Upload
kathleen-wilson
View
214
Download
0
Embed Size (px)
Citation preview
Social Science & Medicine 55 (2002) 2017–2031
Exploring the determinants of health for First Nationspeoples in Canada: can existing frameworks
accommodate traditional activities?
Kathleen Wilsona,*, Mark W. Rosenbergb
aSchool of Geography and Geology, McMaster University, 1280 Main Street West, Hamilton, Ontario, Canada L8S 4K1bDepartment of Geography, Queen’s University, Kingston, Ontario, Canada K7L 3N6
Abstract
While much research has examined First Nations peoples’ health in Canada, few studies have explored the role of
traditional activities in enhancing health. Using data from the 1991 Aboriginal Peoples Survey (APS), this paper
incorporates a set of measures of traditional activities within a determinants of health framework for understanding
First Nations peoples’ health. Results from the analyses undertaken show that many of the determinants of health
identified in analyses of the Canadian population in general hold for First Nations peoples. While only a few statistically
significant relationships between health status and traditional activities were identified, taking into account the
limitations of the APS and other conceptual issues, we argue that there is the potential to move from the analysis of
traditional activities to a more nuanced analysis of cultural attachment.r 2002 Elsevier Science Ltd. All rights reserved.
Keywords: Aboriginal peoples; First Nations peoples; Traditional activities; Determinants of health; Canada
Aboriginal people from almost every culture believe
that health is a matter of balance and harmony
within the self and with others, sustained and ordered
by spiritual law and the bounty of Mother Earth.
Royal Commission on Aboriginal Peoples Volume 3
(1996, p. 184).
Introduction
In exploring First Nations peoples’ health in Canada
there are two distinct bodies of literature that can be
consulted.1 The first is primarily epidemiologic and
utilizes quantitative methods to examine health and
illness/disease in the context of the determinants of
health (Harris, Caulfield, Sugamori, Whalen, & Hen-
ning, et al., 1997; Hegele et al., 1997; Moffatt, 1995; for
a good review see Young, 1994). This is a significant
body of research for identifying and trying to explain
health inequalities. While important, it is flawed by the
fact that few studies incorporate First Nations peoples’
culture into analyses of health. Some researchers have
taken on the task of trying to operationalize culture with
varying success rates (Bagley, 1991; Foggin & Aurillon,
1989; Neuwalt, Kearns, Hunter, & Batten, 1992; New-
bold, 1997; Thouez, Rannou, & Foggin, 1989; Waldram,
1990; Young, 1998). The second is a body of cultural
literature, characterized by qualitative research meth-
ods, which links culture and health (Adelson, 1998;
Borr!e, 1991, 1994; Garro, 1988, 1995; Hagey, 1989;
Jilek, 1982; Waldram, 1993, 1997).2 As is often the case
with research based on qualitative methods, there is a
lack of generality to these studies.
*Corresponding author.1When using the term Aboriginal we are referring to the
descendants of the original inhabitants of Canada, as defined by
the Constitution Act 1982; Indians, Inuit and M!etis. Many
‘Indians’ prefer the terms First Nations when referring to
themselves as a collective group. Therefore we use the term
First Nations when referring to this segment of the Aboriginal
population.
2This is not an exhaustive list of the research being conducted
on the health of First Nations peoples.
0277-9536/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved.
PII: S 0 2 7 7 - 9 5 3 6 ( 0 1 ) 0 0 3 4 2 - 2
In 1991 Statistics Canada carried out the Aboriginal
Peoples Survey (APS), a national survey of individuals
living on reserves, in settlements and off reserve
areas who self-reported their Aboriginal identity
(Canada, 1993a). Included in this survey are questions
related to language and tradition as well as health,
lifestyle and social issues. Utilizing data from the APS
this paper explores whether measures of ‘traditional’
Aboriginal activities contribute to our understanding of
Aboriginal peoples’ health within a determinants of
health framework. Going beyond conventional mea-
sures that a determinants of health approach embraces,
a set of variables are tested which look at the importance
of traditional activities for the health of Aboriginal
peoples.
The first section of this paper discusses the general
health status of First Nations peoples in Canada,
documenting their lower levels of health, as compared
to the non-Aboriginal population. The second section
summarizes the different conceptualizations used to
explore the determinants of health. In the third section,
the data used in this study are outlined as well as the
methods employed in the analysis. The fourth section
describes the results of the research. In the final section,
we reflect on how taking into account the limitations of
the APS and other conceptual issues might move research
on First Nations peoples’ health beyond the dichotomy
which now exists between epidemiological and cultural
studies to a more nuanced analysis of cultural attachment.
The health of First Nations peoples in Canada
It is a well known fact that First Nations peoples in
Canada suffer from a poorer quality of life, as measured
by mortality and morbidity, as compared to their non-
Aboriginal counterparts (Enarson & Grzybowski, 1986;
Hammond, Rutherford, & Malazdrewicz, 1988; Young,
1991).3 The average life expectancy of Registered
Indians is approximately six years less than the overall
Canadian population (Canada, 1998).4 In 1990, the life
expectancy of male and female Registered Indians was
66.9 and 74.0 years respectively. In contrast, the life
expectancy for the total male and female Canadian
population was 73.9 and 80.5 years (Canada, 1996a).
Mao, Moloughney, Semenciw and Morrison (1992) note
that Registered Indians living on reserves suffer higher
mortality rates from coronary heart disease, suicide and
cirrhosis as compared to the rest of the Canadian
population. The infant mortality rate among Registered
Indians is approximately two times higher than for
Canadians generally (Canada, 1996a).
Research on morbidity shows that a much higher
proportion of First Nations peoples suffer from certain
illnesses than do non-Aboriginal Canadians. Enarson
and Grzybowski (1986) examined differences in tuber-
culosis rates across the country for three groups; Inuit,
registered Indians and others of mainly European
origin. Their findings showed that tuberculosis rates
were 16 times higher among Registered Indians and 24
times higher among Inuit as compared to the third
group. Mao et al. (1992) showed that the stroke rate for
Registered female Indians is 2 times higher than for non-
Aboriginals. While rates of infectious diseases among
Aboriginal peoples are declining, rates of chronic
illnesses are on the rise. Research by Young, McIntyre,
Dooley, and Rodriguez (1985) and Evers, McCracken,
Antone, and Deagle (1987) documents the very high
rates of type II Diabetes in the Aboriginal population as
compared to the non-Aboriginal population. According
to 1991 figures the disability rate for Aboriginal peoples
was 31% as compared to 13% for the non-Aboriginal
population (Ng, 1996).
The preceding examples represent only a small subset
of the differences in health status which distinguish First
Nations peoples from the remainder of the Canadian
population. Since First Nations peoples suffer from
lower levels of health as compared to their non-
Aboriginal counterparts, this begs the question ‘What
determines First Nations peoples’ health?’ Given the
importance of this question there are two goals we seek
to address in this paper. Firstly, to understand the
determinants of health for First Nations peoples within
the context of a social determinants of health frame-
work. Secondly, given the uniqueness of First Nations
peoples’ health status as well as their cultural system of
beliefs, this paper further enhances our understanding of
the ‘conventional’ social determinants of health by
3As Waldram et al. (1995) state, there is no convenient single
source of mortality data for Aboriginal people living in Canada.
