Health Reports
Smoking correlates among Inuit men and women in Inuit Nunangat

Warning View the most recent version.

Archived Content

Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please "contact us" to request a format other than those available.

by Evelyne Bougie and Dafna E. Kohen

Release date: March 21, 2018

The high prevalence of smoking among Inuit in Canada is well-documented.Note 1Note 2Note 3Note 4Note 5Note 6 Although rates of daily smoking among Inuit appear to be decreasing,Note 6 they remain much higher than among the total Canadian population.Note 3Note 4 This is especially the case in Inuit Nunangat. For instance, in 2012, 63% of Inuit aged 15 or older in Inuit Nunangat reported smoking cigarettes daily, compared with 16% of the Canadian population overall.Note 4 Adverse health outcomes associated with smoking include cancer, respiratory illness, heart disease, and stroke.Note 7 Lung cancer is a growing public health concern among Inuit in the Arctic.Note 8

Few empirical population-based studies have investigated factors associated with cigarette use among Inuit. Research is needed to identify characteristics of Inuit who are at risk of smoking in order to inform culturally appropriate cessation and prevention programs. Using data from the 2012 Aboriginal Peoples Survey (APS), an Inuit-specific social determinants of health framework,Note 5 and a gender-based analysis, the present study examines correlates of smoking among Inuit men and women aged 18 or older living in the four regions collectively known as Inuit Nunangat (Nunavik in Northern Quebec, Nunatsiavut in Northern Labrador, Nunavut, and the Inuvialuit Region of the Northwest Territories).

Smoking among Inuit: A social determinants approach

Nicotine addiction plays an important role in cigarette use, but individual, social, and economic factors also influence smoking behaviours.Note 9 Many researchers have emphasized the importance of examining Inuit health from a social determinants perspectiveNote 2Note 10 with a focus on the underlying processes.Note 11 Inuit Tapiriit Kanatami (ITK),Note 5 the national representational organization of Inuit in Canada, has identified 11 interrelated factors as key heath determinants: quality of early childhood development; culture and language; livelihood; income distribution; housing; personal safety and security; education; food security; availability of health services; mental wellness; and the environment. This framework has been used by others to investigate Inuit health.Note 12Note 13

Many of these health determinants have been associated with smoking in the general population, but no studies to date have focused specifically on Inuit. For instance, smoking is related to socioeconomic factors, such as unemployment, low educational attainment, and low income, in both the generalNote 9Note 14 and AboriginalNote 15Note 16Note 17 populations. Smoking is also related to mental wellness—smoking and depression often co-occur in both the generalNote 18 and AboriginalNote 19Note 20 populations. Co-occurrence of smoking and alcohol addiction is frequentNote 21 and has been observed in the Aboriginal population as well.Note 15Note 17Note 22 Smoking has also been related to low social support.Note 16Note 19Note 23

Other determinants of Inuit health identified in ITK’s framework could be associated with smoking but have not been empirically investigated. Overcrowding, a particular concern among Inuit,Note 1Note 4 may act as a stressor increasing the likelihood of substance abuse and other social problems.Note 10 A recent study reported an association between crowding and higher psychological distress among Inuit women.Note 12 While evidence indicates that living in a smoke-free home reduces cigarette use,Note 24 crowded conditions may increase exposure to smokers in the home.

Food insecurity is also well-documented among Inuit, especially in Inuit Nunangat.Note 1Note 5Note 25 It exists when, because of a lack of money, one or more members of a household do not have access to the variety or quantity of food that they need.Note 26 In some Inuit communities, the cost of store-bought healthy food is at least twice as high as in southern Canada.Note 1 Food insecurity among Inuit has been connected to socioeconomic conditions (such as unemployment and crowding), weaker extended family ties, and less favourable self-rated health,Note 25 as well as to higher psychological distress.Note 12 Because of relationships between smoking and both socioeconomic conditions and mental wellness, an association between smoking and food insecurity might also be expected.

It has been argued that cultural erosion negatively affects Aboriginal people’s well-being.Note 27Note 28 Inuit culture is an important determinant of health.Note 5 Inuit who experience cultural continuity by participating in traditional activities might be less likely to engage in smoking behaviour, although a recent study showed the opposite to be true among off-reserve First Nations and Métis adults.Note 17 Another cultural factor is the legacy of the residential school system,Note 5 which has directly and indirectly affected Aboriginal people’s health and well-being.Note 10Note 11Note 29Note 30 Because smoking may be a way to cope with stress, distress, and disadvantage,Note 19Note 23 each of which has been related to residential schooling,Note 29 an association might exist between cigarette use and residential school experiences.

