You are here

  • Home
  • Online First
  • Neighbourhood chronic stress and gender inequalities in hypertension among Canadian adults: a multilevel analysis

Article Text

Article menu

other Versions

Download PDFPDF

Research report
Neighbourhood chronic stress and gender inequalities in hypertension among Canadian adults: a multilevel analysis
  1. Flora I Matheson1,2,
  2. Heather L White1,3,
  3. Rahim Moineddin4,5,
  4. James R Dunn1,6,
  5. Richard H Glazier1,2,3,4,5
  1. 1Centre for Research on Inner City Health, The Keenan Research Centre in the Li Ka Shing Knowledge Institute of St. Michael's Hospital, Toronto, Ontario, Canada
  2. 2Department of Public Health Sciences, University of Toronto, Ontario, Canada
  3. 3Department of Health Policy, Management and Evaluation, University of Toronto, Ontario, Canada
  4. 4Department of Family and Community Medicine, University of Toronto, Ontario, Canada
  5. 5The Institute for Clinical Evaluative Sciences, Toronto, Ontario, Canada
  6. 6Department of Geography, University of Toronto, Ontario, Canada
  1. Correspondence to Heather L White, Department of Health Policy, Management and Evaluation, Faculty of Medicine, University of Toronto, 506-169 St. George Street, Toronto, ON M5R 2M4, Canada; heatherlynn.white{at}utoronto.ca

Abstract

Background A growing body of literature regarding the contextual influences of ‘place’ effects on health increasingly demonstrates that living in neighbourhoods with high levels of deprivation is associated with worse cardiovascular outcomes; however, little research has explored whether neighbourhood deprivation has a differential impact on the cardiovascular health of men and women. The purpose of this study was to explore gender differences in the association between neighbourhood deprivation and the prevalence of hypertension among non-institutionalised Canadian adults.

Methods Individual-level data from the Canadian Community Health Survey (2000–2005) were combined with area-level data from the 2001 Canada Census to assess the relationship between gender, neighbourhood deprivation and hypertension using multilevel regression.

Results Of the 103 419 respondents, 20 705 reported having hypertension (17.6%). In multilevel models, neighbourhood deprivation was significantly associated with hypertension and this effect remained significant after adjusting for individual-level demographic, socioeconomic and lifestyle characteristics (OR 1.12, 95% CI 1.10 to 1.15). Neighbourhood deprivation appears to be a stronger predictor of hypertension among women, such that women living in areas of high deprivation were 10% more likely to report having hypertension in comparison with men living in the same neighbourhoods and with women living in the least impoverished neighbourhoods.

Conclusions Although future research is needed to determine whether interventions at the area-level are effective in reducing inequalities in health outcomes across neighbourhoods, policies aimed at reducing area-level deprivation may have a differential benefit on the cardiovascular health of men and women.

  • Cardiovascular diseases
  • gender
  • hypertension
  • residence characteristics
  • socioeconomic factors
  • deprivation
  • gender inequalities
  • multilevel modelling
  • social inequalities
  • urban health

https://doi.org/10.1136/jech.2008.083303

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    In recent years, there has been an explosion of interest in the public health and sociological literature regarding the ways in which neighbourhoods and residential environments may have an effect on a variety of health outcomes, including cardiovascular disease (CVD). Although the idea that ‘place’ matters for individual health is not new, there has been relatively little investigation into gender differences in contextual effects, examining whether the specific aspects of environments that are important for health vary according to a resident's sex.

    Numerous studies have now demonstrated a relationship between CVD risk and social class, reporting an increase in CVD incidence, prevalence and mortality with decreasing social class (see reviews in1–3) Simultaneously, a growing body of contextual literature has further demonstrated a relationship between neighbourhood deprivation and CVD, suggesting that neighbourhood socioeconomic characteristics may have an important influence on cardiovascular health over and above the impact of individual-level determinants.1 4–6 Interpretation of these findings often attributes the increased risk of CVD among individuals living in deprived neighbourhoods to two primary sources: 1. increased exposure to health-damaging environments, including high levels of neighbourhood poverty, crime, delinquency and low job availability,6–11 and 2. decreased access to health-promoting goods and services including health and social resources, healthy and affordable foods, safe places to exercise and convenient transportation.6 12 13 Several contextual studies have suggested that neighbourhood deprivation may be a stronger predictor of CVD among women;14–17 however, no research to date has examined the interaction between gender and cardiovascular inequality, exploring whether neighbourhood deprivation has a differential impact on the cardiovascular health of men and women.

    Although there is a general acknowledgement of the importance of examining gender in the context of health, most contextual studies published to date simply adjusted for gender, assuming that the influence of neighbourhood deprivation is equivalent for both men and women. Consequently, studies that integrate structural inequalities and gender differences in health are rare, especially those examining whether gender differences vary by socioeconomic position. Recognising that gender and socioeconomic status have the potential to frame the structure of our lives as well as our access to personal, material, social and psychological resources, it is important to examine how they interact to influence cardiovascular outcomes.