The Medical Services Branch (MSB) collects vital statistics
based on its administrative regions. However, it only serves
75% of the on-reserve registered Indian population. As a result,
individuals living off-reserve and the non-status population are
unaccounted for by MSB. Mortality figures for Registered
Indians are also obtained from the Indian Register. However,
late reporting of births and deaths adversely affects the quality
of this information (Waldram et al., 1995). The Canadian
Census is another source from which health information can be
obtained. However, the refusal of Indian bands to participate in
the Census has resulted in incomplete enumeration. Further,
since ethnic status within the Census is based on self-
identification, the failure of individuals to identify with any of
the three Aboriginal categories results in inaccurate population
counts.
4While the term ‘Indian’ is generally inappropriate, we use it
for reasons of clarity only in two circumstances. First, it is used
when discussing the results of other studies that have employed
this term. Second, since the term ‘Indian’ is utilized within the
APS, we use it when referring to our analysis of the APS.
K. Wilson, M.W. Rosenberg / Social Science & Medicine 55 (2002) 2017–20312018
exploring the role of traditional activities in shaping the
health of First Nations peoples.
Conceptualizing the determinants of health
There is much discussion in the Canadian literature
surrounding the determinants of health. Many of these
documents are broad in their discussions of both health
and the determinants of health.
The 1974 Lalonde report entitled New Perspective on
the Health of Canadians was one of the earliest health
documents to identify factors, other than the health care
system (human biology, environment, life style, and
health care organization), which contribute to health
(see Lalonde, 1974, p. 31). While it is an important
document, the Lalonde report has been criticized for its
failure to acknowledge the effect of social environments
on health (Ontario, 1991). Despite this criticism, the
Lalonde report was the impetus for a sequence of
national reports that were aimed at ‘‘Achieving Health
for All’’, a phrase coined by Jake Epp and the
Ottawa Charter for Health Promotion (Epp, 1986; see
also ACPH, 1996; ACPH, 1994; Ontario, 1991; Sas-
katchewan, 1997). Epp’s (1986) paper identified
specific challenges faced in achieving health for all
Canadians: (i) reducing inequalities, (ii) increasing
the prevention effort, and (iii) enhancing people’s ability
to cope. The report attempted to strike a balance
between health promotion, disease prevention and
health care.
In 1991 the Healthy Public Policy Committee of the
Premier’s Council on Health Strategy produced a
document entitled Nurturing Health: A Framework on
the Determinants of Health. The purpose of this
document was to develop public policy initiatives aimed
at improving the health of Ontario’s citizens (Ontario,
1991). Similar to other policy documents, Nurturing
Health provided a broad look at the determinants
of health, covering a variety of factors ranging from
the physical and social environment to biological
endowment and health care. The Federal, Provincial,
and Territorial Advisory Committee on Population
Health (ACPH) produced two documents, Strategies
for Population Health: Investing in the Health of
Canadians (1994) and Report on the Health of Canadians
(1996), that identified similar key factors that influence
health. The main factors these documents identified were
living and working conditions, physical environment,
personal health practices, health services, and biology
and genetic endowment (see ACPH, 1994; ACPH,
1996).
In addition to these documents, numerous non-
governmental initiatives have outlined the ways in which
researchers can effectively study the determinants of
health. Evans and Stoddart (1990, p.1349) propose a
framework for the determinants of health that they
argue is complex and ‘‘sufficiently comprehensive and
flexible to represent a wider range of the relationships
among the determinants of health’’. They provide a
comprehensive model of the determinants of health,
which includes the social and physical environment,
genetic endowment, and health care systems. Their
discussion of the social and individual factors that
influence health tends to focus on the importance of
social conditioning as well as the dynamics of social
status and class and their inter-relationships with well
being.
Hertzman, Frank and Evans (1994) developed a
framework for discussing heterogeneities in health status
and the determinants of population health. Their
framework allows for the partitioning of the population
to define subgroups that differ in their average health
status (see p. 67). Their conceptual model is presented as
a cube with the three axes of the face of the cube
representing the key dimensions for studying hetero-
geneities in health status: (i) stages of the life cycle, (ii)
subpopulation partitions, and (iii) sources of hetero-
geneity. In terms of subpopulation partitions, Hertzman
et al. (p.75) argue that ‘‘the interesting partitions are
those which consistently demonstrate clear heterogene-
ity of health status across their subgroups in many
diverse settings’’. According to their conceptual model,
populations are partitioned by socioeconomic status,
ethnicity/migration, geography, gender, and special
populations.
The frameworks outlined above differ in their
approach to health and well being. Despite their
differing approaches, they are almost exhaustive in their
discussions and examinations of the determinants of
health. These approaches provide the conceptual frame-
work from which we explore the determinants of health
for First Nations peoples. However, we take our
analysis beyond the conventional determinants of health
outlined above by examining the role of traditional
activities in shaping health. More specifically, using
logistic regression analysis this paper examines if
measures of traditional activities that were included in
the APS further our understanding of First Nations
peoples’ health within the context of the social
determinants of health.
Data and methods
The APS is a national survey of individuals who self-
reported their Aboriginal identity, living on reserves, in
settlements and off reserves (Canada, 1993a). The
observations in the APS are drawn from the Aboriginal
population of Canada based on responses to two
questions in the 1991 Canadian Census. To be selected
for the APS, individuals had to indicate whether: (i) they
K. Wilson, M.W. Rosenberg / Social Science & Medicine 55 (2002) 2017–2031 2019
had Aboriginal origins (North American Indian, Inuit,
M!etis); and/or (ii) they were registered under the Indian
Act of Canada. Those individuals selected for the APS
were then asked whether: (i) they identified with an
Aboriginal group; and/or (ii) if they were registered
under the Indian Act. If an individual answered in the
affirmative to either of these questions, they were asked
to complete the remainder of the survey.5
The total sample size of the APS is 36,635 persons
who reported having Aboriginal origins and/or being
registered under the Indian Act. Of that total sample
size, 25,122 individuals report identifying with their
Aboriginal origins and 11,513 do not. The sample size
may under-represent some groups because of their
refusal to participate in the 1991 census or the APS.
The results of the APS are unavailable for those
communities that were incompletely enumerated in the
1991 Census (Canada, 1995a). In total, 78 reserves and
settlements were incompletely enumerated in the census
and 181 were incompletely represented in the APS
(Canada, 1995a).
The APS groups Aboriginal identity into three
categories; North American Indians, M!etis and Inuit.
This analysis explores the determinants of health for
only those individuals identifying themselves as regis-
tered North American Indians (a sub-sample size of
16,249). North American Indians who are not registered
under the Indian Act or who did not indicate their
registration status within the APS are excluded in the
analysis. Only those who self-reported themselves as
registered North American Indians were chosen for
analysis because they, unlike their non-status counter-
parts and the M!etis, receive both standard medical
services as well as non-insured medical services such as
eye care, dental care and drugs through the First
Nations and Inuit Health Branch (FNIHB) of Health
and Welfare Canada.
The APS is divided into eight thematic sections and
also contains standard sociodemographic information
for each individual (see Table 1). The breadth of
questions included in the APS provides the opportunity
to examine the relative importance of traditional
activities in relation to more conventional measures of
the determinants of health (e.g., age).
Based on the variables included in the APS, the
dependent variable in the logistic regression models is
self-assessed health status, with individuals reporting
their health as excellent, very good, good, fair or poor.
