and how should it be managed?
Clinical Epidemiology
What is the relative likelihood that this
patient has, or is at risk for, this
disease, and what is the evidence to
support its management?
Starfield 02/04
H 5946
Social Medicine
Why does this patient have this
disease at this particular time, and
how might this affect management?
Community Medicine
Is this disease important? If so, how
important is it, to whom, and what is
the overall benefit of management to
the community?
Starfield 02/04
H 5947
Public Health
What characteristics are most
salient in improving overall
health and the distribution of
health in populations, and what
does evidence suggest should
be priorities for intervention?
Starfield 02/04
PH 5942
Clinical,
Epidemiological, and
Social Views towards
Health
Starfield 02/04
H 6177
COMMUNITY
LEVEL
OCCUPATIONAL &
ENVIRONMENTAL
EXPOSURES
INDIVIDUAL
LEVEL
Social Influences on the
Health of Individuals
PHYSIOLOGICAL
STATE
MATERIAL
RESOURCES
SOCIODEMOGRAPHIC
CHARACTERISTICS
DEVELOPMENTAL
HEALTH
DISADVANTAGE
WEALTH: LEVEL &
DISTRIBUTION
POLITICAL
AND
POLICY
CONTEXT
SOCIAL
RESOURCES
POWER
RELATIONSHIPS
HEALTH*
BEHAVIORS
BEHAVIORAL &
CULTURAL
CHARACTERISTICS
CHRONIC
STRESS
HEALTH SYSTEM
CHARACTERISTICS
HEALTH SERVICES
RECEIVED
Shading represents degree to which characteristics are measured at the ecological
level (lighter color) or at the individual level aggregated to community.
GENETIC &
BIOLOGICAL
CHARACTERISTICS
*“Health” has two aspects: occurrence
(incidence) and intensity (severity).
Starfield 07/07
IH 6784 n
Community Medicine
and Public Health Views
towards Health
Starfield 02/04
H 6178
Societal Influences on Population Health and Equity
COMMUNITY CONTEXT
notes continued on IH 6891 bn
ENVIRONMENTAL
CHARACTERISTICS
POLICY CONTEXT
OCCUPATIONAL &
ENVIRONMENTAL
POLICY
POLITICAL
CONTEXT
SOCIAL
POLICY
EQUITY IN
HEALTH*
WEALTH: LEVEL &
DISTRIBUTION
POWER/STATUS
RELATIONSHIPS
HISTORICAL
HEALTH
DISADVANTAGE
RATES OF DISCOMFORT
AND DISEASE
ECONOMIC
POLICY
BEHAVIORAL &
CULTURAL
CHARACTERISTICS
RATES OF DISABILITY
AND DEATH
HEALTH
POLICY
HEALTH SYSTEM
CHARACTERISTICS
Dashed lines indicate the existence of pathways through individual-level
characteristics that most proximally influence health.
Shading represents degree to which characteristics are measured at the ecological
level (lighter color) or at the individual level aggregated to community.
DEMOGRAPHIC
STRUCTURE
*“Health” has two aspects: occurrence
(incidence) and intensity (severity).
Starfield 01/08
IH 6891 an
Societal Influences on Population Health and Equity (continued)
COMMUNITY CONTEXT
notes continued from IH 6891 an
ENVIRONMENTAL
CHARACTERISTICS
POLICY CONTEXT
OCCUPATIONAL &
ENVIRONMENTAL
POLICY
POLITICAL
CONTEXT
SOCIAL
POLICY
EQUITY IN
HEALTH*
WEALTH: LEVEL &
DISTRIBUTION
POWER/STATUS
RELATIONSHIPS
HISTORICAL
HEALTH
DISADVANTAGE
RATES OF DISCOMFORT
AND DISEASE
ECONOMIC
POLICY
BEHAVIORAL &
CULTURAL
CHARACTERISTICS
RATES OF DISABILITY
AND DEATH
HEALTH
POLICY
HEALTH SYSTEM
CHARACTERISTICS
Dashed lines indicate the existence of pathways through individual-level
characteristics that most proximally influence health.
Shading represents degree to which characteristics are measured at the ecological
level (lighter color) or at the individual level aggregated to community.
DEMOGRAPHIC
STRUCTURE
*“Health” has two aspects: occurrence
(incidence) and intensity (severity).
Starfield 01/08
IH 6891 bn
Implications for Data/
Information Systems:
I. Linkages
• Linking individual/aggregated individual
data with contextual/ecological data
i.e., clinical approaches with systems
approaches
Starfield 02/04
INSI 6170
Implications for Data/
Information Systems:
II. Areas
• Characteristics of areas in which people
live and work
• Social and political (power) characteristics
of the people in an area
• Characteristics of health system
Starfield 02/04
INSI 6171
Implications for Data/
Information Systems:
III. Health Services
• Problems (ICPC)
• Diagnoses (comorbid diagnoses)
• Management (disease or morbidityoriented?)
• Reassessment (disease or problemoriented?)
Starfield 02/04
INSI 6172
Implications for Data/
Information Systems:
IV. Disease, Morbidity, or
Health Oriented?