    Hypertension is a well-known risk factor for CVD and when uncontrolled can lead to stroke, heart attack and/or kidney failure.18 Although individual-level effects of social class on blood pressure are known to exist,1 19 contextual studies have further suggested an association between neighbourhood deprivation and increases in blood pressure.15 20–29 Although this literature is far from conclusive regarding the specific pathways linking neighbourhood effects to hypertensive risk, both physical and social features of neighbourhoods may be relevant.6 25 26 The accessibility and safety of recreational spaces and the presence of sidewalks/street lighting may influence participation in leisure-time physical activity whereas the availability and cost of healthy foods, tobacco and liquor products may affect dietary, smoking and drinking patterns.6 12 13 30 In addition, neighbourhood aesthetic quality (eg, the presence of green space and cleanliness) can contribute to a resident's experience of stress and subsequent ability to cope with stressors.6 31 Finally, features of the social environment are also recognised as important determinates of hypertensive risk.26 32 According to social disorganisation theory, disadvantaged neighbourhoods that lack social and economic resources predispose their residents to high levels of physical and social disorder and low levels of informal social support.33 Although residential stability and neighbourhood design can foster the development of trusting relationships and social interactions with neighbours, contextual features of deprived neighbourhoods including crime, crowding, population density, high residential turnover and the low availability or lack of safe and affordable housing can isolate residents, weakening their capacity for collective action and reinforcing the perception that their local environment is dangerous, threatening and stressful.7–9 34 35 Given that stressful experiences are often interpreted and mediated through gender differences in socialisation, it might be expected that neighbourhood deprivation has a stronger impact on the cardiovascular health of women, primarily because high material deprivation and associated residential instability are likely to interfere with the maintenance of social networks and supports, both of which are known to be protective for women's cardiovascular health.34–37

    The present investigation expands on previous research by examining the prevalence of hypertension among men and women living in 3663 urban neighbourhoods in Canada. Using multilevel modelling, the aim was to understand how neighbourhood deprivation affects hypertension prevalence after adjusting for known individual-level risk factors. Further, the interaction between gender and neighbourhood deprivation was tested to examine whether neighbourhood deprivation has a differential impact on the cardiovascular health of men and women.

    Hypotheses:

    1. Higher levels of neighbourhood deprivation are associated with an increased risk of reporting hypertension.

    2. The association between neighbourhood deprivation and hypertension differs by gender such that women living in neighbourhoods with higher levels of deprivation report an increased prevalence of hypertension relative to men living in the same neighbourhoods, as well as to women living in neighbourhoods with lower levels of deprivation.

    3. Gender differences in risk will remain statistically significant after controlling for individual-level demographic, socioeconomic and lifestyle characteristics.

    Methods

    Data sources

    Data used in this investigation were obtained from two sources, both of which are collected and maintained by Statistics Canada. Individual-level data were obtained from the Canadian Community Health Survey (CCHS), a nationally representative cross-sectional survey designed to collect information on health determinants, outcomes and health service utilisation among non-institutionalised Canadian adults.38 Data from three waves of the CCHS were combined (2001, 2003 and 2005 surveys) to create a large dataset of 400 055 respondents representative of the general population of Canada. The combined sample was restricted to include respondents between the ages of 30 and 84 years residing in Canada's 25 Census Metropolitan Areas (CMAs) at the time of data collection (n=117 086). Women who reported being pregnant at the time of the survey were excluded from the analysis in order to reduce potential selection bias in cases where hypertension may be attributed to pre-eclampsia (n=809).

    At the neighbourhood-level, survey responses were linked to the 2001 Canada Census by census tract (n=3663) using the Postal Code Conversion File Plus available through Statistics Canada. The neighbourhood deprivation index used in this investigation was developed based on previous area-based research and theory linking neighbourhood deprivation with poor health39–43 and has been published in detail elsewhere.40 Eighteen measures were initially extracted from the 2001 census data describing the socioeconomic and demographic character of the included neighbourhoods. To avoid the problem of multicollinearity that would result from analysing correlated census measures simultaneously in a regression model, weighted factor analysis using oblique rotation was conducted where the weight assigned to each indicator was determined from the statistical relationships that existed among each of the indicators within the 25 CMAs. The following six census measures expressed as proportions, were eventually selected to comprise the deprivation index (eigenvalue 2.8): population aged 20+ without high school graduation, lone parent families, families receiving government transfer payments, the population aged 15+ who are unemployed, the population living below the low income cut-off (adjusted for community size, family size and inflation) and homes needing major repairs. Factor loadings ranged from 0.51 to 0.91 with a Cronbach's α of 0.92. The deprivation index represents a standardised continuous score ranging from a score of −2, which represents a neighbourhood with low deprivation, to a score of +6, which represents a neighbourhood with high deprivation. The 25 CMAs were further categorised by population size to account for potential diversity in lifestyles in the larger metropolitan areas relative to smaller cities: large cities (population 500 000 or greater), mid-size cities (200 000–499 999) and small cities (10 000–199 999).