Due to the low response rate in some of the categories,
individuals reporting their health as excellent, very good
or good are combined into a ‘healthy’ category while
individuals reporting their health as fair or poor are
combined into an ‘unhealthy’ category. Approximately
86% of the population fall into the healthy category
while 13% fall into the unhealthy category. This
distribution is similar to what is found in other national
and provincial surveys of the non First Nations
population (Canada 1994–95, 1996–97; Ontario, 1990;
Qu!ebec, 1987, 1992–93).
Building upon the frameworks presented earlier in the
paper, the determinants of health included in this study
can be divided into five categories. The first four
categories represent conventional determinants of
health: demographic status, socioeconomic status, utili-
zation of health care services, and place of residence.
The fifth category includes variables that are designed to
measure attachment to traditional activities. Gender,
age and marital status are used to measure demographic
status while total income, highest level of schooling and
employment status are used to measure socioeconomic
status. A variable measuring whether or not an
individual had visited a physician in the past two weeks
measures utilization of health care services. Place of
residence is the contextual geographic variable and is
divided into four levels: reserve/settlement; Census
Metropolitan Area (CMA); other urban; and other
rural.
Three variables were chosen as measures of attach-
ment to traditional activities: participation in tradi-
tional activities; spent time on the land; acquired food
through hunting, trapping or fishing. Participation in
traditional activities is defined in the APS as ‘‘traditional
ways of doing things such as hunting, fishing, trapping,
storytelling, traditional dancing, fiddle playing, jigging,
arts and crafts, pow-wows, etc’’(APS, 1991). In terms
Table 1
Thematic areas included in the 1991 Aboriginal Peoples Survey
Section Theme
A Identity
B Language and tradition
C Disability
D Health, lifestyle and social issues
E Mobility
F Schooling
G Work and related activities
H Expenditures and sources of income
5The definition of the Aboriginal population differs between
the 1991 Canada Census and the 1991 Aboriginal Peoples
Survey. The Census defines the population based on Aboriginal
ancestry (i.e., persons who report at least one Aboriginal
origin). The APS selects individuals who not only indicate
Aboriginal ancestry but also consider themselves to be
Aboriginal. For example, a respondent may report in the
1991 Census that they have Inuit origin from an ancestor but
report in the APS that they do not identify with an Aboriginal
group. As such, the population counts between the 1991 Census
and the APS are different.
K. Wilson, M.W. Rosenberg / Social Science & Medicine 55 (2002) 2017–20312020
of spending time on the land, the APS asked respon-
dents if they spent part of the last 12 months living
on the land to hunt, fish, trap or teach traditional
ways to their children. Finally, the survey asked
respondents what amount of meat, fish or poultry
consumed in the house is obtained through hunting,
fishing, or trapping. The survey coded responses into
four categories: none, some, half, or most. For purposes
of this research, the responses are recoded into a binary
variable (i.e., no meat, fish or poultry acquired through
hunting and some, half or most acquired through
hunting).
Given the dichotomous nature of the dependent
variable (0,1), logistic regression is used to estimate all
the models. Coefficients are estimated using the max-
imum likelihood method (MLM) of estimation (Aldrich
& Nelson, 1984). The independent variables are recoded
into categorical indicator variables. One value of each
variable is chosen to be the reference category. To
produce a more reliable statistical model, in most cases,
the value chosen to be the reference category is the one
with the highest frequency (see Table 2). For ease of
interpretation, the results are discussed in terms of the
odds ratios. The odds ratio is a measure that approx-
imates how much more likely (or unlikely) it is for the
outcome, in this case being unhealthy, to be present
among those with a given attribute relative to the
reference category and controlling for all other attri-
butes.
Results
In the first part of this section, we use a series of cross-
tabulations to set the context for the logistic regression
analysis that follows. In doing so, we examine the health
of First Nations peoples by age and sex as well as
attachment to traditional activities by age, sex and place
of residence.
Fig. 1 shows that for each of the three youngest age
cohorts, a higher proportion of men report being
healthy (i.e., excellent, very good, good) as compared
to women. For adults aged 40–64 years, 74% of women
are healthy as compared to approximately 81% of men.
In addition, this figure shows that as age increases the
proportion of men and women reporting excellent, very
good or good health decreases.
Figs. 2–5 illustrate how attachments to traditional
activities vary by sex and place of residence. In Fig. 2
we see that a much higher proportion of men have spent
some amount of time living on the land in the past year
relative to women in every age cohort. Similar to the
results of Fig. 2, Figs. 3 and 4 illustrate that higher
proportions of men than women report acquiring food
through hunting, fishing or trapping as well as participat-
ing in traditional activities. The findings of Figs. 2–4 may
reflect a gender division of labour, in that spending time
on the land, participating in traditional activities and
acquiring food through hunting, fishing and trapping are
activities generally associated with male band members.
Table 2
Categories assigned to each variable (reference categories indicated by bold)
Variable groupings Categories
Health status
Self assessed health status Healthy (excellent/very good/good)
Unhealthy (fair/poor)
Demographic status
Age 15–24, 25–39, 40–64, 65+
Sex Female, Male
Marital status Single, Divorced/separated/widowed, Married
Socioeconomic status
Total income 0–9,999, 10,000–19,999, 20,000–39,999, 40,000+
Education Less than high school, High school, Post-secondary
Employment status Employed, Unemployed not in the labour force
Health care
Visited physician in past year Yes, No
Geographic
Place of residence Reserve, CMA, Other urban, Other rural
Attachment to traditional activities
Spent time living on land in No, Yes
Past year
Hunt for food Yes, No
Participate in traditional Yes, No
Activities
K. Wilson, M.W. Rosenberg / Social Science & Medicine 55 (2002) 2017–2031 2021
Fig. 5 demonstrates that attachment to traditional
activities varies by place of residence. In particular, a
much higher percentage of individuals living on reserves
report spending time on the land, participating in
traditional activities and acquiring food through
hunting, fishing, trapping, relative to individuals living
in CMAs and other urban areas. For example, 85% of
individuals living on reserves report acquiring food
through hunting, fishing and trapping as compared to
only 47% in CMAs. In addition, approximately 30% of
respondents living on reserves stated they spent time
living on the land relative to only 14% of their
counterparts living in CMAs. This may reflect differ-
ential levels of access to traditional activities between
First Nations peoples living on reserves and those living
in urban areas.
Having explored some key relationships between
traditional activities, sex and place of residence, we
now turn to examine the complexity of these relation-
ships in the context of the social determinants of health
using logistic regression analysis.
The first model explores the determinants of health at
the national level (see Table 3).6 All of the variables are
inserted into the model, with the exception of the
variables measuring attachment to traditional activities.
In general, the relationships between health status and
the conventional determinants of health are as expected.
0102030405060708090
100
15-24 24-39 40-64 65+
Age Groups
Perc
ent
Female
Male
Fig. 1. Percentage of ‘healthy’ registered Indians by sex.
05
101520253035404550
15-24 25-39 40-64 65+
Age Groups
Perc
ent
Female
Male
Fig. 2. Spent time living on the land by sex.
0
10
20
30
40
50
60
70
80
90
100
15-24 25-39 40-64 65+
Age Groups
Perc
ent
Female
Male
Fig. 3. Percentage of registered Indians who report acquiring
food through hunting, fishing and trapping.
