Introduction
Rapid population aging and mass migration to cities are the greatest demographic shifts of our time (Leeson, Reference Leeson2018; Woetzel et al., Reference Woetzel, Remes, Law, Means, Coles and Krishnan2016; World Health Organization, 2007). In Canada, older adults outnumbered children for the first time in 2015 (Statistics Canada, 2015), and population projections indicate municipalities will see an increase in their older adult populations (Hartt & Biglieri, Reference Hartt and Biglieri2018). Overwhelmingly, older adults in Canada and across the world have expressed a desire to age in place (78% in a recent survey conducted in 2020) (March of Dimes Canada, 2021). The desire to age in place is also tightly related to the importance of place attachment in later life and how place attachment can promote familiarity, a sense of security and belonging, as well as the maintenance of positive self-identity, preserving social networks, and maintenance of daily rhythms (Aliakbarzadeh et al., Reference Aliakbarzadeh Arani, Zanjari, Delbari, Foroughan and Ghaedamini Harouni2022; Lewicka, Reference Lewicka2011). The positive well-being benefits of aging in place and place attachment, however, diminish when one’s built environment does not support them. This begs the question: How well do the places we live support the health and well-being of older adults? As over 50 per cent of Canadian older adults (over age 65) reside in suburban neighbourhoods (Hartt, Biglieri, Rosenberg, & Nelson, Reference Hartt, Biglieri, Rosenberg and Nelson2021), it is critical to consider the role of the built environment in later life. To begin to answer this question, this macro-scale investigation seeks to understand the “double risk” that many older adults live with: the potential of being disadvantaged by socio-demographic risk factors and being further disadvantaged by living in a non-supportive built environment. In other words, where do these older adults in Canada live, and are those neighbourhoods classified as more or less supportive built environments?
This article presents the findings of a macro-scale investigation into Canadian older adults living in more vulnerable conditions and offers preliminary data on what types of built environments they live in, which in turn can influence their independent mobility and physical, mental, and social health. We begin by examining two bodies of empirical literature. First, we look at existing research on the experiences of older adults in different typologies of neighbourhoods (as assessed for their walkability), and their influence on their health. Second, we examine research on the socio-demographic risk factors that affect the experiences of older adults. These literatures were used to inform the selection and analysis of two data sets (neighbourhood-scale built environment typologies and socio-demographic factors) to understand where the most vulnerable older adults in Canada live, judging by these double risk factors. Statistical and spatial analyses of the relationships provide insights that may help municipalities target neighbourhoods for interventions that address these inequities in risk factors through built environment and other programmatic interventions.
In macro-scale investigations, municipalities tend to be treated as single entities, as opposed to having diverse neighbourhood contexts, which can impact health and well-being in diverse ways (Corburn, Reference Corburn2013; Hartt & Hollander, Reference Hartt and Hollander2018). The impact of neighbourhood-level built environment factors on health and well-being, as well as associated inequities within cities, have been well-documented and present a pressing need to further understand these relationships (Corburn, Reference Corburn2009, Reference Corburn2013; Duncan & Kawachi, Reference Duncan and Kawachi2018; Egger & Dixon, Reference Egger and Dixon2014; Frohlich & Abel, Reference Frohlich and Abel2014; Giles-Corti et al., Reference Giles-Corti, Vernez-Moudon, Reis, Turrell, Dannenberg and Badland2016). For instance, previous studies evaluated the associations between neighbourhood differences and socio-economic status and health (Bisung, Kangmennaang, & Luginaah, Reference Bisung, Kangmennaang and Luginaah2018; Mitchell, Richardson, Shortt, & Pearce, Reference Mitchell, Richardson, Shortt and Pearce2015; Vincens, Emmelin, & Stafström, Reference Vincens, Emmelin and Stafström2018), access to healthy foods (del Canto, Engler-Stringer, & Muhajarine, Reference del Canto, Engler-Stringer and Muhajarine2015), physical activity and obesity levels (Kowaleski-Jones et al., Reference Kowaleski-Jones, Zick, Smith, Brown, Hanson and Fan2018; Witten, Pearce, & Day, Reference Witten, Pearce and Day2011), as well as exposure to harmful pollutants (Kravitz-Wirtz et al., Reference Kravitz-Wirtz, Teixeira, Hajat, Woo, Crowder and Takeuchi2018) and brownfields (Bambra et al., Reference Bambra, Robertson, Kasim, Smith, Cairns-Nagi and Copeland2014). There are also inequities between neighbourhoods in terms of support provided by the built environment, especially in terms of access to public infrastructures (transportation systems, parks and greenspaces, social services, educational facilities, affordable housing), which influences social inclusion, health outcomes, and quality of life (Duncan & Kawachi, Reference Duncan and Kawachi2018; Lo et al., Reference Lo, Anisef, Wang, Preston, Basu and Wang2015; Masuda et al., Reference Masuda, Teelucksingh, Zupancic, Crabtree, Haber and Skinner2012). These inequities in the built environment are significant for older adults, who are more vulnerable to the health and social impacts of their neighbourhood when compared with other age groups (Ghani, Rachele, Loh, Washington, & Turrell, Reference Ghani, Rachele, Loh, Washington and Turrell2018). For this research, we expand the typical investigation of major Canadian cities (Toronto, Vancouver, Montreal) to add 30 other Canadian municipalities, most of which are classified as suburban. This combination of municipalities allows us to interrogate the idea of suburbs as monolithic (Keil, Reference Keil2018) and to expand upon work investigating inequitable distributions of infrastructure (and thus, vulnerability) in Canadian suburban municipalities, which are argued to be attributable to two decades of neoliberal disinvestment in outer suburban areas due to austerity climates; rapidly growing vulnerable populations in these spaces; and suburban land use typology of segregated uses, automobile-oriented, and low densities exacerbating public infrastructure needs (Keil, Reference Keil2018; Lo et al., Reference Lo, Anisef, Wang, Preston, Basu and Wang2015). Understanding vulnerability at the neighbourhood level is essential as earlier case studies have revealed disparities within municipalities like York Region, Ontario, which create “complex geographies that call for innovative planning responses” (Lo et al., Reference Lo, Anisef, Wang, Preston, Basu and Wang2015, p. 148).