    Respondents were classified as having hypertension if they reported having been told by a health professional that they had high blood pressure (no vs yes). Results from a validation study conducted in the USA indicated that sensitivity and specificity for this method of self-reported hypertension was over 80%.44 Individual-level variables reflecting underlying demographic, socioeconomic and lifestyle characteristics were included to adjust for their known relationship with hypertension.1 19 Marital status was categorised into respondents who were married or living common law compared with respondents who were single and divorced, separated or widowed. Ethnicity, based on self-report, was dichotomised to include Caucasian respondents versus other, and the presence of dependent children was dichotomised to include respondents with children under the age of 25 living in the household versus those without children or whose children were aged 25 and older. Although it is recognised that the highest burden of domestic work is most often attributable to the presence of younger children in the household, the CCHS does not provide disaggregated measures of household composition. Consequently, it was not possible to create more refined categories for the age of children living at home. Respondents were grouped into the following categories based on their highest level of education attainment: postsecondary graduation, some postsecondary education, high school graduation and some high school education. Annual household income was measured in the following categories: $0–$19 999, $20 000–$39 999, $40 000–$59 999, $60 000 to $79 999 and $80 000 or greater. As a large number of respondents chose not to report their annual income, a dummy variable was created to incorporate missing respondents into the analysis. Lifestyle factors associated with increased blood pressure include current smokers (no vs yes), high self-perceived stress (no vs yes), being overweight, defined as having a body mass index greater than 25 (no vs yes), and level of drinking risk (non-drinker: 0 drinks/week; low-risk drinker: males: 1–14 drinks/week; females: 1–9 drinks/week; high-risk drinker: males: ≥15 drinks/week; females >10 drinks/week). To derive leisure-time physical activity, respondents' energy expenditure (EE) was calculated based on the frequency and duration of each activity they engaged in during their leisure time within the 3 months prior to being surveyed and its value of metabolic energy cost, expressed as a multiple of resting metabolic rate (MET) according to the approach adopted by the Canadian Fitness and Lifestyle Research Institute.45 Using the categories defined and reported by Statistics Canada, respondents whose estimated leisure-time EE was below 1.5 kcal/kg/day were considered to be physically inactive; respondents with EE values between 1.5 kcal/kg/day and 2.9 kcal/kg/day were considered to be moderately active, whereas respondents with EE values of 3 kcal/kg/day or greater were considered to be active.45 As this measure for leisure-time activity does not account for energy expenditure while at work or while doing household chores, an additional measure for usual daily activity level was also included (usually sit, stand or walk quite a lot, usually lift or carry light loads, and do heavy work or carry very heavy loads). To assess whether findings could be explained by underlying gender differences in health seeking behaviour, a variable was included on whether respondents sought care from a family or medical doctor within the previous 12 months (yes vs no). Records with missing values (with the exception of income) were removed (n=12 613) leaving a final sample of 103 419 respondents. There were no significant differences between removed cases and those retained for analysis. Ethics approval was obtained from the Research Ethics Boards of St. Michael's Hospital and Statistics Canada data publication guidelines were followed throughout the analysis.

    Statistical analyses

    Descriptive statistics were calculated to describe the sample population stratified by gender. To allow for the hierarchical clustering of individual respondents within neighbourhoods, multilevel modelling was used to conduct maximum likelihood logistic regression to model the likelihood of reporting hypertension. A multilevel logistic model with one explanatory variable at level one (individual-level) and one explanatory variable at level two (area-level) can be written as follows:

    logit(pij)=π0j+π1jxijπ0j=γ00+γ01zj+u0jπ1j=γ10+γ11zj+u1j

    where [u0ju1j]=N([00],[σ02σ01σ01σ12]).

    Here, Pij is the probability that individual i in area j will experience the outcome, xij is an explanatory variable on the individual-level, and zj is an explanatory variable at the area-level. This model can be written in the following single equation: logit(pij)=γ00+γ10xij+γ01zj+γ11xijzj+u0j+u1jxij. The segment γ00+γ10xij+γ01zj+γ11xijzj represents the fixed effects and the segment u0j+u1jxij represents the random effects of the model. The cross-level interaction term can be represented by γ11 zj xij in which the coefficient γ11 shows how π1j, the slope at the individual-level, varies with zj, the area-level variable. Finally, the intraclass correlation coefficient (ICC) can be calculated as ρ=σμ2/(σμ2+π2/3), where σμ2 is the variance of the random intercept in a fully unconditional logistic model. The ICC provides an estimate of the total variance in hypertension prevalence that can be attributed to differences between neighbourhoods.

    Five models were fit using random intercepts to allow the prevalence of hypertension to vary randomly across neighbourhoods. First, an unconditional logistic regression model was fit to determine the amount of variation in hypertension that could be attributed to the neighbourhood-level (ICC). This was then followed with two models introducing the contextual effect of neighbourhood deprivation (model 1) and a cross-level interaction between gender and neighbourhood deprivation (model 2). The individual-level effects of demographic (model 3), socioeconomic (model 4) and lifestyle characteristics (model 5) were successively added. Finally, a variable for physician consultation within the previous 12 months and its interaction with gender was added to model 5 to rule out whether the findings reflect gender differences in health-seeking behaviour. Fixed effect coefficients and random variances are presented for each model and final estimates are adjusted for survey cycle and city size. Confidence intervals (95% CI) are reported throughout the manuscript.

    To depict the spatial clustering of neighbourhood deprivation and hypertension rates by census tract, bivariate choropleth maps were constructed of three select cities representing eastern (Halifax), central (Toronto) and western (Vancouver) Canada. Choropleth maps provide an easy way to visualise the magnitude of a variable within defined geographic boundaries and are particularly useful for comparing how measures vary across regions and/or population subgroups. Hypertension rates for each census tract were generated using prevalence estimates from the unconditional logistic regression model. For the purposes of mapping, hypertension rates and neighbourhood deprivation levels are depicted as class combinations based on 1st and 3rd quartile values. ArcMap46 and MapInfo47 GIS software were used to generate the maps.