0102030405060708090
100
15-24 25-39 40-64 65+
Perc
ent
Female
Male
Age Groups
Fig. 4. Percentage of registered Indians who report participat-
ing in traditional activities.
0102030405060708090
100
Huntin
g/Fish
ing/T
rappin
gLan
d
Partici
patio
n
Perc
ent
ReserveCMAOther UrbanOther Rural
Fig. 5. Participation in traditional activities by place of
residence.
6 Initially, two variables measuring smoking and drinking
behaviours were included in the model to represent lifestyle.
The results showed no statistically significant relationships
between these two variables and health status. Further,
supplementary analyses explored an interaction of smoking
and drinking behaviours but they did not show a significant
relationship between the combined effects of smoking and
drinking on health status. As a result, these two variables were
removed from subsequent analyses.
K. Wilson, M.W. Rosenberg / Social Science & Medicine 55 (2002) 2017–20312022
With respect to demographic status, the odds ratios
indicate that the likelihood of assessing one’s health as
unhealthy increases with age. For example, compared to
those between the ages of 15–24, those aged 25–39 are
two times more likely to report being unhealthy and
those aged 65 and over are six times more likely to
report being unhealthy. Relative to those who are single,
the odds ratios show that individuals who are married
are less likely to report being unhealthy. This finding is
consistent with previous research that suggests social
support positively affects health (Bloom, 1990). Inter-
estingly, sex is not significantly related to health status.
In terms of the socioeconomic variables, the odds
ratios for income reveal that as income increases, the
likelihood of reporting being unhealthy decreases. The
odds ratio for employment status is reflective of the
‘healthy worker’ effect (see Dahl, 1993), with those
individuals not in the labour force reporting they are
more likely to be unhealthy than the employed. In terms
of education, individuals with high school education or
Table 3
Geographic Model of the Determinants of Health for First Nations Peoples
Variable 95% Confidence intervals
For exp (b)
Estimated coefficient Odds ratio exp (b) Lower Upper
Sex (Ref: Male) �0.080 0.923 0.798 1.068
Age (Ref: 15–24)
25–39 0.837*** 2.309 1.811 2.944
40–64 1.583*** 4.871 3.760 6.311
65+ 1.853*** 6.379 4.660 8.730
Marital status (Ref: Single)
Divorced/separated/widowed �0.103 0.902 0.727 1.119
Married �0.245* 0.783 0.650 0.943
Total income (Ref: 0–9,9999)
10,000–19,999 �0.082 0.922 0.782 1.086
20,000–39,999 �0.537*** 0.584 0.456 0.748
40,000+ �1.669*** 0.189 0.097 0.368
Education (Ref: Less than high school)
High school �0.687*** 0.503 0.420 0.603
Post secondary �0.612*** 0.542 0.445 0.661
Employment (Ref: Employed)
Unemployed �0.168 0.845 0.643 1.110
Not in labour force 0.597*** 1.817 1.499 2.202
Physician use (Ref: Yes) �0.718*** 0.488 0.352 0.676
Place of residence (Ref: Reserve)
CMA �0.211* 0.810 0.661 0.993
Other urban �0.123 0.884 0.742 1.053
Other rural 0.009 1.009 0.807 1.262
Constant �2.172***
Model chi-square 980.590***.
Percentage of correct predictions 83.1%.
Sensitivity 12.7%.
Specificity 97.6%.
Rho 0.151.
N 6885.
*po0:05; � � po0:01; � � �po0:001:
K. Wilson, M.W. Rosenberg / Social Science & Medicine 55 (2002) 2017–2031 2023
post secondary education are less likely to report being
unhealthy, as compared to those with less than high
school education.
The results also show that individuals who have not
visited a doctor are less likely to report being unhealthy
than those who have seen a doctor. With respect to place
of residence, individuals who live in Census Metropo-
litan Areas (CMA) are less likely to report being
unhealthy, as compared to those individuals who live
on reserves. The fact that major metropolitan areas have
better and more health services (in addition to other
resources) that cannot be found on many reserves, may
explain this finding. However, this result may also reflect
a process of self-selection among migrants to cities.
Research has shown that many First Nations peoples
migrate to the city in search of education, employment
and health care (Canada, 1993b; Clatworthy, 1994).
Other statistics show that registered Indians living off
reserve enjoy higher levels of socioeconomic status as
compared to registered Indians living on reserve. For
example in 1990, the average individual income for
individuals living off reserve was $12,551 as compared to
$8,812 for individuals living on reserve (DIAND, 1995).
Similar disparities exist for education level (Clatworthy,
1994; DIAND, 1995). Therefore, the finding that
individuals living in CMAs are less likely to report
being unhealthy than individuals living on reserves may
capture a package of advantages of city life (e.g., better
housing opportunities, job opportunities, etc.) including
higher socioeconomic status.
The model chi-square is significant, which allows for
the rejection of the null hypothesis that the predictor
variables are not related to health status. To assess the
fit of this model we must examine the rho-squared value.
While the rho-squared value for this model (0.151),
appears low, analyses based on large sample sizes tend
to deflate the upper bound of the rho-squared by an
unknown amount (McFadden, 1974). The sensitivity of
this model is low (12.7%), while the specificity is high at
97.6%. Overall, the model correctly classified 83.1% of
respondents.
The second model explores the determinants of health
at the national level and includes the variables chosen to
measure attachment to culture/traditional activities (see
Table 4). Once again we see that increased age is
associated with increased odds of being unhealthy, and
higher levels of income and education are associated
with lower odds of being unhealthy. Married respon-
dents are less likely to report being unhealthy relative to
their single counterparts. In addition, those not in the
labour force have higher odds of reporting being
unhealthy as compared to the employed.
The variables chosen as measures of traditional
activities revealed some interesting findings. Individuals
who reported spending time living on the land in the
past year were less likely to report being unhealthy, as
compared to individuals who had not spent time living
on the land. Acquiring food through traditional ways
was also significantly related to health status. The odds
ratio revealed that individuals who reported that they
had not obtained any of their food through hunting,
fishing or trapping were less likely to be unhealthy
compared to those who had reported that they had. The
third proxy measure, participation in traditional activ-
ities, was not statistically significant.
The odds ratio for the hunting/fishing/trapping
variable, at first glance, is opposite to what one might
expect and therefore counter to the argument that
attachment to traditional activities are necessary for
good health. One possible explanation for this counter-
intuitive result is that the hunting variable is serving as a
proxy for a lack of resources to purchase food rather
than any form of attachment to traditional ways. To test
this hypothesis a supplementary model was examined. A
variable measuring whether an individual reported not
having enough food in the past year was added to the
model and a set of interaction effects between the food
adequacy variable and each measures for attachment to
traditional ways was also created. The direct effect for
food adequacy was statistically significant showing that
those who reported not having enough food in the past
year were two times more likely to report being
unhealthy as hypothesized. However, the interaction
effects between food adequacy and the proxy measures
were not statistically significant. While hunting, fishing
and trapping may indeed be supplementing the diets of
respondents as well as maintaining attachments to a
traditional way of life, the lack of an interaction effect
makes it difficult to interpret the hunting variable in a
clear way.