Understanding city neighbourhoods in terms of access to services and amenities is one way to more accurately reflect heterogeneous experiences of everyday life in suburban areas. For older adults, there is a need to identify these inaccessible and potentially unsupportive built environments to address these issues of inequity and direct targeted policy responses (Buffel, Phillipson, & Scharf, Reference Buffel, Phillipson and Scharf2012; Lo et al., Reference Lo, Anisef, Wang, Preston, Basu and Wang2015). Our approach is framed by an environmental justice perspective (Day, Reference Day2010) and thus examines vulnerability risk factors through a neighbourhood-by-neighbourhood lens to understand inequities in double risk. The neighbourhood-by-neighbourhood approach is a way of planning that is rooted in the contextual experiences of the people living in that neighbourhood and the neighbourhood’s unique built environment, service, and policy landscape. It means tailoring community engagement and resultant policies and interventions to the specific needs of that neighbourhood. This is in contrast to city-wide policies that are high level and a one-size-fits-all approach, which often misses these contextual needs. Overall, we found that an extremely large proportion of older adults living with vulnerable socio-demographic risk factors is also living in potentially unsupportive built environments in Canada. This means most older adults living with more socio-demographic vulnerable conditions are also facing significant barriers in their neighbourhood surroundings when it comes to independent mobility (which then negatively impacts their physical, mental, and social health). This double risk must be identified, measured, and taken into consideration to ensure the health and well-being of older Canadians.
Overview of the Aging and Built Environment Literature
The Person-Environment Fit theory argues that, as we age, we are more likely to experience impairment and therefore more likely to be negatively impacted by the environment (Lawton, Reference Lawton, Windley and Byerts1982), requiring a study of person-place relations at both the micro and macro scale (Wahl & Gitlin, Reference Wahl, Gitlin and Birren2007). This is echoed in the geographies of aging literature – the need to conduct multi-scalar inquiries, in addition to a move to more relational understandings of aging and place (Skinner, Cloutier, & Andrews, Reference Skinner, Cloutier and Andrews2015). Older adults are more likely than other age groups to spend time in their immediate neighbourhoods (Glass & Balfour, Reference Glass, Balfour, Kawachi and Berkman2003; Kerr, Rosenberg, & Frank, Reference Kerr, Rosenberg and Frank2012), are more sensitive to the built environment than other age groups (Ghani et al., Reference Ghani, Rachele, Loh, Washington and Turrell2018), and a systematic review concluded that the “neighbourhood environment is important for older adults’ health and functioning” (Yen, Michael, & Perdue, Reference Yen, Michael and Perdue2009, p. 455). In addition, research indicates that, as we age, our life spaces change – they effectively shrink, making it important to understand the impact of one’s immediate built environment on the well-being of the individual (Rosso, Auchincloss, & Michael, Reference Rosso, Auchincloss and Michael2011).
The importance of the neighbourhood in the health and well-being of older adults has been recognized through the World Health Organization policy on Age-Friendly Cities (AFCs). AFC policies have been endorsed by federal and provincial levels of government in Canada to encourage municipalities to plan for their aging populations by conducting extensive public consultation and assessing eight domains of age-friendliness (outdoor spaces and public buildings, transportation, housing, social participation, respect and social inclusion, civic participation and employment, communication and information, and community supports and health services) (Ontario Seniors Secretariat, 2013). However, criticism of the policy remains, primarily because it tends to be applied in a uniform way that lacks important context-specific detail and negates larger processes of urban change (Buffel et al., Reference Buffel, Phillipson and Scharf2012; Buffel & Phillipson, Reference Buffel and Phillipson2016; Scheidt & Windley, Reference Scheidt, Windley, Birren, Schaie, Abeles, Gatz and Salthouse2006). AFC planning is typically done at the municipal level, but some jurisdictions (like New York City) have done this type of planning at the neighbourhood scale (Steels, Reference Steels2015). Research on population projections and AFC policy uptake in Ontario shows that cities with the greatest projected demographic share of older adults are the least likely to have started age-friendly planning (Hartt & Biglieri, Reference Hartt and Biglieri2018). Fortunately, 33 per cent of Ontarian municipalities have undertaken some AFC planning; however, the quality and status of those plans are unknown (Hartt & Biglieri, Reference Hartt and Biglieri2018).
The Chief Public Health Officer of Canada (2017) recently released a report highlighting the need to focus more research on the built environment’s impacts on health – specifically, the importance of encouraging physical activity, promoting healthy food options, and supporting mental wellness, especially for vulnerable populations like older adults. While there is still much to be learned on how the places we live impact our health, research shows that walkable, mixed-use neighbourhoods with good public transport, and easy access to amenities, services, family, and friends tend to produce higher levels of physical activity in older adults (Hirsch, Winters, Clarke, Ste-Marie, & McKay, Reference Hirsch, Winters, Clarke, Ste-Marie and McKay2017; Kerr et al., Reference Kerr, Rosenberg and Frank2012) and encourage higher levels of social interaction and social capital (Levasseur et al., Reference Levasseur, Généreux, Bruneau, Vanasse, Chabot and Beaulac2015; Leyden, Reference Leyden2003). The Improving Health by Design Report produced by the Chief Medical Officers of Health in the Greater Toronto and Hamilton Area (2014) called on public health policy makers and planners to promote walkable mixed-use neighbourhoods as a way to combat chronic disease and encourage active transportation. The concept of supportive, walkable, and widely accessible neighbourhoods is particularly salient when considering the transportation options for older adults who have lost their driver licences and tend to live in places with poor public transit and faraway amenities. Our study examines whether older adults with vulnerable socio-demographic conditions are living these policy- and research-purported supportive built environments.
