A Method for Assessing

1
A Method for Assessing Broad View of Health
Among Healthy Cities and Communities Participants

• Robert E. Aronson, Dr.P.H.,
• UNC-Greensboro
• Michelle C. Kegler, Dr.P.H.,
• Emory University
• Barbara L. Norton, M.P.H.
• Univ. of Oklahoma
• Heather Paulsen, M.A.
• Center for Civic Partnerships

2
Healthy Cities Communities Movement
The Healthy Cities movement was born in 1986 when
the World Health Organization (WHO) initiated a
pilot initiative in 11 European cities. By 1998
there were over 6500 self-declared healthy
cities and close to 40 cities directly sponsored
by the WHO. Within the U.S. statewide
initiatives have been launched in the states of
Indiana, Colorado and California, along with
countless individual cities. Healthy Cities
projects work with communities to foster
health-promoting environments through
inter-sector cooperation, extensive community
participation, and the design and implementation
of action-based plans.
3
Healthy Communities Principles

• A broad definition of health
• A broad definition of community
• Shared vision from community values
• Address quality of life for everyone
• Diverse citizen participation and widespread
  community ownership
• Focus on systems change
• Build capacity using local assets and resources
• Benchmark and measure progress and outcomes

Norris T, Pittman M. The Healthy Communities
Movement and the Coalition for Healthier Cities
and Communities, Public Health Reports, Vol. 115,
2000, p. 122.
4
What is a broad view of health?

• A definition of health that is holistic, and
  applies to individuals as well as contexts
• A perspective of the determinants of health
  that recognizes and emphasizes the contribution
  of lifestyle and environment.

5
Broad Definition of Health

• Health is more than the absence of disease.
• Health can be described as an optimum state of
  well being, including mental, physical,
  emotional and spiritual.
• Health includes a sense of belonging to community
  and experiencing control over ones circumstances
  and fate.
• Health applies not only to individuals, but also
  to contexts, such as organizations, cities and
  communities.

6
What is a Healthy City or Community?

• A healthy community is one that is continually
  creating and improving those physical and social
  environments and expanding those community
  resources which enable people to mutually support
  each other in performing all the functions of
  life and in developing to their maximum
  potential.

Hancock T, Duhl L. 1986. Healthy Cities
Promoting Health in the Urban Context.
Copenhagen WHO Europe, p. 24.
7
What Makes Populations Healthy?
Source Health and Welfare Canada
8
Broad Perspective on Determinants of Health

• Biologic and genetic factors
• Behavior and lifestyle
• Social environment and relationships
• Physical environment and living conditions
• Health care system

9
Assessing Broad View of Health

• members of the steering committee must grasp
  and articulate the root and multi-causal nature
  of community health problems to ensure that
  issues beyond behavioral risk factors are
  addressed.
• Kegler MC, Twiss JM, Look V. Assessing community
  change at multiple levels The genesis of an
  evaluation framework for the California Healthy
  Cities Project. Health Education and Behavior,
  Vol. 27(6) p. 768, 2000.

10
Assessing Broad View of Health in California
Healthy Cities and Communities

• Seen as an aspect of the evaluation of change
  among members of the local steering
  committee/local coalition.
• Administered as a part of the steering
  committee/coalition survey during years 1 and 3.
• Question structured as a paired comparison
  activity to identify individual and composite
  rank ordering of determinants of health.

11
Ways to collect rank-order data

• Ask research participants to number items
  sequentially, in order of importance.
• Ask research participants to compare two items at
  a time and select the most important item until
  all items are compared with every other item
  (paired comparison).