Once again the model chi-square ratio statistic is
significant. The inclusion of the cultural variables has
resulted in a slightly higher rho-squared value (0.154), as
compared to the previous model. In addition, the
specificity has increased (98.1%) while the sensitivity
(12.5%) decreased slightly. Overall, this model correctly
classified 83.8% of respondents.
It is also important to highlight the statistically
significant difference in health status observed for
those living on reserves as compared to those living in
CMAs (see Table 4). Exploring the determinants of
health by location is important because the proportion
of the Aboriginal population living in urban centres is
steadily increasing. In 1991, 44% (320,000) of the total
population was living in urban areas (Canada, 1996d)
and this figure is predicted to grow by another 43% over
the next 25 years (Kerr, Siggner, & Bourdeau, 1995).
Much of the relocation literature focuses on the
movement of whole communities. However, a smaller
body of literature, which discusses individual relocation,
argues that individuals who move away from the
reserve/community also face problems (Duck, 1993;
K. Wilson, M.W. Rosenberg / Social Science & Medicine 55 (2002) 2017–20312024
Redwolf, 1995; Richardson, 1993). The literature
contends that individuals who move from the reserve
to urban area face lower levels of access to their
traditional activities. As Peters (1996, p. 321) notes,
few urban Aboriginal institutions have ‘‘as their primary
mission the promotion or support of aboriginal culture
and identity’’. Despite the improved levels of socio-
economic status associated with living in urban
areas, urban migrants face diminished levels of access
to traditional activities, identity and the land, all of
which can cause psychological and emotional health
problems.
Table 4
Cultural model of the determinants of health for First Nations peoples
Variable 95% Confidence intervals
For exp (b)
Estimated coefficient Odds ratio exp (b) Lower Upper
Sex (Ref: Male) �0.084 0.920 0.788 1.074
Age (Ref: 15–24)
25–39 0.844*** 2.327 1.797 3.012
40–64 1.615*** 5.026 3.819 6.615
65+ 1.820*** 6.172 4.419 8.622
Marital status (Ref: Single)
Divorced/Separated/Widowed �0.143 0.866 0.689 1.089
Married �0.257* 0.773 0.635 0.942
Total income (Ref: 0–9,9999)
10,000–19,999 �0.081 0.922 0.775 1.096
20,000–39,999 �0.493*** 0.611 0.471 0.793
40,000+ �1.964*** 0.140 0.063 0.313
Education (Ref: Less than high school)
High school �0.715*** 0.489 0.403 0.594
Post secondary �0.551*** 0.576 0.468 0.710
Employment (Ref: Employed)
Unemployed �0.236 0.790 0.589 1.059
Not in labour force 0.605*** 1.832 1.495 2.245
Physician use (Ref: Yes) �0.740*** 0.477 0.337 0.676
Place of residence (Ref: Reserve)
CMA �0.212* 0.809 0.649 1.009
Other urban �0.157 0.855 0.703 1.039
Other rural 0.043 1.044 0.825 1.320
Land (Ref: No) �0.203* 0.816 0.673 0.990
Hunt (Ref: Yes) �0.212* 0.809 0.681 0.961
Traditional activities (Ref: Yes) 0.074 1.077 0.927 1.251
Constant �2.122***
Model chi-square 904.174***.
Percentage of correct predictions 83.8%.
Sensitivity 12.5%.
Specificity 98.1%.
Rho 0.154.
N 6269.
*po0:05; **po0:01; ***po0:001:
K. Wilson, M.W. Rosenberg / Social Science & Medicine 55 (2002) 2017–2031 2025
The model results, the demographic trends and the
relocation literature all point to the need to disaggregate
the data by place of residence and run separate models
for those individuals living on reserves and those living
in CMAs. As Table 5 shows, in both locations, increased
age is associated with higher odds of being unhealthy
and those not in the labour force have higher odds of
being unhealthy. The results also show differences in the
determinants of health between the two locations. For
example, visiting a physician is only significantly related
to health status on reserves. The odds ratio shows that
individuals who have not visited a physician are less
likely to be unhealthy, as compared to those that have
visited a physician. In addition, marital status is only
significantly related to health status in CMAs and the
odds ratios reveal that relative to single respondents,
Table 5
Determinants of health for First Nations peoples by place of residence
Variable Reserve CMA
Estimated
coefficient
Odds ratio
exp (b)Lower Upper Estimated
coefficient
Odds ratio
exp (b)Lower Upper
Sex (Ref: Male) �0.045 0.956 0.796 1.148 �0.420 0.657 0.373 1.157
Age (Ref: 15–24)
25–39 0.797*** 2.218 1.574 3.125 1.304** 3.684 1.716 7.908
40–64 1.801*** 6.059 4.271 8.595 2.158*** 8.654 3.609 20.755
65+ 2.218*** 9.187 6.141 13.745 1.923** 6.838 1.864 25.081
Marital status (Ref: Single)
Divorced/Sepa-
rated/widowed
�0.087 0.917 0.685 1.227 �0822* 0.439 0.221 0.874
Married �0.178 0.837 0.656 1.068 �0.886* 0.412 0.202 0.840
Total income (Ref: 0–9,9999)
10,000–19,999 �0.013 0.987 0.796 1.225 0.355 1.426 0.837 2.430
20,000–39,999 �0.151 0.859 0.612 1.207 �0.487 0.615 0.240 1.576
40,000+ �0.574 0.563 0.245 1.296 �1.788 0.167 0.008 3.440
Education (Ref: Less than high school)
High school 0.532*** 1.702 1.334 2.173 �0.688* 0.502 0.256 0.984
Post-secondary 0.152 1.164 0.867 1.562 �1.140** 0.320 0.154 0.666
Employment (Ref: Employed)
Unemployed 0.001 1.001 0.691 1.450 0.158 1.171 0.476 2.879
Not in labour
Force
0.753*** 2.123 1.621 2.781 1.265*** 3.543 1.776 7.067
Physician use (Ref:
Yes)
�0.753*** 0.471 0.321 0.690 �0.867 0.420 0.062 2.849
Land (Ref: No) �0.096 0.908 0.735 1.121 �0.875 0.417 0.149 1.170
Hunt (Ref: Yes) �0.079 0.924 0.727 1.174 0.176 1.192 0.723 1.968
Traditional activities
(Ref: Yes)
0.234* 1.263 1.049 1.522 �0.412 0.663 0.404 1.087
Constant �3.238*** �2.659***
Model chi-square 630.778*** 116.557*** *po0:05 **po0:01 ***po0:001Percentage of correct predictions 79.7% 88.0%
Sensitivity 24.5% 12.4%
Specificity 94.4% 98.1%
N 3732 812
Rho 0.16 0.20
K. Wilson, M.W. Rosenberg / Social Science & Medicine 55 (2002) 2017–20312026
married and divorced respondents are less likely to be in
poor health.
When exploring the variables that act as measures
for traditional activities, the results show that only
one variable is significantly associated with health
status. For individuals living on reserves, those who
have not participated in traditional activities are 1.3
times more likely to report being unhealthy, as compared
to individuals who have participated in traditional
activities providing additional support for the importance
that traditional ways play in the lives of First Nations
peoples.