Examining the Literatures + Selecting Data Sets: Environments and Socio-Demographic Factors
Using the Person-Environment Fit theory alluded to earlier, this research looks at the two “halves” of the theory to elucidate relationships between selected socio-demographic vulnerability risk factors and built environments in Canadian neighbourhoods (Figure 1). By examining the relationality between these two sets of factors, it can give us some insight into the ability of older adults to conduct desired activities and their distribution across municipalities. Activity is the third part of the model utilized by this research (Iwarsson & Ståhl, Reference Iwarsson and Ståhl2003). Personal and built environment factors combine to have an impact on a person’s ability to do and access activities in their lives. From a planning and local policy perspective, activity is interpreted similarly to mobility – the ability to access places in relation to one’s home and the type of transportation mode required to get there (e.g., walking, car, public transit). In terms of built environment features, we used a neighbourhood typology (ranging from walkable and more likely to be supportive to auto-dependent and less likely to be supportive), and for socio-demographic factors, we used age (over 65, over 85), socio-economic status, and living alone. We first examined the categorization of supportive and unsupportive built environments via Gordon and Janzen’s (Reference Gordon and Janzen2013) neighbourhood typology.
Theory and measures concept graphic.
Built Environment Literature: Relationships Between Walkability and Health/Well-being of Older Adults
Research on older adults has indicated that walkable neighbourhoods are more likely than suburban neighbourhoods to be supportive for older adults. First, walkable neighbourhoods often have better access to services and amenities (which can be accessed on foot, rather than by car), which can foster a sense of independence and dignity (Levasseur et al., Reference Levasseur, Généreux, Bruneau, Vanasse, Chabot and Beaulac2015). In car-dependent neighbourhoods, while being able to drive makes older adults feel independent, enabling them to complete daily activities across the city, the question becomes what happens when these individuals can no longer drive (Zeitler & Buys, Reference Zeitler and Buys2015). Car-dependent neighbourhoods “discourage participation in the community” for those who are unable to drive (Lord et al., 2011, as cited in Zeitler & Buys, Reference Zeitler and Buys2015, p. 805). There have been numerous studies on the negative impacts of losing one’s driver’s licence, such as increased depressive symptoms (Ragland, Satariano, & Macleod, Reference Ragland, Satariano and Macleod2005), negative effects on social integration (Mezuk & Rebok, Reference Mezuk and Rebok2008), and decline in physical, social, and cognitive function (Chihuri et al., Reference Chihuri, Mielenz, DiMaggio, Betz, DiGuiseppi and Jones2016). For many older adults, driving means being able to perform tasks and maintaining socialization and a sense of autonomy, independence, and self-worth. With the loss of driving comes diminished agency and an increased dependence on others and perceptions of the self as a burden (Dickerson et al., Reference Dickerson, Molnar, Eby, Adler, Bedard and Berg-Weger2007; Kerschner & Aizenberg, 1999, as cited in Dumbaugh, Reference Dumbaugh2008). However, theoretically, if you lost your driver’s licence and you lived in an area where you could accomplish most things on foot, the impact of losing your license might be less. Access to public transport in Boston, MA, USA, was found to mitigate the negative impacts associated with losing one’s driver’s licence, and older adult non-drivers in the city ended up making more trips than older adult drivers and non-drivers in the suburbs (Coughlin, 2001, as cited in Dumbaugh, Reference Dumbaugh2008). Other studies have shown that walkability is an important correlate of food insecurity, specifically, mobility-related food insufficiency for older adults (Chung et al., Reference Chung, Gallo, Giunta, Canavan, Parikh and Fahs2011). Walkable neighbourhoods are also overall safer for pedestrians; one study found that the built form of a neighbourhood had an impact on both older adult pedestrian and motorist deaths (Dumbaugh & Zhang, Reference Dumbaugh and Zhang2013).
Walkable neighbourhoods also have the potential to increase physical activity and active transport for older adults (Cerin, Nathan, Van Cauwenberg, Barnett, & Barnett, Reference Cerin, Nathan, Van Cauwenberg, Barnett and Barnett2017; Hirsch et al., Reference Hirsch, Winters, Clarke, Ste-Marie and McKay2017; Nyunt et al., Reference Nyunt, Shuvo, Eng, Yap, Scherer and Hee2015). Studies have shown that for older adults, those living in a walkable neighbourhood were more physically active than those living in a car-dependent neighbourhood and were three times as likely to meet Canadian national physical activity guidelines (Winters et al., Reference Winters, Voss, Ashe, Gutteridge, Mckay and Sims-Gould2015). A systematic review and meta-analysis found that walkability was best associated with health-enhancing levels of physical activity for older adults in terms of features of the built environment (Barnett, Barnett, Nathan, Van Cauwenberg, & Cerin, Reference Barnett, Barnett, Nathan, Van Cauwenberg and Cerin2017). Physical activity has also been found to be a protective factor in regard to developing Alzheimer’s and Parkinson’s diseases (Paillard, Rolland, & de Souto Barreto, Reference Paillard, Rolland and de Souto Barreto2015) and of future cognitive ability (Stubbs, Chen, Chang, Sun, & Ku, Reference Stubbs, Chen, Chang, Sun and Ku2017). Finally, walkable neighbourhoods hold potential for increased social interaction and connection (Ferreira, Johansson, Sternudd, & Fornara, Reference Ferreira, Johansson, Sternudd and Fornara2016; Kaczynski & Glover, Reference Kaczynski and Glover2012). A study done in Montreal, Canada, compared older adults living in an urban versus a suburban neighbourhood and found that those in urban neighbourhoods reported greater levels of social interaction (Richard, Gauvin, Gosselin, & Laforest, Reference Richard, Gauvin, Gosselin and Laforest2008). The presence of “third places” (spaces for social interaction that are not home or work) like coffee shops, commercial areas, libraries, and parks has been identified as places important for social health of older adults (Alidoust, Bosman, & Holden, Reference Alidoust, Bosman and Holden2018; Oldenburg, Reference Oldenburg1989). Further, a study comparing neighbourhood typologies in Australia found that third places are more likely to be located, and are perceived as more accessible, in walkable neighbourhoods when compared with automobile-dependent ones (Alidoust et al., Reference Alidoust, Bosman and Holden2018).