    Results

    Descriptive statistics describing the sample population are presented in table 1. To account for unequal probabilities of selection, non-response and to reflect age and gender distributions of the Canadian population at the time of the survey, descriptives are presented as unweighted sample frequencies with weighted population proportions. The sample was roughly split by gender, comprising 46 633 men (48.8%) and 56 786 women (51.2%) distributed in 3663 neighbourhoods. Overall, the prevalence of reporting hypertension among respondents was 17.6% (n=20 705), with 17.0% of men (n=8597) and 18.2% of women (n=12 108) reporting a medical diagnosis of hypertension. In bivariate analysis, women were 20% more likely to report having hypertension when compared to male respondents (OR 1.20, 95% CI 1.16 to 1.24, data not shown).

    View this table:
    Table 1

    Descriptive statistics of adult respondents aged 30–84 stratified by gender, Canada 2000–2005 (n=103 419)

    First, an unconditional logistic regression model was considered in order to decompose the variance in hypertension across the two levels of analysis. According to this empty model, the estimate of the total variance in hypertension that occurs between neighbourhoods was 0.07 (p<0.001; data not shown), yielding an ICC of 0.021. This suggests that on average, 2.1% of the variance in self-reported hypertension can be attributed to neighbourhood-level contextual effects. In interpreting the value of the ICC, it is important to keep the following points in mind: although most of the variation in hypertension occurs at the individual-level, the magnitude of the ICC is consistent with previous multilevel investigations of CVD14 16 22 24 31 and is important given that the size of the ICC does not rule out the relatively large effects of neighbourhood-level measures and their ability to impact the lives of individuals living within them.48

    Table 2 presents the results of a series of multilevel logistic regression models depicting the adjusted likelihood of reporting hypertension. As shown in model 1, neighbourhood deprivation was significantly associated with self-reported hypertension such that respondents were 12% more likely to report a diagnosis of hypertension with each unit increase in neighbourhood deprivation (OR 1.12, 95% CI 1.10 to 1.15). For areas grouped by rank of area-level deprivation, the average deprivation scores for the first and third quartiles are −0.7 and 0.5 respectively; therefore, a one-unit increase in neighbourhood deprivation is roughly equivalent to comparing a neighbourhood at the lower 25th percentile with a neighbourhood at the upper 75th percentile. This finding confirms the first hypothesis of the present study, indicating that higher levels of neighbourhood deprivation are associated with an increased risk of reporting hypertension. Model 2, which includes the cross-level interaction for gender and neighbourhood deprivation indicates that the average difference in hypertension prevalence between males and females depends on the extent of deprivation present within the neighbourhood, with each unit increase in neighbourhood deprivation associated with a 16% increased likelihood of reporting hypertension among women (OR 1.16, 95% CI 1.12 to 1.20). This finding supports the second hypothesis of the present study, suggesting that women living in areas of higher deprivation are more likely to report having hypertension when compared with men living in the same neighbourhoods as well as with women living in the least deprived neighbourhoods. Adjusting for individual-level demographic and socioeconomic indicators in models 3 and 4 slightly attenuates the cross-level interaction among gender, neighbourhood deprivation and hypertension (OR 1.12, 95% CI 1.08 to 1.16). As anticipated, there is a strong gradient between increasing age and the likelihood of reporting hypertension; however, in stratifying the final model by age, no significant differences were found in the magnitude of the interaction between neighbourhood deprivation and gender for respondents under the age of 64 and those aged 65 and older (OR 1.07, 1.08 respectively, data not shown).

    View this table:
    Table 2

    Multilevel logistic regression models: ORs of reporting hypertension by neighbourhood deprivation, gender, sociodemographic and lifestyle characteristics, Canada 2000–2005 (n=103 419)

    The full model (model 5), which includes lifestyle factors and health behaviours associated with hypertension, suggests that the inclusion of lifestyle characteristics further attenuates the cross-level interaction among gender, neighbourhood deprivation and hypertension (OR 1.07, 95% CI 1.03 to 1.11), Despite this reduction in risk, the coefficient remains statistically significant (p<0.001), providing support for the third hypothesis of the present study that gender differences in the contextual effects of neighbourhood deprivation on hypertensive risk remain statistically significant despite controlling for individual-level demographic, socioeconomic and lifestyle characteristics.

    To assess whether these findings could be explained by underlying gender differences in health-seeking behaviour, physician consultation and its interaction with gender were added to model 5. Although consultation with a family or medical doctor within the previous 12 months was associated with a lower likelihood of reporting hypertension (OR 0.32, 95% CI 0.30 to 0.35), the interaction term with gender was not significant (OR 1.07, 95% CI 0.92 to 1.08). More importantly, the inclusion of these variables did not alter the magnitude of the cross-level interaction between neighbourhood deprivation and gender (OR 1.07, 95% CI 1.03 to 1.11). Taken together, these findings suggest that although accessing healthcare is an important determinant of self-reported hypertension prevalence, gender differences in care-seeking behaviour do not account for gender inequalities in the neighbourhood effect. In addition to neighbourhood deprivation, the presence of dependent children in the household was also associated with gender differences in hypertension prevalence such that women with dependent children were 25% less likely to report having a diagnosis of hypertension (OR 0.75, 95% CI 0.68 to 0.82). Additional interaction terms with both age and gender were tested and found to be non-significant for the remaining individual-level variables.

    Figure 1 allows us to explore the cross-level interaction between gender, neighbourhood deprivation and hypertension in more depth. As seen in the figure, there appears to be little variation in the probability of reporting hypertension among male respondents regardless of where they live. In contrast, female respondents living in the most deprived neighbourhoods had more than a 10% higher probability of reporting hypertension when compared with women living in the least deprived neighbourhoods. In addition, these findings further point to a 10% greater likelihood of women living in the most deprived neighbourhoods to report a diagnosis of hypertension in comparison to men living in the same neighbourhoods.