Both models had higher rho-squared values (0.16
(reserves) and 0.20 (CMAs)) and good specificity,
94.4% for reserves and 98.1% for CMAs. Sensitivity
was much higher for the reserves (24.5%), as compared
to CMAs (12.4%), suggesting that the determinants of
health are better predictors of ‘unhealthy’ health
status on reserves than in CMAs. Overall, the models
correctly classified 88% (CMAs) and 79.7% (reserves) of
respondents.
From traditional activities to cultural attachment
The purpose of this study was to examine if the
inclusion of variables measuring traditional activities in
a social determinants of health framework could further
our understanding of First Nations peoples’ health. The
results from this analysis raise a number of issues. First,
the majority of the variables that were statistically
significantly related to health (i.e., age, income, educa-
tion, employment, utilization of health care and place of
residence) are similar to variables that are commonly
shown to be determinants of health for the general
Canadian population (see Badgley, 1993; Canada,
1995b, 1996c; Canada, 1998; Grayson, 1993, Wilson,
Jerrett, & Eyles, 2001). This implies that the health of
First Nations peoples is dependant on similar determi-
nants of health as those found for the Canadian
population in general and all other things being equal
it follows that the same policies and programs used to
improve the health of Canadians in general should be
sufficient to improve the health of First Nations peoples.
Indeed, there is superficial evidence to support this
viewpoint when one examines the convergence in
morbidity and mortality rates between First Nations
peoples and the Canadian population during the
twentieth century (see Waldram, Herring, & Young,
1995).
Second, the findings of this study revealed only a
limited number of significant relationships between
traditional activities and health even when the data
were disaggregated to examine only those individuals
living on reserves compared to those living in CMAs. As
noted earlier, First Nations peoples suffer from higher
levels of morbidity and mortality as compared to the
general Canadian population even though convergence
is taking place over time. The same research (i.e.,
Waldram et al., 1995), does, however, also show that
there remains a gap in the morbidity and mortality rates
between First Nations peoples and the general Canadian
population and gaps between First Nations peoples
living on isolated reserves mainly in the northern parts
of most provinces and First Nations peoples living on
reserves in southern Canada and near to major urban
places. Without taking these trends into account, the
findings in this paper might lead one to conclude: (i) that
traditional activities are not important determinants of
health; and/or (ii) First Nations peoples’ culture, in
general, cannot be conceptualized within a determinants
of health framework. Each of these ‘conclusions’ needs
to be considered.
First, utilization and access to health care services are
important determinants of health. Within the APS, the
utilization of conventional health care services (e.g.,
physician, nurse, eye specialist) is emphasized. Incorpor-
ating a conventional measure of health care, such as
physician utilization, into this analysis is problematic
since there are no regular physician services and access is
highly circumscribed for those First Nations peoples
living on isolated reserves. In addition, the survey under-
emphasizes the importance of traditional health prac-
tices. While the APS measures the utilization of
traditional healers, it does not take into account other
traditional healing practices, such as the use of sweat
lodges (see Adair, Deuschle, & Barnett, 1988; Kunitz,
1989; Waldram, 1997, 1993; Waldram et al., 1995;
Wilbush, 1988). Unfortunately, only 8% of respondents
in the APS indicated that they had used a traditional
healer, but there is no way to determine how many
reserves actually have a traditional healer and whether
respondents from isolated reserves are under-repre-
sented where they might be most likely to use traditional
healers and other traditional health practices (see
below).
Second, with respect to measures of traditional
activities and health status, there is a certain level of
ambiguity in how one interprets the direction of the
relationships. For example, the odds ratio for the
hunting variable revealed that individuals who reported
that they had not obtained any of their food through
hunting, fishing or trapping were less likely to be
unhealthy compared to those who had reported that
they had. This finding may indicate that hunting, fishing
and trapping supplement diets and therefore the
relationship represents a positive covariate for health.
Conversely, it may be a proxy measure for poverty and
therefore can be interpreted as a negative covariate for
health. This level of ambiguity indicates a need for more
measures of traditional activities that do not conflate
traditional activities in one aggregated variable. For
K. Wilson, M.W. Rosenberg / Social Science & Medicine 55 (2002) 2017–2031 2027
example, the APS defines traditional activities as a
wide range of activities: ‘‘hunting, fishing, trapping,
storytelling, traditional dancing, fiddle playing, jigging,
arts and crafts, pow-wows, etc.’’ (Canada, 1991, p. 6).
As such, there is no way to tease out which of these
traditional activities may be a significant determinant of
health.
Third, the analysis is also constrained by the
limitations inherent in using a public use microdata
file and more general data collections encountered by
Statistics Canada in developing the APS. Due to reasons
of confidentiality and anonymity the data cannot be
divided into northern and non-northern regions nor
can specific bands or communities be identified.
Consequently, it is not possible to explore how the
relationship between traditional activities and health
varies by location. For example, it is reasonable to
believe that the health experiences of a First Nations
person living on an isolated reserve far from any
urban place, an urban First Nations person living in a
northern Ontario town such as Kenora, and an
urban First Nations person living in Toronto would all
differ.
A final limitation of the survey is its categorization of
Aboriginal peoples into three groups; North American
Indian, M!etis and Inuit. Contemporary feminist litera-
ture has explored the impact of essentialized categoriza-
tion on gender, race and sexuality (Kobayashi & Peake,
1994; McDowell, 1991; McDowell, 1992; Nicholson,
1995; Penrose et al, 1992; Pile, 1994; Rose, 1993). The
very basis of these categories is difference, with
categories set up in opposition to one another (Kobaya-
shi & Peake, 1994). This structure has at least two
inherent problems. First, naturalized and essentialized
categories limit differences to polarized opposites.
Second, as Rose (1993) argues, such categories often
deny and do not allow for radical difference within
oppositions.
At first glance, the analysis shows that, in general,
traditional activities are not conclusively linked to health
status. Due to the essentialized categorization of
Aboriginal identity, it is not possible to explore this
relationship for different First Nations groups, such as
the Ojibwa, Mohawk and Dene. The categories North
American Indian, M!etis, and Inuit are not sufficient for
acknowledging the distinct identities within each group.
As such, the survey denies the existence of multiple
Aboriginal identities. Thus, this categorization is a crude
measure of ethnicity that creates three homogeneous
categories of Aboriginal peoples, which do not actually
exist.
If the analysis of First Nations peoples’ health is to
move from a focus on traditional activities to cultural
attachment, future population health surveys like the
APS will need to take into account other potentially
important cultural determinants of health. First Nations
cultures are deeply rooted in spirituality. Many First
Nations peoples participate in activities or ceremonies
such as sweat lodges, the use of sacred medicines,
and the offering of tobacco, as part of their spiritual
way of life (Akiwenzie-Damm, 1996; Canada, 1996a,
1996b, 1996d, 1996e). Taking into account the other
limitations of this analysis as discussed above would
also help to breakdown the dichotomy between
the medical literature, which is characterized by
quantitative and epidemiologic research surrounding
the prevalence of illness and disease and on the other
side a qualitative body of research that is concerned with
the specificity of First Nations peoples, their culture and
their health.
Conclusions
This paper explored the importance of traditional
activities for First Nations peoples’ health based on a
determinants of health framework and employing
sample data from the APS. The research shows the
significance of income, education and employment for
health similar to other analyses of the Canadian
population in general where a determinants of health
framework has been employed. While the inclusion of a
subset of variables measuring traditional activities
resulted in an attenuated set of meaningful results,
taking into account the limitations of the APS in future
surveys has the potential to move research beyond the
division which now exists between epidemiological and
cultural studies.