Selecting the Data Set: Categorizing Supportive Versus Unsupportive Built Environments in Canada
The question becomes: How do we classify built environments in Canada? After conducting dozens of empirical experiments, Gordon and Janzen (Reference Gordon and Janzen2013) found that the most effective and precise way to define the Canadian urban and suburban landscapes was through transportation behaviour models, which “produce ‘roughly correct’ definitions for practical policy making, similar to Statistics Canada’s set of six definitions of ‘rural’ (du Plessis et al., Reference du Plessis, Beshiri, Bollman and Clemenson2002, p. 200).”Footnote 1) Their extensive study, which manually checked their models using Google Earth and conducted extensive engagement with academic and policy experts, took over five years and examined every neighbourhood in all 33 census metropolitan areas (CMAs). This resulted in a four-part typology, and generally,Footnote 2 the four types can be defined as:
Exurbs: very low-density rural areas where more than half the workers commute to the central core.
Auto Suburbs: neighbourhoods where almost all travel is done by automobile; there is negligible transit, walking, or cycling to work.
Transit Suburbs: neighbourhoods where a higher proportion of people travel by public transit.
Active Cores: neighbourhoods where a higher proportion of people uses active transportation (walk or cycle) to get to work.
In our study, we use Gordon and Janzen (Reference Gordon and Janzen2013) and Gordon and Shirokoff’s (Reference Gordon, Shirokoff, Agarwal, Filion, Grant, Harris, Hess and Luka2014) analysis of Canadian urban neighbourhoods as operational variables of levels of built environment support. For this research, we have proposed conceptualizing their suburban classifications as a spectrum ranging from supportive to unsupportive built environments. We consider “Active Core” neighbourhoods to be the most supportive built environments, followed by “Transit Suburbs,” whereas “Auto Suburbs” and, lastly, “Exurbs” are considered the least supportive built environments. The rationale behind conceptualizing these classifications as a spectrum acknowledges that these categories are not absolutes, but rather an indication of the level of support provided by the type of built environment. According to Gordon and Janzen (Reference Gordon and Janzen2013) and Gordon and Shirokoff (Reference Gordon, Shirokoff, Agarwal, Filion, Grant, Harris, Hess and Luka2014), trips are more likely to be accomplished on foot within Active Cores, whereas the other typologies are increasingly more automobile dependent. The spectrum model of built environment support allows us to take what is known about classifying more/less supportive neighbourhoods for older adults, and then identify their locations using a neighbourhood-by-neighbourhood approach. This allows us to identify more supportive neighbourhoods that are not necessarily in the centre of the city and less supportive neighbourhoods that are, thereby avoiding the categorization of cities as monolithic entities. Data for the built environment type were compiled from the Suburban Nation project website (www.canadiansuburbs.ca) where typology data at the census tract level for all 33 CMAs are publicly available. Typology data were produced by the Suburban Nation research team using 2016 census data following Gordon and Shirokoff’s (Reference Gordon, Shirokoff, Agarwal, Filion, Grant, Harris, Hess and Luka2014) technical definitions.1
Socio-Demographic Factors Literature: Capturing Older Adults Living with More Risk Factors
We have discussed the importance of the built environment for older individuals, but what about the socio-demographic vulnerability risk factors that they face? Before outlining how we measured vulnerability, it is important to unpack the term itself. Vulnerability as a concept “differs from other social science concepts that describe ‘negative states,’ such as poverty, neglect and exclusion, in its potentiality and therefore the avoidability of its undesirable outcomes” (Schröder-Butterfill & Marianti, Reference Schröder-Butterfill and Marianti2006, p. 14). Vulnerability is also constructed by social systems that have created inequitable outcomes, and often this is a result of power relations – like classism, racism, ageism, ableism, sexism, and so forth (Bankoff, Reference Bankoff2003). We use this term because it underscores our perspective that, while this research identifies where those with the most risk factors may live, there is an opportunity to address these inequities by changing health and social outcomes through targeted interventions in the built environment and through policy in programs/services (e.g., through the implementation of the Age-Friendly City framework, supplementation through programming or small-scale interventions in the built environment). We also acknowledge that vulnerability itself is not a “personal characteristic” and is instead a combination of characteristics and interactions between exposures, threats, and diverse coping mechanisms throughout the life course (Schröder-Butterfill & Marianti, Reference Schröder-Butterfill and Marianti2006, p. 18). This study is thus a first step in understanding vulnerability of older adults across Canada, by investigating exposures to certain socio-demographic risk factors and built environments. Future, in-depth work to understand intersections and particularities of exposure, threats, and coping mechanisms is still required. Thus, for this study, our operational definition of vulnerability risk factors is based on three socio-demographic factors:
-
1. Age – despite a heterogeneity of aging experiences, being older, especially in the over 85 age range, is associated with a greater risk for issues with mobility and is a good indication of activity space size (Hodge, Reference Hodge2008). Further, ageism (discrimination against older people) has been identified as a social determinant of health and is pervasive in policies, infrastructure, programming, services, and interpersonal interactions (Mikton, de la Fuente-Núñez, Officer, & Krug, Reference Mikton, de la Fuente-Núñez, Officer and Krug2021).
-
2. Low socio-economic status has long been considered a social determinant of health, and living on a low income is a strong predictor of mortality and perceived health (Raphael, Reference Raphael2011).
-
3. Living alone has been associated with being at risk for social isolation and therefore worse physical and mental health (Cornwell & Waite, Reference Cornwell and Waite2009). Nicholson (Reference Nicholson2012) found that there was an “overabundance of evidence demonstrating numerous negative health outcomes and potential risk factors related to social isolation [in older adults]” (p. 137). Further, a meta-analytic review found that living alone for older adults was associated with a 32 per cent increased likelihood of mortality (Holt-Lunstad, Smith, Baker, Harris, & Stephenson, Reference Holt-Lunstad, Smith, Baker, Harris and Stephenson2015).
Social scientists have cautioned that the studying of vulnerability is quite complex and should be understood as a confluence of a whole history of life experiences and coping mechanisms requiring a detailed qualitative review (Schröder-Butterfill & Marianti, Reference Schröder-Butterfill and Marianti2006, p. 14). However, from a public policy perspective, Klinenberg’s (Reference Klinenberg2002) social autopsy of the heat wave disaster in Chicago in 1995 (in which an estimated 739 people died from heat-related causes, the majority of whom were older, poor residents) shows that studying proxy measures of vulnerability are vital to avoid risking the undercounting of vulnerable older adults. Klinenberg (Reference Klinenberg2002) notes that the surveying of vulnerable older adults is a nearly impossible task, as those who come out to public meetings and participate in research are likely to be older adults who are considered quite active in the community. The concept of staying in one’s own home has been glorified in society, and many older adults fear losing their independence. As a result, those struggling with issues living at home may be less likely to ask for help or become politically involved (Klinenberg, Reference Klinenberg2002). This research argues that policy makers need to be aware of where people live so that services can be targeted and retrofits to less supportive neighbourhoods can be prioritized (as opposed to only relying on the opinions of those who attend public meetings).
Selecting the Data Set: Exploring Socio-Demographic Risk Factors in the Canadian Census
The question then becomes how to match the information from the built environmental typologies to these socio-demographic factors as collected by the Canadian Census (administered by Statistics Canada). Due to limitations on how data are collected, the potential vulnerability of older adults was measured using a range of available socio-demographic variables based on the literature review: population over 65 years of age, population over 65 years of age living alone, population over 65 years of age living on low incomes, population over 85 years of age, population over 85 years of age living alone, and population over 85 years of age living on low incomes. All variables were considered both as absolute values and as proportions relative to the census tract total. Low income was defined by Statistics Canada’s after-tax low-income measure (LIM-AT). All data were collected at the census tract level from the 2016 Canadian Census. According to Statistics Canada, “Census tracts (CTs) are small, relatively stable geographic areas that usually have a population between 2,500 and 8,000 persons. They are located in census metropolitan areas and in census agglomerations that had a core population of 50,000 or more in the previous census. A committee of local specialists (for example, planners, health and social workers, and educators) initially delineates census tracts in conjunction with Statistics Canada” (2018).
Findings: Combining the Data Sets
The principal objective of this research was to explore the prevalence of the double risk facing older Canadians. Empirically, the goal was to determine whether older Canadians disadvantaged by social determinants of health are more likely to be living in more or less supportive environments. To do so, we examined every neighbourhood in all 33 of Canada’s major cities (as defined by their CMA, and henceforth referred to as cities). The location of the case study cities is depicted in Figure 2. Census tracts, as defined by Statistics Canada, were used as a proxy for neighbourhoods. In total, 5,299 census tracts were included in the analysis.
Map of 33 case study cities (CMAs).
Descriptive statistics were used to examine the number and proportion of older adults living in different built environments. A series of Kruskal-Wallis tests were used to analyse older adult vulnerability across built environment types, and Pearson correlations were used to ascertain the relationship between older adults living alone and those living on low incomes.
Taking Stock of Older Canadians Across the Country
According to the built environment typology, the Canadian urban landscape is made up of 763 Active Core (most likely to be supportive) neighbourhoods, 554 Transit Suburbs, 3,439 Auto Suburbs, and 500 Exurbs. Toronto, Ontario (ON), had the most neighbourhoods on the most supportive end of the spectrum (Active Core) at 137, whereas Abbotsford–Mission, British Columbia (BC), had the fewest (0). Table 1 provides univariate descriptive statistics of older adults living in Canadian cities and levels of older adults living in vulnerable conditions at the census tract level. In 2016, almost four million Canadians were 65 years of age or older. Approximately one quarter of them were living alone and one eighth were living on low incomes. Over half a million Canadians were 85 years of age or older, with similar proportions of living alone and living on low incomes.
Descriptive statistics of older adults living in vulnerable conditions in every census tract (n = 5,299) in every Canadian CMA (n = 33)
Source: Statistics Canada (2018).
In examining older adults and their environments, we found that an extremely large proportion of older adults with vulnerable conditions are residing in built environments that are on the unsupportive end of the built environment spectrum. Table 2 shows the absolute number of older residents with vulnerable conditions in all 33 cities. In total, we found that almost 70 per cent of adults 65 years of age or older living alone were living in unsupportive Auto Suburbs or Exurbs. In absolute numbers, that equates to almost 650,000 Canadians. Similarly, we found that over 65 per cent of adults 65 years of age or older (over 300,000 residents) with low incomes were living in less supportive Auto Suburbs or Exurbs. For older adults 85 years of age and above, a group more likely to be living in more vulnerable conditions, the results were similar. Over 80 per cent of adults 85 years of age and above in Canadian cities living alone (approximately 125,000 residents) and 80 per cent of those 85 and older with low incomes (over 50,000 residents) were found to be living outside of the most supportive Active Core built environments.
Number and proportion of older adults living by vulnerability category residing in Active Core, Transit Suburb, Auto Suburb, and Exurb built environments in Canadian CMAs
Note: Older adult vulnerability categories are not mutually exclusive.
When examining the absolute numbers in each vulnerability category across Canada, we found roughly 11 to 20 per cent are living in the most supportive Active Core, 10 to 17 per cent in the second most supportive Transit Suburb, 60 to 70 per cent in the less supportive Auto Suburb, and 4 to 8 per cent in the least supportive Exurb. In short, older Canadians living in vulnerable conditions in cities were overwhelmingly found to be living in built environments at the unsupportive end of the spectrum.
Neighbourhood Differences Within and Between Cities
Figure 3 breaks down the proportion of older adults living in vulnerable conditions residing outside of Active Core neighbourhoods for each of the 33 cities, namely Transit Suburbs, Auto Suburbs, and Exurbs. In most cities, the proportion of older adults living with vulnerable conditions in these types of less supportive built environments was close to 80 per cent. Interestingly, Regina, Saskatchewan (SK), with a population of about 229,000, was found to have the least potential double risk. Although still substantial relative to the Canadian urban landscape, Regina had the smallest proportion of older adults living in vulnerable conditions in less supportive built environments. Thus, out of the 33 CMAs studied, Regina has the highest percentages of older adults living in the most supportive neighbourhoods like Active Core and Transit Suburbs. Still, for those 65 years of age and older in Regina, just over 50 per cent of those with low incomes and just under 60 per cent of those living alone were in either Auto Suburbs or Exurbs – the least supportive built environments. For the 85-years-and-older group, less than 32 per cent of those with low incomes and 50 per cent of those living alone were found to be in the least supportive built environments (Auto Suburb and Exurb). See Figure 4 for visual representation of these findings.
Proportion of older adults living alone and with low incomes in unsupportive neighbourhoods (Transit Suburbs, Auto Suburbs, and Exurbs) in every Canadian CMA.
Visualizing the data in Regina, SK.
As Figure 3 and Table 3 show, the number of vulnerable older adults living in the least supportive neighbourhoods in Regina, SK, is noticeably lower than that of most Canadian cities. On the other extreme sits Abbotsford–Mission, BC, a mid-sized city of about 148,000 residents. With no supportive Active Core nor Transit Suburb neighbourhoods in the city, 100 per cent of its older adult population live in built environments on the least supportive end of the spectrum. Perhaps most interestingly, even in cities with larger numbers of supportive neighbourhoods, the majority of older adults living in vulnerable conditions were found to be living in the less supportive built environments in those communities. For example, Toronto, ON, has (as one might expect) the greatest number of more supportive Active Core (137) and Transit Suburb (162) neighbourhoods, yet almost 90 per cent of older adults in all vulnerable condition categories were found to be living in non-Active Core built environments, with about 15–25 per cent in Transit Suburbs, 61–73 per cent in Auto Suburbs, and 1–4 per cent in Exurbs. Peterborough, ON, an example of another mid-sized city in Canada with a population of about 82,000, has nearly double the percentage of older adults in all vulnerable condition categories living in more supportive neighbourhoods than Toronto does, with between 17 and 40 per cent living in Active Core neighbourhoods. These examples demonstrate that the potential double risk of older adults living in vulnerable conditions in unsupportive built environments is not strictly a large or mid-sized city phenomenon – it is a considerable issue that plays out in unexpected ways across the variegated Canadian urban landscape. However, considering the suburban nature of many of Canada’s cities, an important distinction remains to be made: Are vulnerable older adults disproportionately residing in less supportive neighbourhoods, or are there simply many more less supportive than more supportive neighbourhoods in Canadian cities?
Number and proportion of older adults living by vulnerability category residing in Active Core, Transit Suburb, Auto Suburb, and Exurb built environments in Regina, Peterborough, Toronto, and Abbotsford
Note: Older adult vulnerability categories are not mutually exclusive.
When validating the results, a series of Kruskal–Wallis tests were used to examine the median proportion of older adults living alone or with low incomes in the four built environment types. The results summarized in Table 4 show that, in all four vulnerability variables, there was a significant difference between the median proportion of older adults in distinct built environments. Generally, the proportion of older adults living in vulnerable conditions followed the urban hierarchy with the highest proportions in Active Core neighbourhoods, followed by Transit Suburbs, then Auto Suburbs, and finally Exurbs. On average, there was almost double the proportion of both groups of older adults living alone and with low incomes in Active Core and Transit Suburb built environments. This finding is distinctly different than when analysing the Canadian urban population as a whole (see Table 2), which showed the majority of vulnerable older adults living in Auto Suburbs. Therefore, we can surmise that the overwhelming number of older adults living in vulnerable conditions in less supportive environments (depicted in Table 2) is not an issue of underrepresentation in supportive environments, but rather the fact that the vast majority of urban environments in Canada are inherently less supportive.
Kruskal–Wallis tests comparing the median proportion of older adults living in vulnerable conditions by built environment type
Although we cannot say for certain how many older individuals are both living alone and with low income in each built environment due to the structure of the census data, we can examine the overarching relationship between the two. Pearson correlations examining the relationship between the proportion of older adults living alone and the proportion of those with low incomes were run for all 5,299 census tracts included in the analysis. Correlation results show that there is a strong positive correlation between living alone and with low incomes for both older adult groups (over age 65 and over age 85). Therefore, neighbourhoods with high proportions of older adults living alone also tend to have high levels of older adults living with low income. It cannot be said for certain how many individuals are experiencing both; however, the results suggest that many older Canadians are living alone with a low income.
Discussion
Beyond Assumptions About City Size and Aging in Place
Through analysing double risk for older adults in Canada, we reveal the reality that Canadian cities and the types of neighbourhoods that older adults live in are more complicated than common perceptions about large versus mid-sized cities might have us believe. While most older Canadians in the 33 Census Metropolitan Areas across all the socio-demographic vulnerability risk factors were found to be living in less supportive built environments, which is consistent with Gordon and Janzen’s (Reference Gordon and Janzen2013) “Canada is a suburban nation” literature, the most interesting findings came to light when the numbers were parsed out between the different typologies on the supportive/unsupportive built environment spectrum, and between municipalities. For instance, the fact that Regina and Peterborough have greater proportions of older adults living with more vulnerable conditions in supportive neighbourhoods than Toronto upends the assumption of Toronto as a completely urban, walkable municipality. The research is congruent with decades of work on suburbanization, where poverty and diversity are increasingly concentrated in the peripheries of cities, and there is a lack of social, health, and built infrastructures, with concentrations of wealth, whiteness, and privilege in the centre (Biglieri, De Vidovich, & Keil, Reference Biglieri, De Vidovich and Keil2020; Hulchanski, Reference Hulchanski2010; Keil, Reference Keil2018). The findings from the Toronto context can add to the literature already documenting these processes – recording the peripheralization of older adults living in more vulnerable socio-demographic conditions, away from the supportive built environments that could positively impact their health and well-being.