12
Our Survey Question
This next question asks you to think about the
relative importance of different factors that
affect the health of individuals.
8. The following is a list of pairs of items, all
considered by some to be important factors that
influence the health of individual persons.
(Although some of the pairs may sound
repetitive, no pair of items appears more than
once.) For each pair of items, please circle
the item which you think is a more important
influence on health. Please circle only one
from each pair.
13
Our Survey Question (Continued)
Lifestyle decisions vs. Genetics/heredity Qualit
y health care vs. Supportive
relationships Living conditions vs. Genetics/he
redity Quality health care vs. Genetics/heredity
Living conditions vs. Lifestyle
decisions Genetics/heredity vs. Supportive
relationships Living conditions vs. Supportive
relationships Supportive relationships vs. Lifes
tyle decisions Quality health care vs. Lifestyle
decisions Quality health care vs. Living
conditions
14
Univariates on Determinants of Health (n309)
Which is a more important influence on
health.... Lifestyle decisions vs.
Genetics/heredity 85.1 14.9 Quality health
care vs. Supportive relationships 56.0 44.0
Living conditions vs. Genetics/heredity 82.5
17.5 Quality health care vs. Genetics/heredity
75.1 24.9 Living conditions vs. Lifestyle
decisions 25.9 74.1
15
Univariates on Determinants of Health Continued
Which is a more important influence on
health.... Genetics/heredity vs. Supportive
relationships 33.7 66.3 Living
conditions vs. Supportive relationships 46.3
53.7 Supportive relationships vs. Lifestyle
decisions 32.0 68.0 Quality health
care vs. Lifestyle decisions 38.5 61.5 Qual
ity health care vs. Living conditions 53.7 46
.3
16
Pairs with the most disagreement
Quality health care vs. Supportive
relationships 56.0 44.0 Living
conditions vs. Supportive relationships 46.3
53.7 Quality health care vs. Living
conditions 53.7 46.3
17
What is the resulting rank order?

• Can the aggregate responses to all ten of these
  pairs be converted into a rank order of the five
  items?

Using the software ANTHROPAC 4.0, a composite
rank-order can be determined.
but
we must further analyze the data to determine
the degree of agreement with the rank order.
18
Analysis of Paired Comparison Data Using Anthropac

• Using the software Anthropac, the data is
  converted into the following files
• Individual dominance matrices
• Matrix of item by item dominance scores
  (for each pair which is more important) for
  every respondent (309 matrices)
• Aggregate dominance matrix
• Single matrix of item by item dominance scores
  based upon an average of all individual
  dominance matrices
• Unrandomized respondent by item matrix
• If the order of the pairs were randomized, an
  unrandomized matrix containing respondents in
  rows and pairs in columns is generated.
• Composite Rank Order
• This file is an item-vector that averages the
  respondent-by-item matrix across all
  respondents, and produces a z-score for each item
  (Torgerson, WS, 1958. Theory and Methods of
  Scaling. New York John Wiley.).

19
Composite Rank Order of Determinants of Health
Rank Item z-score _____________________________
_____________ 1 Lifestyle decisions
0.50 2 Quality health care 0.13 3 Living
conditions 0.02 4 Supportive
relationships -0.03 5 Genetics/heredity -0.63
20
Interpreting this Composite Ranking

• Does the rank ordering indicate a broad view of
  health?
• What kind of variability is there in this rank
  ordering?
• Does the rank order differ from community to
  community?
• What kinds of changes might we expect to see
  after three years of involvement in the Healthy
  Cities and Communities project?

21
Next Step in the Analysis Consensus

• Use ANTHROPACs Consensus Analysis function to
  determine if a single or multiple rank order
  solutions are possible.
• Consensus Analysis is a factor analytic
  approach that examines patterns of agreement
  between respondents.

22
EIGEN VALUES FROM CONSENSUS ANALYSIS
Factor Value Percent Cum Ratio _______________
___________________________________ 1 89.939 48.2
48.2 1.594 2 56.435 30.2 78.5 1.404
3 40.195 21.5 100.0 Average correlation
among respondents 0.169 Respondent reliability
0.984 Range of individual factor loadings
(-0.891, 0.891)
23
Interpreting Consensus Outputs

• There is little agreement between respondents on
  the rank-order of the determinants of health.
• We cannot consider the composite rank-order to be
  a meaningful indicator of how people involved in
  the CA Healthy Cities and Communities initiative
  view the determinants of health.
• There are likely to be two or more common ways to
  rank-order these items within our sample.
• Even analyzing results community-by-community,
  shows little consensus on rank-order in all but
  six communities.

24
Whats Next

• We can look to see if there is greater consensus
  among respondents from the year 3 survey.
• If there is consensus in the year 3 survey
  results, we can assess the composite rank order
  in terms of a broad perspective on the
  determinants of health.
• If there is no consensus, we can still examine
  the univariates for each pair of items from the
  year 3 survey to see if there are any changes.

25
References for these Methods

• Borgatti. 1992. ANTHROPAC 4.0. Columbia
  Analytic Technologies.
• Bernard HR. 1994. Research Methods in
  Anthropology. Thousand Oaks Sage Publications.
• De Munck VC, Sobo EJ, eds. 1998. Using Methods
  in the Field. Walnut Creek Altamira Press.
• Weller SC, Romney AK. 1988. Systematic Data
  Collection. Qualitative Research Methods, Vol.
  10. Newbury Park, CA Sage Publications.