    Figure 1
    Figure 1

    *Adjusted probability of reporting hypertension by neighbourhood deprivation and gender for respondents aged 30–84, Canada 2000–2005 (n=103 419). *Adjusted estimates obtained from the full model (model 5). †Material deprivation index represents a standardised measure constructed from factor loadings.

    Figure 2 displays the choropleth maps of neighbourhood deprivation and hypertension rates in Halifax, Toronto and Vancouver by census tract. Looking at figure 2, a strong clustering effect is seen between neighbourhoods with high levels of deprivation and high rates of hypertension among female respondents, which remains consistent across the three cities. In many high-deprivation neighbourhoods where males reported low rates of hypertension, females reported some of the highest rates of hypertension; triangulating the findings of the regression analysis that neighbourhood deprivation exerts a differential impact on the hypertension prevalence of men and women. These maps also identify several low-deprivation neighbourhoods with high hypertension rates among female respondents. Although these findings may in part be explained by the overall higher prevalence of hypertension among female populations, it remains an area for further inquiry.

    Figure 2
    Figure 2

    Clustering of neighbourhood deprivation and hypertension rates in Halifax, Toronto and Vancouver municipalities by census tract, stratified by gender, Canada 2000–2005.

    Discussion

    The findings of this investigation suggest that the contextual effects of neighbourhood deprivation exert a significant influence on hypertensive risk and that this influence remains significant despite the inclusion of individual-level demographic, socioeconomic and lifestyle characteristics. Furthermore, the results suggest that neighbourhood deprivation is likely to exert a stronger influence on the cardiovascular health of women given that women residing in the most impoverished areas were on average, 10% more likely to report having hypertension when compared with men living in the same neighbourhoods. These findings add to the growing body of research that suggests that the places in which we live can be important determinants in the pathways leading to cardiovascular health and disease.

    This study has produced novel results regarding gender differences in the relationship between neighbourhood deprivation and hypertension and is the first study to explicitly consider individual-level characteristics, which could potential account for this gender inequality. To the authors' knowledge, this is the first multilevel study in a Canadian population that has sought to investigate the contextual effects of neighbourhood deprivation and the individual-level effects of demographic, socioeconomic and lifestyle characteristics on hypertension prevalence.

    The magnitude of the neighbourhood—gender interaction is in agreement with previous results presented by Winkleby et al, who reported a 7% higher incidence of CVD among women in comparison to men living in high-deprivation neighbourhoods,17 and although a 10% increase in risk may appear small, its influence at the population-level is not trivial. Using the present study's conservative prevalence rate for hypertension of 18% among urban Canadian adults, a 10% increase in prevalence translates to approximately 460 000 additional women living with high blood pressure, based on population estimates from the 2006 Canada Census.

    Despite these contributions, several limitations of the study should be noted. The prevalence of hypertension was based on self-reported diagnosis and therefore may not reflect the true prevalence of hypertension in the Canadian population. Given the high proportion of individuals living with undiagnosed hypertension, the estimates presented are likely to be conservative in nature and further work is needed to confirm these findings using more stringent clinical diagnoses. Numerous investigations have shown that individual-level socioeconomic inequalities are linked to an increased risk of cardiovascular disease.1–3 Despite the fact that the socioeconomic gradients for educational attainment and household income reported in these findings are small, both variables remained significant predictors of hypertension over and above the impact of neighbourhood-level deprivation and their effect size is consistent with previous multilevel studies of neighbourhood socioeconomic characteristics and hypertension.24 29 Although it is possible that the present findings reflect some degree of gender differences in reporting behaviour, including variables relating to healthcare-seeking behaviour did not account for the increased prevalence of hypertension among women living in deprived neighbourhoods. Moreover, in a previous analysis exploring neighbourhood deprivation and gender inequalities in self-reported health, no significant differences were found in the way men and women perceived their health across neighbourhoods (unpublished; available from author upon request). This analysis was restricted to urban centres as census tract characteristics are not collected for rural areas of Canada. Contextual effects are likely to differ for these areas in relation to the availability of health services, access to public transportation and public space, social resources and support networks and methods are continually sought for investigating the impact of neighbourhood deprivation on the health of rural populations. Finally, as with all cross-sectional studies regarding neighbourhood effects and individual health outcomes, it is not possible to get a clear picture of the extent to which neighbourhood characteristics actually cause the health problem under investigation. Although the impact of deprivation on hypertension may imply a direct harmful effect on risk, in actuality this may stem from intermediary causes including decreased access to healthy resources within the local community, leading to subsequent increases in harmful risk factors observable at the individual-level (ie, higher BMI and physical inactivity).