Based on availability in the 2012 APS, the present study uses an Inuit-specific health frameworkNote 5 to analyze smoking in the context of labour force status, participation in traditional activities, education, household income, crowding, presence of a regular smoker in the home, strength of family ties, food security, diagnosed mood and/or anxiety conditions, heavy drinking, and residential school experiences. In line with calls for gender-based analysis in health and tobacco research,Note 31Note 32 factors associated with smoking were examined separately for Inuit men and women.



The 2012 APS is a nationally representative cross-sectional survey of First Nations people living off reserve, Métis, and Inuit, developed by Statistics Canada. Participation was voluntary. Data were collected directly from respondents through personal interviews or through computer-assisted interviews. Interviews by proxy were allowed. The overall response rate for Inuit Nunangat was 76%. Response rates for the individual regions were: Nunatsiavut, 81%; Nunavik, 77%; Nunavut, 75%; and the Inuvialuit Region, 71%.

Study sample

The study sample consisted of APS respondents aged 18 or older who self-identified as Inuit and who were living in Inuit Nunangat at the time of data collection. About 7% of the initial study sample had missing smoking data and were excluded. The sample size was 2,614 Inuit—1,263 men (mean age 36.7) and 1,351 women (mean age 38.8). About 7% resided in Nunatsiavut, 22% in Nunavik, 62% in Nunavut, and 9% in the Inuvialuit Region. Around 11% were interviewed by proxy. Percentages of missing data for smoking status were similar for proxy and non-proxy respondents, although smoking prevalence rates were lower among proxy than non-proxy respondents (data not shown).


The APS smoking question was: “At the present time do you smoke cigarettes daily, occasionally, or not at all?” The dependent variable for this study was being a current smoker (daily or occasional) versus being a non-smoker. This way of categorizing smokers is in line with recent research that used the 2012 APS to examine smoking among off-reserve First Nations and Métis adults.Note 17

Respondents’ labour force status during the APS reference week was coded as “employed,” “unemployed,” or “not in labour force.” Employed persons were those who, during the reference week, did any paid work. Unemployed persons were those who were looking for work, were on temporary layoff, or had a job to start within four weeks. Persons not in the labour force were neither employed nor unemployed, including those who were either unable to work or unavailable for work (for example, retirees, homemakers, students, or persons permanently unable to work).

Participation in traditional activities (yes/no) indicated whether, in the past year, respondents had done any of the following: made clothing or footwear; arts or crafts; hunted, fished, or trapped; gathered wild plants.

Education was coded as “currently attending school” or highest level attained: “less than high school,” “high school,” or “postsecondary education.”

Household income was operationalized as the after-tax income (in dollars) of all household members, adjusted by a factor accounting for household size.

Household crowding was coded binary (yes/no) to indicate whether respondents lived in dwellings with more than one person per room.Note 33

Regular smoker in the home (yes/no) indicated whether respondents reported anyone smoking inside their home every day or almost every day.

Strength of family ties was measured with the question: “On a scale from 1 to 5, how strong are the ties among members of your family living in your city, town or community but in another household?”, coded as “stronger” (5 or 4) or “weaker” (3, 2, or 1). Research has shown the importance of strong family ties for Inuit health,Note 12Note 13 but the studies did not pertain specifically to smoking.

Food security was measured using the U.S. Household Food Security Survey Module Six-Item Short FormNote 34 and coded as “high or marginal” or “low or very low.”

Diagnosed mood and/or anxiety disorder was coded as binary (yes/no) if respondents reported having ever been diagnosed with a mood disorder (such as depression, bipolar disorder, mania, or dysthymia) and/or an anxiety disorder (such as a phobia, obsessive-compulsive disorder or a panic disorder).

Heavy drinking was defined as five or more drinks on a single occasion at least once a month in the past 12 months (yes/no). This definition is not consistent with drinking guidelines for women,Note 35 but it has been used in past research using the APS.Note 16Note 17

Because the residential school experiences derived variable on the APS master file had very high rates of missing data, two indicators were created for this study. The first indicated whether respondents had personally attended a residential school (yes/no); the second, whether respondents’ parents (mother and/or father) had attended a residential school (yes/no/don’t know).

The percentages of missing data for the independent variables ranged between 0.2% and 6%. Missing data were excluded from the denominator for descriptive analyses and from the multivariate analysis. A total of 1,001 Inuit men and 1,093 Inuit women had valid data on all independent variables.

As expected, several independent variables were correlated—but not highly. Correlation coefficients for the overall sample ranged from 0.01 to 0.27. The strongest were between being employed and household income (0.27); being employed and food security (0.21); household income and not being in the labour force (-0.22); household income and not having a high school diploma (-0.23); household income and food security (0.28); and not being in the labour force and not having a high school diploma (0.20). Preliminary analyses showed no concern for collinearity among selected independent variables, with all tolerance values over 0.50Note 36 and all variance inflation factor values under 2.Note 37


Smoking prevalence rates for Inuit men and women were calculated overall and by age group (18 to 24, 25 to 54, and 55 or older) and Inuit region. Preliminary regression analyses (not shown) found better model fit and more similar results when occasional smokers were combined with daily smokers (as opposed to combined with non-smokers or excluded from the model). Bivariate logistic regression analyses examined the unique association of each socio-demographic characteristic with smoking. Multivariate logistic regression analyses examined the association of each socio-demographic characteristic with smoking when the other covariates were held constant. The full model included all socio-demographic characteristics and was adjusted for age, Inuit region, and proxy reporting. Sampling weights were applied to account for the sample design, non-response, and known population totals. A bootstrapping technique with Fay adjustment was used when calculating estimates of variance.Note 38 Statistical significance was set at < 0.05 for all analyses.


Smoking prevalence

In 2012, 75% of Inuit men and 74% of Inuit women aged 18 or older in Inuit Nunangat reported that they smoked. Most were daily smokers—the daily/occasional breakdown was 64% and 11% for men, and 64% and 9% for women (Table 1). A chi-square test showed no significant sex difference in the distribution of smoking prevalence rates (χ2 = 1.3, p = 0.51).

Smoking prevalence rates were distributed differently across age groups (χ2 = 75.7, p < 0.0001 for men and χ2 = 80.2, p < 0.0001 for women), with Inuit men and women aged 55 or older being more likely to be non-smokers. Smoking prevalence rates were also distributed differently across the four Inuit regions (χ2 = 14.0, p < 0.05 for men and χ2 = 13.1, p < 0.05 for women)—residents of Nunatsiavut were generally more likely to be non-smokers.

Bivariate analyses

The odds of smoking were significantly lower among Inuit who lived in food-secure households or in higher-income households (Tables 2 and 3). Among Inuit men, the odds of smoking were significantly lower for those who had a high school diploma. Among both sexes, the odds were significantly higher for those who were unemployed, lived in crowded conditions, lived in homes where a regular smoker was present, and engaged in heavy drinking. Parental residential school experience was also significant: the odds of smoking were higher among Inuit women whose parents had attended residential school (versus those whose parents had not), and among Inuit men who did not know if their parents had attended residential school (versus those whose parents had not).

Multivariate logistic regression analyses

When all the selected socio-demographic characteristics were taken into account (Table 4), significantly lower odds of smoking were observed among Inuit men who lived in higher-income households (OR = 0.85: 95% CI: 0.75 to 0.97). Relative to men who did not have a high school diploma, those who were high school graduates had significantly lower odds of smoking (OR = 0.53; 95% CI: 0.31 to 0.89). Inuit men had significantly higher odds of smoking if they lived in crowded conditions (OR = 1.5; 95% CI: 1.01 to 2.31) or in homes where a regular smoker was present (OR = 1.9; 95% CI: 1.23 to 2.82).

When all the selected socio-demographic characteristics were held constant, the odds of smoking were significantly lower for Inuit women in food-secure households (OR = 0.41; 95% CI: 0.27 to 0.62) (Table 4). Relative to women who were not high school graduates, those with a high school diploma (OR = 0.47, 95% CI: 0.28 to 0.78) or postsecondary education (OR = 0.59; 95% CI: 0.38 to 0.91) had significantly lower odds of smoking. The odds of smoking were significantly higher for Inuit women living in crowded conditions (OR = 1.6; 95% CI: 1.02 to 2.36) or in homes where a regular smoker was present (OR = 2.3; 95% CI: 1.43 to 3.59). Relative to those who had not personally attended a residential school, Inuit women who reported having done so had significantly higher odds of smoking (OR = 2.4; 95% CI: 1.28 to 4.33).

The bivariate associations between smoking and being unemployed, smoking and heavy drinking, and smoking and parental residential school experience were no longer significant in the multivariate analyses.


According to the 2012 APS, three-quarters of Inuit men and women aged 18 or older in Inuit Nunangat smoked cigarettes; most of them were daily smokers. Despite declining rates of daily smoking among Inuit since 1991,Note 6 the prevalence of smoking is still much higher than among the Canadian population as a whole.

Findings from this study are consistent with research that has identified associations between smoking and socioeconomic factors in both the general and Aboriginal populations.Note 9Note 14Note 15Note 16Note 17 In the fully adjusted models, Inuit men and women who were high school graduates were less likely than those who had not obtained a diploma to be smokers. As well, Inuit women with postsecondary education were less likely to smoke than were those who did not have a high school diploma. Smoking was associated with lower household income among Inuit men.

Among Inuit of both sexes, significant relationships were apparent between smoking and household crowding and the presence of a regular smoker in the home. In the present study, 45% of Inuit men and 46% of Inuit women reported living in crowded conditions (more than one person per room); these figures compare with 3% for the total population of Canada.Note 4 Overcrowding in Inuit Nunangat is associated with issues such as lack of affordable housing and high cost of construction.Note 1Note 5 The relationship between crowding and health is complex,Note 39 and many contextual factors (such as other housing and psycho-social stressorsNote 5) were beyond the scope of the present study. Future research that disentangles the specific pathways between smoking and crowding in Inuit Nunangat would be fruitful.

Crowded conditions may increase an individual’s exposure to smokers. In this study, about one-third of Inuit men and a quarter of Inuit women reported living in homes where a regular smoker was present. Research conducted in Nunavik indicated that in 2004, 84% of Inuit homes had some smoking restrictions,Note 40 and in 2007/2008, smoking indoors was not allowed by 76% of Nunavut homes.Note 41 Maintaining a smoke-free home can be difficult. One qualitative study has described the challenges among First Nations women in the northwest region of British Colombia, which included overcrowding; unemployment and more time spent at home; a northern climate not amenable to smoking outside in winter; and difficulty being assertive with guests and family members who smoke.Note 42 Given the association between smoking and crowding, as well as between smoking and living in homes where a regular smoker is present, further research focusing on Inuit Nunangat and perceived difficulty maintaining a smoke-free home would be informative.

Two associations were specific to Inuit women. Even when the other selected socio-demographic variables were taken into account, Inuit women were more likely to smoke if they lived in food-insecure households. Research has shown that Inuit women are more likely than Inuit men to live in food-insecure circumstances, and food insecurity in Inuit Nunangat has been related to unemployment, crowded conditions, and weaker extended family ties.Note 25 A qualitative study of Inuit women in Igloolik (Nunavut) identified food insecurity as an outcome of multiple stresses including climate change and a decline in full-time hunting, reduced availability of country foods and weakened sharing networks, as well as short-term and reactive coping mechanisms that actually increase women’s vulnerability to food insecurity in the long run.Note 43 Food insecurity, therefore, could signal the presence of multiple risk factors, including an association with smoking. Research is warranted to identify the mechanisms at work in this relationship.

The second association with smoking specific to Inuit women was personally having attended a residential school. The factors underlying this association are complex and were not tested in the present study. However, the results point to the importance of considering this aspect of Inuit history when attempting to explain smoking behaviour.


This analysis has a number of limitations. Canadian estimates of smoking prevalence based on self-reports are similar to estimates based on urinary cotinine concentrations—a biomarker of tobacco smoke exposure.Note 44 However, the validity of self-reported cigarette smoking data has not been determined for Inuit.

Associations between smoking and the variables examined are strictly correlational; causality cannot be inferred, and some relationships could be bi-directional. Another limitation is shared method variance; all measures—smoking status and selected covariates—were based on self-reported data collected at a single point in time from the same participants. Also, the covariates are inter-related, and although associations between smoking and these factors are discussed as independent relationships, the complex interplay among them in real life is not captured in the present analyses.

Among Inuit, the number of occasional smokers was much smaller than the number of daily smokers. The 2012 APS only asked follow-up questions (frequency and duration of smoking) if respondents were daily smokers. Therefore, it was not possible to establish similarities between occasional and daily smokers.

The content of the 2012 APS limited the ability to examine some Inuit-specific social determinants of health from ITK’s framework (for instance, the environment or early childhood development).

Lastly, proxy reporting could have introduced bias. Preliminary analyses revealed that smoking rates were lower among proxy than non-proxy respondents, suggesting that the prevalence rates reported in this study may be underestimated.


Findings from this analysis empirically document the relationship between smoking and a selection of socio-demographic characteristics relevant to Inuit in Inuit Nunangat. By identifying who is at risk for smoking and by highlighting co-occurring factors, these findings can help inform culturally appropriate prevention and cessation programs and guide efforts at raising awareness of the complex interplay of factors involved in cigarette use among Inuit men and women in Inuit Nunangat.


This study was sponsored by the First Nations and Inuit Health Branch (FNIHB), formerly with Health Canada, now formally transferred to the new Department of Indigenous Services Canada. The authors thank Inuit Tapiriit Kanatami for their input on data analysis.

Date modified: