Housing and health among Inuit children

by Dafna E. Kohen, Evelyne Bougie and Anne Guèvremont

Inuit children are generally in poorer health than other children in Canada. This disparity has been associated with socioeconomic and household characteristics,Note 1,Note 2 which include housing conditions.

The impact of the physical and psychosocial housing environment on children’s health has long been recognized.Note 3 The physical environment involves safety issues (for example, fire hazards) and biological and chemical exposure (for example, to toxins such as tobacco smoke and mold). The psychosocial environment is related to feelings of security, control, and attachment such as homeownership and housing satisfaction.

Early childhood is particularly important for the study of associations between housing conditions and health. Young children spend much of their time indoors, and early exposures can have long-term health consequences.

According to the Canada Mortgage and Housing Corporation, "acceptable" housing should be suitable (not crowded), adequate (not in need of major repair), and affordable (cost less than 30% of before-tax household income).Note 4 According to results of the 2006 Census,Note 5 substantial numbers of Inuit lived in housing that did not meet one or more of these criteria.Note 6,Note 7 For example, 31% lived in dwellings with more than one person per room, and 28% were in dwellings that needed major repair; the corresponding figures for non-Aboriginal people were 3% and 7%.

For young children, crowding and housing in need of repair have been related to  conditions such as asthma, respiratory tract infections and influenza,Note 8-16 injuries,Note 16,Note 17 and behaviour  problems such as aggression, conflict,  social withdrawal, psychological distress, poor school adjustment, and lower social competence.Note 14,Note 18-22  Unaffordable housing has been associated with child behaviour problems.Note 19,Note 22

Environmental tobacco smoke has been linked to ear infections, respiratory infections, asthma, bronchitis, pneumonia, and influenza .Note 9-11,Note 15,Note 23-27 This is notable because the prevalence of daily smoking among Inuit is three times that of adults in Canada.Note 28 As a result, Inuit children are more likely than non-Aboriginal children to be exposed to second-hand smoke.Note 29

Housing satisfaction has been related to better parental ratings of child health.Note 1 Finally, homeownership has been associated with fewer behaviour problems.Note 30,Note 31

However, many of the studies that reported associations between physical and psychosocial housing conditions and child health were based on small samples or on one community, focused on a single housing factor in relation to a single health outcome, or did not account for socioeconomic factors  such as income and education. By contrast, the present study examines several physical and mental health outcomes in a population-based sample of Inuit children aged 2 to 5 in relation to a number of physical and psychosocial housing conditions, controlling for demographic and socioeconomic characteristics.


Data source

The data are from the 2006 Aboriginal Children’s Survey (ACS), a national survey of Aboriginal children (First Nations off reserve, Métis, and Inuit) younger than age 6.Note 32 Because the frame for the ACS was obtained from the census roster, housing characteristics from the 2006 Census were included in the ACS dataset.

The study sample consisted of 1,233 children aged 2 to 5, who were identified by parents or guardians as Inuit (single identity or in combination with another Aboriginal identity). Three-quarters of these children lived in Inuit Nunangat, the collective name for the aggregate of the four Inuit regions created through the signing of land claim agreements:  49% were in Nunavut; 20% in Nunavik;  3% in Nunatsiavut; and  4% in Inuvialuit. The remaining 24% lived outside Inuit Nunangat.  For comparisons with non-Aboriginal children, data from the 2006 Census and the 2006/2007 National Longitudinal Survey of Children and Youth (NLSCY) were used.

Physical housing  characteristics

Three housing characteristics were derived from the 2006 Census.  Crowding was defined as more than one person per room (results of analyses using a 1.5-person-per-room cut-off were similar and are available by request). Dwelling in need of major repair was based on the question: “Is this dwelling in need of any repairs? (no, regular maintenance, minor repairs, major repairs).” Unaffordable housing was defined as reported housing costs greater than 30% of before-tax household income.Note 4 From the ACS, regular exposure to smoke in the home was determined based on responses to the question, “Including both household members and regular visitors, does anyone smoke inside your home every day or almost every day (yes/no)?”

Psychosocial housing characteristics

Two psychosocial housing characteristics were considered: homeownership and satisfaction with housing. Homeownership (“Is this dwelling owned by you or a member of this household, even if it is still being paid for?”) was derived from the 2006 Census. Parental satisfaction with housing conditions (“How would you rate your feelings about your housing conditions?:very satisfied, satisfied, dissatisfied, very dissatisfied”) was derived from the ACS.

The analysis accounted for five socioeconomic and demographic variables:   parent’s/guardian’s highest level of education, household income (adjusted for household size), area of residence (in or outside Inuit Nunangat), and child sex and age.