These data could also serve as a starting point for understanding why some municipalities have higher proportions of older adults with more vulnerability risk factors living in more supportive neighbourhoods than others – like Regina and Peterborough – perhaps investigating why this is through an examination of the municipality’s policies, housing stock, programs/services for older adults, and in-depth qualitative research into the perceptions of older adults and community members in that municipality. This could, in turn, help policy makers and academics understand how supportive environments for older adult populations are/have been produced in these locations and identify diverse factors (e.g., policies, macro economic considerations, governance, infrastructures, community organizing) that might be considered in communities without as many supportive environments. For instance, in municipalities with no supportive built environments like Abbotsford, this kind of investigation could raise questions about the diversity of housing available in the municipality and lead to residential intensification of their commercial areas to change those neighbourhoods to being more supportive, in addition to increasing programming, services, and other supports.
Practical Applications: Built Environment Spectrum Neighbourhood-by-Neighbourhood Approach
Conceptualizing built environments as more or less supportive is a nuanced approach that, when combined with the socio-demographic statistics, means we can reveal diversity within cities, and in doing so, upend centralist bias and, importantly, begin to understand the inequitable distribution of double risk for older adults (Gordon & Janzen, Reference Gordon and Janzen2013, p. 214; Keil, Reference Keil2018). It also allows for future research to conduct deeper qualitative relational work, further investigating aspects like place attachment and using a more nuanced understanding of diversity within these neighbourhood classifications themselves (Andrews, Evans, & Wiles, Reference Andrews, Evans and Wiles2013). On a practical level, it can also enable policy makers to tailor their policies and approaches to diverse neighbourhoods, responding directly to critiques of AFCs as lacking context-specific detail and negate larger processes of urban change (Buffel et al., Reference Buffel, Phillipson and Scharf2012; Buffel & Phillipson, Reference Buffel and Phillipson2016; Scheidt & Windley, Reference Scheidt, Windley, Birren, Schaie, Abeles, Gatz and Salthouse2006). This is essential, especially when municipalities and community groups charged with implementing AFC plans are faced with climates of austerity (Joy, Reference Joy2020).
For instance, the built environment domains of the AFC model (transportation, outdoor spaces and buildings, housing) could be perceived as the easiest to evaluate with checklists, with domains like social inclusion being more complex. Programs and services to foster social inclusion differ from built environment domains in that they tend to be more easily piloted, evaluated, changed quickly based on feedback, more financially feasible, and present a quicker time frame for results – which makes it easier for organizations and governments to demonstrate success. From this perspective, the built environment domains can be the hardest to change, as they require significant private and/or public investment (which is harder to justify in austerity climates for governments or funding agencies), and change tends to happen over longer periods of time, making it more difficult to pilot, to evaluate, and to respond to feedback. The difference between living in a more supportive and less supportive built environment can be mitigated by the targeted delivery of services (Warner, Homsy, & Morken, Reference Warner, Homsy and Morken2017) or through small improvements to the built environment (like improved public transit, more benches, safer traffic crossings) (Levasseur et al., Reference Levasseur, Généreux, Bruneau, Vanasse, Chabot and Beaulac2015; Macmillan et al., Reference Macmillan, Mackie, Hosking, Witten, Smith and Field2018; Ottoni, Sims-Gould, Winters, Heijnen, & McKay, Reference Ottoni, Sims-Gould, Winters, Heijnen and McKay2016). A longitudinal study in Sweden found that even though over a period of nine years older adults experienced losses of cognitive and physical functioning, small improvements (e.g., levelling pavement, separating cyclists and pedestrians, lowering curbs, lowering speed limits and wider sidewalks) significantly lessened their perceptions of environmental barriers (Hallgrimsdottir & Ståhl, Reference Hallgrimsdottir and Ståhl2018). The findings from this research could be used in decision making for municipalities and to help decide where to direct services for older adults and/or built environment improvements to address inequities in distribution of risk.
A Tool for Understanding and Operationalizing Aging in Community
Finally, this spectrum approach to understanding built environments and socio-demographic risk factors could offer insights into the potential for older adults to age in community, a key idea in the literature, and one often expressed by older adults (Thomas & Blanchard, Reference Thomas and Blanchard2009). The idea is to move away from policies that focus on helping you stay in the house you have always lived in, and instead focus on moving to a more accessible location/unit, located within your community, and maintaining your social support network – thus retaining the well-being impacts produced by place attachment over time (Aliakbarzadeh et al., Reference Aliakbarzadeh Arani, Zanjari, Delbari, Foroughan and Ghaedamini Harouni2022; Lewicka, Reference Lewicka2011). For example, instead of living alone in a 2,500-sq ft house with high maintenance costs and poor walkability, one might consider moving to an apartment in the same municipality that is in a supportive neighbourhood. The built environment spectrum approach can provide indication that in some municipalities in Canada, one cannot move easily to a neighbourhood that is more supportive and maintain a sense of place attachment, social relations, and independent mobility. For example, Abbotsford–Mission, BC, has zero Active Core or Transit Suburb neighbourhoods, meaning that someone living there would have to move to another municipality to be in a more supportive neighbourhood, and therefore might experience a loss of social relationships. A mid-sized municipality like Peterborough, ON, at 53 km2, however, is a great example in which someone who lived in an auto-dependent suburb (n = 24) could theoretically move to a more supportive walkable neighbourhood (n = 6) and still be connected to their social networks. A large city like Toronto has both options as well, but considering its vast geographic scale (~ 630 km2, ~ 21 km north–south, ~ 43 km east–west), moving from one corner of the city to another may pose a challenge and loss of social networks.
These three examples point to the need for the creation of a range of housing types within neighbourhoods, combined with commercial and retail uses accessible on foot. One way to accomplish this is to require retail on corners within single-use residential neighbourhoods or require higher densities along corridors. That way, when someone needs to downsize, they can, without losing their social networks or attachment to place. It is important to note that the ability to move is also highly dependent on one’s socio-economic status, and areas like the Greater Toronto Area or the Greater Vancouver Area pose risks to older adults wishing to stay in the communities where they have formed social ties. The issue of the cost of housing is outside the scope of this paper, but future research needs to ask: If older adults wanted to age in community, would they be able to financially? Finally, just because someone lives in an Active Core neighbourhood does not mean that that person is not socially isolated. Being able to know who is vulnerable is a complex and difficult task, and requires an in-depth understanding of exposures, threats, and coping mechanisms over the life course (Schröder-Butterfill & Marianti, Reference Schröder-Butterfill and Marianti2006). Using census counts as proxy numbers can be a valuable (albeit conservative) data source for monitoring and planning for vulnerable, older populations (Klinenberg, Reference Klinenberg2002). Our study has identified older adults with vulnerable socio-demographic conditions living in less supportive built environments, facing a double risk, and provides important insight to help guide policy interventions.
Limitations and Future Research
This research does not consider self-selection of home location nor whether or not an individual has moved to be in that location (Mokhtarian & Cao, Reference Mokhtarian and Cao2008). It is limited in understanding what type of housing an individual is living in. Additionally, while this study looked at walkability as a way to understand built environment vulnerability risk, we should note that even with the best public and active transportation infrastructure, access might still be prohibitive for folks who have mobility-related disabilities, and more research is needed on their experiences in place. Further, more deep qualitative research on relational interactions between people and their communities is needed to understand place effects on health and well-being for older adults (Andrews et al., Reference Andrews, Evans and Wiles2013; Cummins, Curtis, Diez-Roux, & Macintyre, Reference Cummins, Curtis, Diez-Roux and Macintyre2007; Dean et al., Reference Dean, Biglieri, Drescher, Garnett, Glover and Casello2020; Graham & Healey, Reference Graham and Healey1999; Lee & Dean, Reference Lee and Dean2018). Lastly, this work is not a longitudinal study on the effect of the built environment on an individual’s aging and health trajectory. Research on the long-term impacts of exposure to certain built environments is needed and ongoing (Arcaya et al., Reference Arcaya, Tucker-Seeley, Kim, Schnake-Mahl, So and Subramanian2016; Brook, Setton, Seed, Shooshtari, & Doiron, Reference Brook, Setton, Seed, Shooshtari and Doiron2018). This research is a bird’s eye view of the spectrum of built environment typologies where older adults with the potentially most vulnerable conditions live in Canada, and subsequent policy work should try to understand qualitative experiences of aging and place attachment in supportive and unsupportive built environments in the respective municipalities.
Conclusions
“Canada is a suburban nation” (Gordon & Janzen, Reference Gordon and Janzen2013, p. 197), and perhaps the findings from this research on Canadian cities are not surprising. However, we want to highlight how the design of these automobile-dependent neighbourhoods can make older adults more vulnerable. Further, when combined with other social determinants of health risk factors, the majority of older adults in Canada are left facing a “double risk,” and the distribution of this risk is varied and inequitably distributed across Canada. Consider the life-altering impact of losing your driver’s licence while living in an automobile-dependent neighbourhood – your life space shrinks, and so, too, does your independence and access to social networks. But, if you lived in a walkable neighbourhood, perhaps losing your licence would have less of an impact because you are able to walk or transit to your desired destinations. Travel for older adults is an integral component of their health and well-being – being enabled to get out independently, the quality of the surrounding built environment and proximity/availability of local services/shops, and the importance of local transport systems (including public transport and walking) are all key considerations (Graham et al., Reference Graham, de Bell, Flemming, Sowden, White and Wright2018). However, there is a reason we used the term vulnerability in this paper: because of “its potentiality and therefore the avoidability of its undesirable outcomes” (Schröder-Butterfill & Marianti, Reference Schröder-Butterfill and Marianti2006, p. 14). In identifying these populations living with more vulnerable conditions in areas less supportive than others, we have also opened the door for targeted policy intervention and, therefore, a potential for the reduction of vulnerability.
This research contributes to the larger need to conceptualize suburban space in terms of everyday life and socio-spatial supports for/barriers to well-being for vulnerable populations (Lo et al., Reference Lo, Anisef, Wang, Preston, Basu and Wang2015). It also highlights environmental injustices experienced by older adults in diverse geographic contexts (Day, Reference Day2010). Understanding intersecting vulnerabilities (age, living alone, and living alone with low income) with the level of support provided by a built environment helps untangle misconceptions of homogeneity within and between Canadian cities. Considering that widespread expansion of public infrastructures, services, and built environment retrofits is unlikely due to a climate of austerity, planners and policy makers must take into account the “varied geographies of vulnerability” (Lo et al., Reference Lo, Anisef, Wang, Preston, Basu and Wang2015, p. 150) and use analyses like this research to identify inequities and create targeted policies. Targeted policy intervention is crucial: “How can we build Age-Friendly Cities if we do not know where the most vulnerable residents live? This research found that over 80 per cent of vulnerable older adults in Canada’s metropolitan areas are living in likely unsupportive built environments – meaning they are living with double risk in terms of health and well-being. These findings can offer policy makers insight in understanding where socially isolated older adults might live to target further research, civic engagement, municipal services and/or built environment improvements. Municipalities can target unsupportive built environments as neighbourhoods for retrofitting; update planning policies to encourage neighbourhood node development with more medium density homes (e.g., stacked townhouses, walk-ups, mid-rise); expand public transportation networks; and create more pedestrian friendly streets with small interventions. One way to improve the impact of AFC policies is by helping policy makers assess where the most vulnerable older adults live in their municipalities. The work outlined here is part of the necessary toolkit to begin addressing context-specific issues and improving the lives of the most vulnerable older adults.
Funding
This research was supported by the British Academy/Leverhulme in partnership with the Department for Business, Energy and Industrial Strategy (Grant SRG1819\190264).
Open access