• Comorbidity
• Concept of health
Starfield 02/04
INSI 6173
Comorbidity
Diseases, risk factors, and
influences are not independent
of one another. Data systems
must allow for coordination
among different providers and
types of providers.
Starfield 10/02
CM 5953
Ratios of Observed and Expected (Co-)Occurrences
of Diseases, Overall and for Different Ages
Starfield 09/07
CM 5657 an
Decision Tree
for ACGs
Missing Age
MAC-26
Age < 1
ACG 9900
Entire Population
Age >= 1
To
MAC 26
tree
Split into MACs,
Based on CADGs
To
MAC 24
tree
To
MAC 12
tree
ADG
05 ?
Age
1
Yes
ACG 0100
2-5
ADG
25 ?
ACG 0200
Yes
No
ACG 0600
No
ADG
24?
Yes
ACG 0300
1
Yes
ACG 1900
ACG 1500
2-5
ACG 2000
ADG
05 ?
No
Key
MAC
ADG
CADG
ACG
Major Ambulatory Category
Ambulatory Diagnostic Group
Collapsed ADG
Ambulatory Care Group
No
ADG
24?
Yes
6+
1 or 2
input
files?
Age
1
1
ACG 2900
ACG 2500
ACG 1300
ACG 0500
6+
ADG
25?
Age
Yes
ACG 2200
Source: JHU ACG Case Mix Adjustment System, V. 4.0, 1997.
No
ACG 2100
ACG 5100
Claims
info?
2-5
ACG 3000
Yes
ACG 2700
6+
ACG 2600
12 + ADG
05 ?
ACG 3100
No
ACG 1400
2
No
Yes
ACG 3300
No
ACG 3200
ACG 5110
ACG 5200
Starfield 09/07
ACG 5090 n
What Is Health?
Health is the extent to which an individual or
group is able, on the one hand, to realize
aspirations and satisfy needs and, on the
other hand, to cope with the interpersonal,
social, biological, and physical environments.
It is a resource for everyday life, not the
objective of living; it is a positive concept
embracing social and personal resources as
well as physical and psychological
capacities.
Source: Ottawa Charter for Health Promotion, 1986.
Starfield 10/02
H 5944
How Is Population Health Measured?
Infant mortality rates vs. neonatal and postneonatal
Mortality rates vs. age-adjusted mortality rates
Cause-specific mortality
Life expectancy
Years of potential life lost
(GAO choice among 17)
Disease occurrence/severity
All diseases
Target diseases
(Popular in US and worldwide)
Self-reported health
DALYs
Which of these represent health?
Starfield 11/02
H 5962
Alternatives for Characterizing
Population Health
1. Diagnosed morbidity case mix by
age
2. DALYs
3. Profiles of health derived from
combinations of separate domains
Starfield 11/02
H 5963
Criteria for Defining Profile Types
Profile
Type
Criteria
A
Excellent health
Excellent health on 3 or 4 domains, with no
domains of poor health
B
Good health
At least average health on all domains, with
excellent health on no more than 2 domains
C
Dissatisfied
Poor health only on satisfaction
D
Discomfort
Poor health only on discomfort
E
Low resilience
Poor health only on resilience
F
High risks
Poor health only on risks
G
Dissatisfied/high discomfort
Poor health on satisfaction & discomfort
H
Dissatisfied/low resilience
Poor health on satisfaction & resilience
I
Dissatisfied/high risks
Poor health on satisfaction & risks
J
Discomfort/low resilience
Poor health on discomfort & resilience
K
Discomfort/high risks
Poor health on discomfort & risks
L
Low resilience/high risks
Poor health on resilience & risks
M
Worst health
Poor health on 3 or 4 domains
Starfield 11/02
CHIP 5171
Dealing with Disparities
Stratified analysis rather
than statistical adjustment
Starfield 02/04
H 5643
Social and Political
Context
Important future direction to
understanding the variety of
influences on health
Starfield 02/04
H 5949
Directions for Population
Health Data
Information systems moving towards characterizing
•
•
•
•
•
Health
Contexts
Comorbidity as well as disease
Disparities (systematic differences)
Bridging clinical medicine and public health to use
knowledge about health, risks, and resiliencies
Starfield 02/04
INSI 5951
Data Systems
for Bridging Public Health
and Clinical Medicine
Boundaries between public health
and clinical medicine are fluid. Data
systems are needed to coordinate
the activities of both sectors.
Starfield 10/02
INSI 5952
Population health is not the sum of any
individual measure of health because of
NON-RANDOM DISTRIBUTION OF HEALTH
AND INFLUENCES ON HEALTH
– Age and gender distributions
– Political/economic/social/environmental
contexts
INDIVIDUAL MEASURES OF HEALTH DO
NOT REPRESENT HEALTH.
Starfield 10/02
H 5950
Summary: Health
Population health is NOT the sum
of individual health. Averages do
not represent population health or
provide information about the
context in which systematic
differences in health occur and
how they can be remedied.
Starfield 02/04
H 6174
Summary: Why Population Health
Is NOT the Sum of Individual
Health, in Any Useful Sense
• Comorbidity: individual measures of
health do not represent health
• Non-random distribution of health,
e.g., age, geography, societal
differences cause systematic
differences in population subgroups
Starfield 02/04
H 6175