    Given growing interest in exploring ecological determinants of health, it is important that researchers specially examine gender differences rather than assume that health outcomes and their potential risks are equivalent for both men and women. The present study is suggestive regarding the contextual effect of neighbourhood deprivation on hypertension for women and although physical features of deprived neighbourhoods, reflected through physical inactivity, diet and smoking covariates at the individual-level act as partial mediators, none of these were able to fully explain the nature of this inequality. It is believed that these findings add support to the growing body of literature theorising that social features of neighbourhoods and neighbourhood stressors may be more detrimental for the heart health of women. Neighbouring is considered to be an important indicator of the health and strength of urban communities and as such, neighbours are recognised as an important source of social networks and support (see reviews in49 50) However, not all groups neighbour equally and women are typically more active than men within their neighbourhoods.37 51 Although this increase in neighbouring among women is often attributed to increased leisure time or more time spent within their local community,10 37 52 there is some evidence to suggest that gender roles and gendered patterns in socialisation tend to reinforce women's extensive involvement with others; embedding them more wholly into the social connectedness of their neighbourhoods and communities independent of occupational and domestic responsibilities.34 37 37 Although the relative effects of work and household context are not the focus of this study, the models add some evidence to this literature, indicating that the presence of dependent children in the household and marital status did not account for gender inequalities in hypertension prevalence within deprived neighbourhoods. As the CCHS data do not allow investigation of occupational and caregiver status, further work is required to determine whether the issue of double burden can account for the present findings.

    Although the specific pathways through which neighbourhood deprivation may influence cardiovascular health remain unclear, the findings from this study indicate that the association between neighbourhood inequalities and poor cardiovascular health may manifest via gender differences in individual health behaviours and socialisation patterns among men and women. Furthermore, this study highlights the importance of examining cross-level interactions in multilevel analysis, reminding researchers that neighbourhood-level risk factors cannot be assumed to operate identically for population subgroups. Future research is needed to determine whether interventions at the area-level can be effective in reducing inequalities in cardiovascular health across neighbourhoods; however, these findings suggest that policies centred around neighbourhood deprivation may have a differential benefit on the cardiovascular health of men and women. These findings are important for policy-makers as they identify where resources might be better located to improve the cardiovascular health of at-risk populations. In particular, recognising the importance of social networking for women's health, further research is required to understand whether community initiatives fostering supportive social networks and neighbourhood efficacy can improve the heart health for high-risk women.

    What is already known on this subject

    • Several empirical studies in North America and Europe have demonstrated that living in neighbourhoods with high levels of deprivation is associated with worse cardiovascular outcomes.

    • Some of these studies have suggested that neighbourhood deprivation may be a stronger predictor of cardiovascular disease among women;14–17 however, no research has specifically examined the relationship between gender and cardiovascular inequality, exploring whether neighbourhood deprivation has a differential impact on the cardiovascular health of men and women.

    What this study adds

    • The present study demonstrates that the influence of neighbourhood deprivation on hypertension prevalence is both gender-specific and place-specific, such that women living in impoverished neighbourhoods are at greatest risk for poor cardiovascular outcomes.

    • Consequently, neighbourhood policies aimed at reducing area-level deprivation may have a differential benefit on the cardiovascular health of men and women.

    Acknowledgments

    The authors gratefully acknowledge the support of the Ontario Ministry of Health and Long-Term Care. The views expressed in this publication are the views of the authors and do not necessarily reflect the views of the Ontario Ministry of Health and Long-Term Care. The authors also wish to thank J Weyman and P Gozdyra for their work in creating the cloropleth maps.

    References

    1. ↵
      1. Kaplan GA,
      2. Keil JE
      . Socioeconomic factors and cardiovascular disease: a review of the literature. Circulation 1993;88:1973–98.
      OpenUrlAbstract/FREE Full Text
    2. ↵
      1. Marmot MG,
      2. Smith GD,
      3. Stansfeld S,
      4. et al
      . Health inequalities among British civil servants. The Whitehall II Study. Lancet 1991;337:1387–93.
      OpenUrlCrossRefPubMedWeb of Science
    3. ↵
      1. Pickering T
      . Cardiovascular pathways: socioeconomic status and stress effects on hypertension and cardiovascular function. Ann New York Acad Sci 2006;896:262–77.
      OpenUrlCrossRef
    4. ↵
      1. Pickett KE,
      2. Pearl M
      . Multilevel analysis of neighbourhood socioeconomic context and health outcomes: a critical review. J Epidemiol Community Health 2001;55:111–22.
      OpenUrlAbstract/FREE Full Text
    5. ↵
      1. Wilkinson RG,
      2. Pickett KE
      . Income inequality and population health: a review and explanation of the evidence. Soc Sci Med 2006;62:1768–84.
      OpenUrlCrossRefPubMedWeb of Science
    6. ↵
      1. Diez-Roux AV
      . The study of group-level factors in epidemiology: rethinking variables, study designs and analytical approaches. Epidemiol Rev 2004;26:104–11.
      OpenUrlFREE Full Text
    7. ↵
      1. Gillis AR
      . Household density and human crowding: unraveling a non-linear relationship. J Popul 1979;2:104–17.
      OpenUrlWeb of Science
    8. ↵
      1. Boardman JD
      . Stress and physical health: the role of neighbourhoods as mediating and moderating mechanisms. Soc Sci Med 2001;58:2473–83.
      OpenUrlCrossRef
    9. ↵
      1. Gillis AR
      . High-rise housing and psychological strain. J Health Soc Behav 1977;18:418–31.
      OpenUrlCrossRefPubMed
    10. ↵
      1. Stafford M,
      2. Cumming S,
      3. Macintyre S,
      4. et al
      . Gender differences in the associations between health and neighbourhood environment. Soc Sci Med 2005;60:1681–92.
      OpenUrlCrossRefPubMedWeb of Science
    11. ↵
      1. Glazier RH,
      2. Creatore MI,
      3. Gozdyra P,
      4. et al
      . Geographic methods for understanding and responding to disparities in mammography use in Toronto, Canada. J Gen Intern Med 2004;19:952–61.
      OpenUrlCrossRefPubMedWeb of Science
    12. ↵
      1. Kaplan GA,
      2. Pamuk ER,
      3. Lynch JW,
      4. et al
      . Inequality in income and mortality in the United States: analysis of mortality and potential pathways. Br Med J 1996;312:999–1003.
      OpenUrlAbstract/FREE Full Text
    13. ↵
      1. Matheson FI,
      2. Moineddin R,
      3. Glazier RH
      . The weight of place: a multilevel analysis of gender, neighbourhood material deprivation and body mass index among Canadian adults. Soc Sci Med 2008;66:675–90.
      OpenUrlCrossRefPubMedWeb of Science
    14. ↵
      1. Sundquist K,
      2. Malmstrom M,
      3. Johansson SE
      . Neighbourhood deprivation and incidence of coronary heart disease: a multilevel study of 2.6 million women and men in Sweden. J Epidemiol Community Health 2004;58:71–7.
      OpenUrlAbstract/FREE Full Text
    15. ↵
      1. Diez-Roux AV,
      2. Nieto FJ,
      3. Muntaner C,
      4. et al
      . Neighborhood environments and coronary heart disease: a multilevel analysis. Am J Epidemiol 1997;146:48–63.
      OpenUrlAbstract/FREE Full Text
    16. ↵
      1. Stjarne MK,
      2. Diderichsen F,
      3. Reuterwall C,
      4. et al
      . Socioeconomic context in area of living and risk of myocardial infarction: results from Stockholm Heart Epidemiology Program (SHEEP). J Epidemiol Community Health 2002;56:29–35.
      OpenUrlAbstract/FREE Full Text
    17. ↵
      1. Winkleby M,
      2. Sundquist K,
      3. Cubbin C
      . Inequalities in CHD incidence and case fatality by neighbourhood deprivation. Am J Prev Med 2007;32:97–106.
      OpenUrlCrossRefPubMedWeb of Science
    18. ↵
      1. Mackay J,
      2. Mensah GA
      . The atlas of heart disease and stroke. Geneva: World Health Organization, Centers for Disease Control and Prevention, 2004.
    19. ↵
      1. Colhoun HM,
      2. Hemingway H,
      3. Poulter N
      . Socio-economic status and blood pressure: an overview analysis. J Hum Hypertens 1998;12:91–110.
      OpenUrlCrossRefPubMedWeb of Science
    20. ↵
      1. Chaix B,
      2. Ducimetière P,
      3. Lang T,
      4. et al
      . Residential environment and blood pressure in the PRIME study: is the association mediated by body mass index and waist circumference? J Hypertens 2008;26:1078–84.
      OpenUrlCrossRefPubMedWeb of Science
    21. ↵
      1. Smith GD,
      2. Hart C,
      3. Watt G,
      4. et al
      . Individual social class, area-based deprivation, cardiovascular disease risk factors, and mortality: the Renfrew and Paisley study. J Epidemiol Community Health 1998;52:399–405.
      OpenUrlAbstract
    22. ↵
      1. Hart C,
      2. Ecob R,
      3. Smith GD
      . People, places and coronary heart disease risk factors: a multilevel analysis of the Scottish Heart Health Study archive. Soc Sci Med 1997;45:893–902.
      OpenUrlCrossRefPubMedWeb of Science
    23. ↵
      1. Diez-Roux AV,
      2. Link BG,
      3. Northridge ME
      . A multilevel analysis of income inequality and cardiovascular disease risk factors. Soc Sci Med 2000;50:673–87.
      OpenUrlCrossRefPubMedWeb of Science
    24. ↵
      1. Merlo J,
      2. Ostergren P-O,
      3. Hagberg O,
      4. et al
      . Diastolic blood pressure and area of residence: multilevel vs. ecological analysis of social inequity. J Epidemiol Community Health 2001;55:791–8.
      OpenUrlAbstract/FREE Full Text
    25. ↵
      1. Mujahid MS,
      2. Diez Roux AV,
      3. Morenoff JD,
      4. et al
      . Neighborhood characteristics and hypertension. Epidemiology 2008;19:590–8.
      OpenUrlCrossRefPubMedWeb of Science
    26. ↵
      1. Morenoff JD,
      2. House JS,