Physical child health outcomes were: parent-/guardian-rated child health status (excellent/very good versus good/fair/ poor); two or more ear infections in past year; activity limitation (yes/no); two or more chronic conditions (yes/no); and chronic respiratory conditions (allergies, asthma, and/or bronchitis).

Mental/Behavioural outcomes were measured by four subscales (prosocial behaviour, hyperactivity-inattention, emotional symptoms, and conduct problems) from the Strengths and Difficulties Questionnaire (SDQ),Note 33 which has been validated for the ACS sample.Note 34


Descriptive analyses of housing, socioeconomic and demographic characteristics were examined. Separate regression analyses were conducted for each child health outcome. Two models were constructed. Model 1 included the physical housing indicators. Model 2 added psychosocial housing, socioeconomic and demographic characteristics. (Results of a model incorporating psychosocial housing characteristics  prior to inclusion of socioeconomic and demographic controls are available by request.) Multicollinearity among the predictors was assessed via correlations. Analyses were based on survey data weighted to be representative of the 2006 Census counts of children younger than age 6. The bootstrap method was used to account for the complex sampling design of the survey.


Inuit children aged 2 to 5 were more likely than non-Aboriginal children to live in crowded conditions and in dwellings that needed major repair. Specifically, 36% of Inuit children were in households with more than one person per room, and 29% lived in structures that needed major repair (Table 1). The corresponding figures for non-Aboriginal children were much lower:  7% (crowded) and 8% (need major repair). As well, 25% of Inuit children were in homes where they were regularly exposed to smoke.

However, in 2006, 9% of Inuit children were in homes where housing costs amounted to more than 30% of household income, well below the comparable percentage for non-Aboriginal children: 25%.

One-quarter (24%) of Inuit children lived in a dwelling owned by a member of the household, compared with three-quarters (74%) of non-Aboriginal children. A majority—69%—of parents/guardians of Inuit children were satisfied or very satisfied with their housing conditions.

More than half (58%) of Inuit children had parents/guardians with less than secondary school graduation, compared with 7% of non-Aboriginal children. A quarter (25%) of Inuit children were in single-parent families,  versus 15% of non-Aboriginal children.

Descriptive analyses of housing characteristics by Inuit region are available in Appendix Table A.

Physical housing characteristics

Compared with Inuit children in dwellings that were not crowded, those in crowded households had lower odds of being rated in excellent/very good health, higher odds of at least two ear infections in the past year, and higher emotional symptoms and conduct problem scores (Model 1 in Tables 2 and 3). Housing in need of major repair was associated with lower odds of excellent/very good health. Children in unaffordable housing had higher odds of two or more chronic conditions and chronic respiratory conditions. Regular exposure to smoke in the home was associated with conduct problems.

For physical activity limitations and serious injury in the past year, no housing characteristics demonstrated a significant association, so results are not presented in Table 2, but are available by request.

Psychosocial housing characteristics

When psychosocial housing, socioeconomic and demographic factors were taken into account (Model 2 in Tables 2 and 3), many of the relationships between physical housing conditions and the health of Inuit children disappeared. Associations between crowding and excellent/very good health, ear infections, emotional symptoms, and conduct problems were no longer significant. In fact, a significant negative association between crowding and more chronic conditions emerged—crowding was associated with lower odds of having two or more chronic conditions.

The association between housing in need of repair and parental ratings of child health was not statistically significant in the full model. As well, a significant negative association with emotional symptoms was apparent—children in dwellings that needed major repair tended to have lower emotional symptoms scores.

In the full model, regular exposure to smoke in the home remained significantly associated with conduct problems. Unexpectedly regular exposure to smoke was associated with higher odds of a child being rated in excellent/very good health.

Homeownership was associated with higher odds of excellent/very good health and lower scores of emotional symptoms and conduct problems.

Parental satisfaction with housing was associated with high odds of excellent/very good health; low odds of ear infections, chronic conditions, and respiratory conditions; and lower inattention-hyperactivity and emotional symptoms scores. However, additional analyses showed that parental housing satisfaction was closely tied to physical housing conditions (less crowding, not in need of major repair) and homeownership (Table 4).


This study reveals associations between physical and psychosocial aspects of housing and Inuit children’s health. Consistent with the literature,Note 6,Note 7,Note 35 ACS data showed the housing situation of Inuit children to be less favourable than that of non-Aboriginal children in Canada. When crowding, need for repair, and housing affordability were considered alone, the analysis suggested that they were associated with negative physical and mental health outcomes for Inuit children.