In 2001, Statistics Canada began collecting data
for the second Aboriginal Peoples Survey. This will be
an important data set as it will allow us to explore
whether the relationships identified in this paper are
consistent over time. In addition, the new survey holds
out the potential to move from the analysis of
traditional activities to a more nuanced analysis of
cultural attachment. In the 2001 APS, there is disag-
gregation of traditional activities. For example, indivi-
duals are asked if they participate in four categories of
activities (hunting, fishing, gathering wild plants and
trapping) and the reasons for participation in those
activities (e.g., food, pleasure, commercial, and medic-
inal/ceremonial) (Canada, 2001). In this way, one
can distinguish between activities performed for
primarily economic reasons from those aimed primarily
at improving health. In terms of health care, the
survey also explores the availability of traditional
medicines, healing and wellness practices. Given these
improvements, future research using data from the
2001 APS may be better able to enhance our under-
standings of the role of traditional activities in shaping
health.
K. Wilson, M.W. Rosenberg / Social Science & Medicine 55 (2002) 2017–20312028
Acknowledgements
Funding for this project was provided by the Social
Science and Humanities Research Council of Canada.
Special thanks to Bob Earickson and three anonymous
reviewers for their thoughtful comments.
References
Adair, J., Deuschle, K., & Barnett, C. (1988). The people’s
health: Anthropology and medicine in a Navajo Community.
Albuquerque: University of New Mexico Press.
Adelson, N. (1998). Health beliefs and the politics of Cree well-
being. Health: An Interdisciplinary Journal for the Social
Study of Health, Illness and Medicine, 2(1), 5–22.
Akiwenzie-Damm, K. (1996). We belong to this land: A view of
cultural difference. Journal of Canadian Studies, 31(3),
21–28.
Aldrich, J. H., & Nelson, F. D. (1984). Linear probability, logit
and probit models. Beverly Hills: Sage Publications.
Badgley, R. F. (1993). Social and economic disparities under
Canadian health care. International Journal of Health
Services, 21(4), 659–671.
Bagley, C. (1991). Poverty and suicide among native
Canadians: A replication. Psychological Reports, 69(1),
149–150.
Bloom, J. R. (1990). The relationship of social support and
health. Social Science and Medicine, 30(5), 635–637.
Borr!e, K. (1991). Seal blood, Inuit blood and diet: A biocultural
model of physiology and cultural identity. Medical Anthro-
pology Quarterly, 5, 48–62.
Borr!e, K. (1994). The healing power of the seal: The meaning of
Inuit healing practice and belief. Arctic Anthropology, 31(1),
1–15.
Canada (1991). Aboriginal Peoples Survey 1991 Microdata
FileFAdults. Ottawa: Statistics Canada.
Canada (1993a). Information package on Aboriginal products
from the 1991 Aboriginal peoples survey and the 1991
census of population. Ottawa: Statistics Canada.
Canada (1993b). Aboriginal peoples in urban centres: Report of
the National Round Table on Aboriginal Urban Issues.
Ottawa: Royal Commission on Aboriginal Peoples.
Canada (1994–95). National Population Health Survey
(NPHS). Ottawa: Statistics Canada.
Canada (1995a). The 1991 Aboriginal Peoples Survey
Microdata FileFAdults User’s Guide. Ottawa: Statistics
Canada.
Canada (1995b). National Population Health Survey Overview
1994–95. Ottawa: Statistics Canada.
Canada (1996a). Looking Forward, Looking Back: Report of
the Royal Commission on Aboriginal Peoples, Vol. 1.
Ottawa: Ministry of Supply and Services.
Canada (1996b). Restructuring the Relationship: Report of the
Royal Commission on Aboriginal Peoples, Vol. 2, Part 1.
Ottawa: Ministry of Supply and Services.
Canada (1996c). Restructuring the Relationship: Report of the
Royal Commission on Aboriginal Peoples, Vol. 2, Part 2.
Ottawa, Ministry of Supply and Services.
Canada (1996d). Gathering Strength: Report of the Royal
Commission on Aboriginal Peoples, Vol. 3. Ottawa:
Ministry of Supply and Services.
Canada (1996e). Perspectives and Realities: Report of the
Royal Commission on Aboriginal Peoples, Vol. 4. Ottawa:
Ministry of Supply and Services.
Canada (1996–97). National Population Health Survey
(NPHS). Ottawa: Statistics Canada.
Canada (1998). Basic Departmental Data. Ottawa: Department
of Indian Affairs and Northern Development.
Canada (2001). Aboriginal Peoples SurveyF2001. Ottawa:
Statistics Canada.
Clatworthy, S. (1994). Draft final report: The migration and
mobility patterns of Canada’s aboriginal population. Winni-
peg: Four Directions Consulting Group.
Dahl, E. (1993). Social inequality in healthFthe role of the
healthy worker effect. Social Science and Medicine, 36(8),
1077–1086.
Department of Indian Affairs and Northern Development
(DIAND) (1995). Highlights of Aboriginal Conditions
1991, 1986: Demographic, Social and Economic
Characteristics. Ottawa: Information Quality and Research
Directorate Information Management Branch Corporate
Services.
Duck, E. (1993). I forget who I am. In Joel T. Maki (Ed.), Steal
my rage (p. 40). Toronto: Douglas and McIntyre.
Enarson, D. A., & Grzybowski, S. (1986). Incidence of active
tuberculosis in the Native population of Canada. Canadian
Medical Association Journal, 134, 1149–1152.
Epp, J. (1986). Achieving health for all: A framework for health
promotion. Ottawa: Minister of Supply and Services
Canada.
Evans, R. G., & Stoddart, G. L. (1990). Producing health,
consuming health care. Social Science and Medicine, 31(12),
1347–1363.
Evers, S., McCracken, E., Antone, I., & Deagle, G. (1987). The
prevalence of diabetes in Indians and Caucasians living in
southwestern Ontario. Canadian Journal of Public Health,
78(4), 240–243.
Federal, Provincial, and Territorial Advisory Committee on
Population Health (ACPH). (1994). Strategies for Popula-
tion Health: Investing in the Health of Canadians. Prepared
for the meeting of the Ministers of Health, Halifax, Nova
Scotia, September 14–15, 1994. Ottawa: Minister of Supply
and Services Canada.
Federal, Provincial, and Territorial Advisory Committee on
Population Health. (ACPH), (1996). Report on the Health of
Canadians. Prepared for the meeting of Ministers of Health,
Toronto, Ontario, September 10–11, 1996. Ottawa: Minister
of Supply and Services Canada.
Foggin, P. M., & Aurillon, N. (1989). Respiratory health
indicators and acculturation among the Inuit and Cree
of northern QuWbec: A regional approach using geographicseriation analysis. Social Science and Medicine, 29,
617–626.
Garro, L. C. (1995). Individual or societal responsibility?
Explanations of diabetes in an Anishinaabe (Ojibway)
community. Social Science and Medicine, 40, 37–46.
Garro, L. C. (1988). Explaining high blood pressure: Variation
in knowledge about illness. American Ethnologist, 15(1),
98–119.
K. Wilson, M.W. Rosenberg / Social Science & Medicine 55 (2002) 2017–2031 2029
Grayson, J. P. (1993). Health, physical activity level and
employment status in Canada. International Journal of
Health Services, 23(4), 743–761.
Hagey, R. (1989). The Native diabetes program: Rhetorical
process and praxis. Medical Anthropology, 12, 7–33.
Hammond, G. W., Rutherford, B. E., Malazdrewicz, R., &
MacFarlane, N. (1988). Haemophilus Influenzae Meningitis
in Manitoba and the Keewatin District, NWT: Potential for
mass vaccination. Canadian Medical Association Journal,
139, 743–747.
Harris, S. B., Caulfield, L. E., Sugamori, M. E., Whalen, E. A.,
& Henning, B, (1997). The epidemiology of diabetes in
pregnant Native Canadians. A risk profile. Diabetes Care,
20(9), 1422–1425.
Hegele, R. A., Harris, S. B., Hanley, A. J., Sun, F., Connelly, P.
W., & Zinman, B. (1997). Angiotensinogen gene variation
associated with variation in blood pressure in Aboriginal
Canadians. Hypertension, 29(5), 1073–1077.
Hertzman, C., Frank, J., & Evans, R.G., Evans. (1994).
Heterogeneities in health status and the determinants of
population health. In R.G. Evans, M.L. Barer, & T.R.
Marmor (Eds.) Why are some people healthy and others not?
The determinants of health of populations (pp. 67–92). New
York: Aldine De Gruyter.
Jilek, W. (1982). Indian healing: Shamanic ceremonialism in the
Pacific Northwest today. Surrey, B.C: Hancock House.
Kerr, D., Siggner, A., & Bourdeau, J.P. (1995). Canada’s
Aboriginal Population, 1981–1991: A Summary Report.
Research study prepared for the Royal Commission on
Aboriginal Peoples. Ottawa: Ministry of Supply and Services.
Kobayashi, A., & Peake, L. (1994). Un-natural discourse:
‘Race’ and gender in geography. Gender, Place and Culture,
1(2), 244–255.
Kunitz, S. J. (1989). Disease change and the role of medicine:
The Navajo experience. Berkeley: University of California
Press.
Lalonde, M. (1974). A new perspective on the health of
Canadians: A working document. Ottawa: Minister of
Supply and Services Canada.
Mao, Y., Moloughney, B. W., Semenciw, R., & Morrison, H.
(1992). Indian reserve and registered mortality in Canada.
Canadian Journal of Public Health, 83, 350–353.
McDowell, L. (1991). The baby and the bath water: Diversity,
deconstruction and feminist theory in geography. Antipode,
22(2), 123–133.
McDowell, L. (1992). Diversity and differenceFunity in
plurality. Antipode, 24(3), 223–227.
Moffatt, M. E. (1995). Current status of nutritional deficiencies
in Canadian Aboriginal people. Canadian Journal of
Physiology and Pharmacology, 73(6), 754–758.
Neuwalt, P. M., Kearns, R. A., Hunter, D. J. W., & Batten, J.
(1992). Ethnicity, morbidity and health service utilization in
two Labrador communities. Social Science and Medicine,
34(2), 151–160.
Newbold, K. B. (1997). Aboriginal physician use in Canada:
Location, orientation and identity. Health Economics, 6,
197–207.
Ng, E. (1996). Disability among Canada’s Aboriginal Peoples
in 1991. Health Reports, 8(1), 25–32.
Nicholson, L. (1995). Interpreting gender. In L. Nicholson, & S.
Seidman (Eds.), Social postmodernism: Beyond identity
politics (pp. 39–67). Cambridge: Cambridge University
Press.
Ontario (1990). Ontario Health Survey. Toronto: Premier’s
Council on Health, Well-Being and Social Justice.
Ontario (1991). Nurturing Health: A Framework on the
Determinants of Health. Toronto: Premier’s Council on
Health Strategy.
Penrose, J., Bondi, L., McDowell, L., Kofman, E., Rose, G., &
Whatmore, S. (1992). Feminists and feminism in the
academy. Antipode, 24(3), 218–237.
Peters, E. (1996). Aboriginal people in urban areas. In D. A.
Long, & O. P. Dickason (Eds.), Visions of the heart:
Canadian Aboriginal issues (pp. 305–344). Toronto:
Harcourt Brace and Company.
Pile, S. (1994). Masculinism, the use of dualistic epistemologies
and third spaces. Antipode, 26(3), 255–277.
Qu!ebec (1987). Sant!e Qu!ebec. Quebec City: Ministry of Health
and Social Services.
Qu!ebec (1992–93).Sant!eQu!ebec.Qu!ebec City:Ministry of Health
and Social Services.
Redwolf, D. (1995). Native transitional problems. In J. T. Maki
(Ed.), Steal my rage (pp. 34–36). Toronto: Douglas and
McIntyre.
Richardson, B. (1993). People of Terra Nullius: Betrayal and
rebirth in aboriginal Canada. Toronto: Douglas and
McIntyre.
Rose, G. (1993). Feminism and geography: the limits of
geographical knowledge. Minneapolis: University of Minne-
sota Press.
Saskatchewan (1997). Population Health Promotion Model: A
Resource Binder. Regina: Population Health Branch Sas-
katchewan Health.
Thouez, J. P., Rannou, A., & Foggin, P. (1989). The other face
of development: Native population, health status and
indicators of malnutrition-the case of the Cree and Inuit
of northern Quebec. Social Science and Medicine, 29,
965–974.
Waldram, J. B. (1990). Physician utilization and urban native
people in Saskatoon, Canada. Social Science and Medicine,
30(5), 579–589.
Waldram, J. B. (1993). Aboriginal spirituality in corrections: A
Canadian case study in religion and therapy. American
Indian Quarterly, 18(2), 197–215.
Waldram, J. B. (1997). The way of the pipe: Aboriginal
spirituality and symbolic healing in Canadian prisons.
Peterborough: Broadview Press.
Waldram, J. B., Herring, D. A., & Young, T. K. (1995).
Aboriginal health in canada: historical, cultural and epide-
miological perspectives. Toronto: University of Toronto
Press.
Wilbush, J. (1988). Placebo effects associated with sweatlodge
therapy. In D. Young (Ed.), Health Care Issues in the
Canadian North (pp. 102–107). Edmonton: Boreal Institute
for Northern Studies.
Wilson, K., Jerrett, M., & Eyles, J. (2001). Testing relationships
among determinants of health, health policy and self-
assessed health status in Qu!ebec, Canada. International
Journal of Health Services, 31(1), 67–89.
Young, T. K. (1991). Prevalence and correlates of Hypertension
in a Subarctic Indian population. Preventive Medicine, 20,
474–485.
K. Wilson, M.W. Rosenberg / Social Science & Medicine 55 (2002) 2017–20312030
Young, T. K. (1994). The health of native americans: Towards a
biocultural epidemiology. New York: Oxford University
Press.
Young, T. K. (1998). Population health: concepts and methods.
New York: Oxford University Press.
Young, T. K., McIntyre, L. L., Dooley, J., & Rodriguez,
J. (1985). Epidemiologic features of diabetes mellitus
among Indians in northwestern Ontario and northeastern
Manitoba. Canadian Medical Association Journal, 132(7),
793–797.
K. Wilson, M.W. Rosenberg / Social Science & Medicine 55 (2002) 2017–2031 2031