      3. Hansen BB,
      4. et al
      . Understanding social disparities in hypertension prevalence, awareness, treatment, and control: the role of neighborhood context. Soc Sci Med 2007;65:1853–66.
      OpenUrlCrossRefPubMedWeb of Science
    27. ↵
      1. McGrath JJ,
      2. Matthews KA,
      3. Brady SS
      . Individual versus neighborhood socioeconomic status and race as predictors of adolescent ambulatory blood pressure and heart rate. Soc Sci Med 2006;63:1442–53.
      OpenUrlCrossRefPubMedWeb of Science
    28. ↵
      1. Cozier Y,
      2. Palmer JR,
      3. Horton NJ,
      4. et al
      . Relation between neighborhood median housing value and hypertension risk among black women in the United States. Am J Publ Health 2007;97:718–24.
      OpenUrlCrossRefPubMedWeb of Science
    29. ↵
      1. Diez-Roux AV,
      2. Chambless L,
      3. Merkin S,
      4. et al
      . Socioeconomic disadvantage and changes in blood pressure associated with aging. Circulation 2002;106:703–10.
      OpenUrlAbstract/FREE Full Text
    30. ↵
      1. McNeill LH,
      2. Wyrwich KW,
      3. Brownson RC,
      4. et al
      . Individual, social environmental, and physical environmental influences on physical activity among black and white adults: a structural equation analysis. Ann Behav Med 2006;31:36–44.
      OpenUrlCrossRefPubMedWeb of Science
    31. ↵
      1. Mair C,
      2. Diez-Roux AV,
      3. Shen M,
      4. et al
      . Cross-sectional and longitudinal associations of neighborhood cohesion and stressors with depressive symptoms in the multiethnic study of atherosclerosis. Ann Epidemiol 2009;19:49–57.
      OpenUrlCrossRefPubMedWeb of Science
    32. ↵
      1. Kawachi I,
      2. Kennedy BP
      . Health and social cohesion. Why care about income inequality? Br Med J 1997;314:1037–40.
      OpenUrlAbstract/FREE Full Text
    33. ↵
      1. Sampson R,
      2. Raudenbush S,
      3. Earls F
      . Neighborhoods and violent crime: a multilevel study of collective efficacy. Science 1997;227:918–24.
      OpenUrlCrossRef
    34. ↵
      1. Campbell KE,
      2. Lee BA
      . Sources of personal neighbor networks: social, integration, need or time? Social Forces 1992;70:1077–110.
      OpenUrlAbstract/FREE Full Text
    35. ↵
      1. Magdol L
      . Is moving gendered? The effects of residential mobility on the psychological well-being of men and women. Sex Roles 2002;47:553–60.
      OpenUrlCrossRefWeb of Science
    36. ↵
      1. Denton M,
      2. Walters V
      . Gender differences in structural and behavioral determinates of health: an analysis of the social production of health. Soc Sci Med 1999;48:1221–35.
      OpenUrlCrossRefPubMedWeb of Science
    37. ↵
      1. Campbell KE,
      2. Lee BA
      . Gender difference in urban neighbouring. Socio Q 1990;31:495–512.
      OpenUrlCrossRef
    38. ↵
      1. Beland Y
      . Canadian community health survey-methodological overview. Health Rep 2002;13:9–14.
      OpenUrlPubMed
    39. ↵
      1. Kornhauser RR
      . Social sources of delinquency: an appraisal of analytic models. Chicago: University of Chicago Press, 1978.
    40. ↵
      1. Matheson FI,
      2. Moineddin R,
      3. Dunn JR,
      4. et al
      . Urban neighbourhoods, chronic stress, gender and depression. Soc Sci Med 2006;63:2604–16.
      OpenUrlCrossRefPubMedWeb of Science
    41. ↵
      1. Pampalon R,
      2. Raymond G
      . A deprivation index for health and welfare planning in Quebec. Chron Dis Can 2000;21:103–13.
      OpenUrl
    42. ↵
      1. Shaw CR,
      2. McKay HD
      . Juvenile delinquency and urban areas; a study of rates of delinquency in relation to differential characteristics of local communities in American cities. 2nd edn. Chicago: Chicago University Press, 1969.
    43. ↵
      1. Sampson RJ,
      2. Groves WB
      . Community structure and crime: testing social-disorganization theory. Am J Sociol 1989;94:774–802.
      OpenUrlCrossRefWeb of Science
    44. ↵
      1. Martin LM,
      2. Leff M,
      3. Colonge N,
      4. et al
      . Validation of self-reported chronic conditions and health services in a managed care population. Am J Prev Med 2000;18:215–18.
      OpenUrlCrossRefPubMedWeb of Science
    45. ↵
      Health Statistics Division and Special Surveys Division, Statistics Canada. Canadian Community Health Survey (CCHS) cycle 1.1 derived variable (DV) specifications. Ottawa: Statistics Canada, 2003.
    46. ↵
      1. Duncan GJ,
      2. Raudenbush SW
      . Assessing the effects of context in studies of child and youth development. Educ Psychol 1999;34:29.
      OpenUrlCrossRefWeb of Science
    47. ↵
      ESRI Incorporated. ArcMap Version 9.2, Redlands: 2006.
    48. ↵
      MapInfo Corporation. MapInfo Professional Version 7.0 - Release Build 26, Troy: 2002.
    49. ↵
      1. Olson P
      . Urban neighbourhood research: its development and current focus. Urban Aff Q 1982;17:491–518.
      OpenUrlCrossRef
    50. ↵
      1. Hallman H
      . Neighborhoods: their place in urban life. Beverly Hills: Sage, 1984.
    51. ↵
      1. Schieman S
      . Residential stability and the social impact of neighborhood disadvantage: a study of gender- and race- contingent effects. Social Forces 2005;83:1031–64.
      OpenUrlAbstract/FREE Full Text
    52. ↵
      1. Devine CM,
      2. Jasran M,
      3. Jabs JA,
      4. et al
      . “A lot of sacrifices:” work-family spillover and the food choice coping strategies of low-wage employed parents. Soc Sci Med 2006;63:2591–603.
      OpenUrlCrossRefPubMedWeb of Science

    Footnotes

    • Funding This research was supported by the Social Sciences and Humanities Research Council of Canada [Standard Research Grant Number 410-2005-2306].

    • Competing interests None.

    • Ethics approval This study was conducted with the approval of the St. Michael's Hospital Review Board.

    • Provenance and peer review Not commissioned; externally peer reviewed.