Unaffordable housing may have an indirect link to child health, in that shelter costs may reduce resources available for health care, adequate nutrition, and home safety.Note 19,Note 22 As well, chronic financial pressure may affect parenting behaviours and parental stress, and ultimately, influence child outcomes.Note 19,Note 22,Note 36 In fact, when the psychosocial housing, socioeconomic and demographic factors were taken into account, associations between affordability and chronic conditions were no longer significant.

Similarly, when homeownership and housing satisfaction were considered, crowding was no longer significantly associated with less favourable health ratings, ear infections and emotional and conduct problems; and housing in need of major repair was no longer associated with less favourable health ratings.

Relationships between parental housing satisfaction and health conditions persisted over and above socioeconomic and demographic characteristics. Parental housing satisfaction was related to higher odds of excellent/very good health and lower odds of ear infections, and respiratory conditions. Homeownership remained significantly associated with positive health ratings.

The lower odds of chronic conditions among children in crowded households could indicate limited access to health care services,Note 35 and therefore, undiagnosed disease,Note 23 in northern areas where crowding is most prevalent. However, the pattern was similar outside Inuit Nunangat (results not shown). More nuanced definitions of concepts might assist in the interpretation of these findings. For instance, a multi-generational household might be described as “crowded,” but such living arrangements can offer advantages.Note 37,Note 38

The relatively low rate of exposure to indoor smoke (given the prevalence of daily smoking) in this study is similar to results of other researchNote 39 and may indicate successful education and prevention programs.Note 40,Note 41 Moreover, regular exposure to indoor smoke was linked to favourable ratings of child health. Although unexpected, such findings have been reported previously.Note 10,Note 39 Kovesi et al.Note 10 attributed the absence of associations between indoor smoking and respiratory tract infections among Inuit children to a lack of variability in smoking behaviour, since smokers were present in most homes. Egeland et al.Note 39 suggested that despite the high prevalence of smoking among Inuit, public health messages aimed at reducing indoor smoking may have resulted in less exposure, and thus, fewer links with child health. Indeed, the ACS specifically asked about indoor smoking.  It is also possible that diagnosis of respiratory problems is less common among Inuit, and so may have been underreported. Nonetheless, for children exposed to smoke indoors, an association with conduct problems remained statistically significant even when psychosocial housing and sociodemographic factors were taken into account.

In the full model, homeownership and housing satisfaction were important for both the physical and mental health of Inuit children. However, as in other studies,Note 42,Note 43 housing satisfaction was linked to physical housing conditions. Consequently, the condition of housing may indirectly influence child health through associations with psychosocial factors.

Consistent with earlier research,Note 44 the positive effects of homeownership remained for parent-rated child health, emotional symptoms, and conduct problems. Homeownership, in turn, was related to housing quality—homeowners were less likely than tenants to live in crowded conditions or in dwellings that needed major repair (data not shown). Homeownership has also been associated with higher-quality, stimulating, and safe environments, more parental supervision, parental self-esteem, emotional support in the home,Note 31,Note 42 community investments, and geographic and school stability.Note 42 These factors could contribute to associations between homeownership and positive child outcomes.


The results of this analysis should be interpreted in the context of several limitations. The associations between housing characteristics and reported child outcomes are correlational; causality cannot be inferred. Data from the 2006 Aboriginal Children's Survey reflect conditions 10 years ago and not necessarily the most recent housing and sociodemographic conditions.

The limitations also include shared method variance—variance attributable to the measurement method (the survey) rather than to the constructs (housing conditions) the measures are assumed to represent. This is a concern when self-reported data are collected at the same time from the same participants, particularly when the variables are perceptual measures.  For this analysis, all child outcomes were based on parental reports. However, three of the physical housing characteristics (crowding, affordability, need for repair) were drawn from the Census, and the others  were from the Aboriginal Children’s Survey. All were self-reported, but not at the same time or necessarily by the same person. Although this does not rule out the possibility of reporting bias or other confounding influences, correlations of housing characteristics, socioeconomic characteristics, and child outcomes were low to moderate.

The findings warrant replication and point to the need for longitudinal data to examine associations over time and for objective data such as direct measures of indoor air quality to be considered.


This study reveals associations between physical and psychosocial housing characteristics and Inuit children’s health outcomes. Future research would benefit from an understanding of the processes through which housing conditions affect child health. For example, do satisfaction and homeownership affect children via their associations with parental mental health, parental stress, and/or parenting behaviours?Note 45  Inclusion of additional measures, such as objective assessments of housing conditions and standardized or non-parent reported outcomes for children, would benefit future analyses.


This study was funded by Aboriginal Affairs and Northern Development Canada. The authors thank Benita Tam for assistance with manuscript preparation.

Date modified: