Key Age-Friendly Components of Municipalities that Foster Social Participation of Aging Canadians: Results from the Canadian Longitudinal Study on Aging

Municipalities can foster the social participation of aging adults. Although making municipalities age-friendly is recognized as a promising way to help aging adults stay involved in their communities, little is known about the key components (e.g., services and structures) that foster social participation. This study thus aimed to identify key age-friendly components (AFC) best associated with the social participation of older Canadians. Secondary analyses were carried out using baseline data from the Canadian Longitudinal Study on Aging (n = 25,411) in selected municipalities (m = 110 with ≥ 30 respondents), the Age-friendly Survey, and census data. Social participation was estimated based on the number of community activities outside the home per month. AFC included housing, transportation, outdoor spaces and buildings, safety, recreation, workforce participation, information, respect, health, and community services. Multilevel models were used to examine the association between individual social participation, key AFC, and environmental characteristics, while controlling for individual characteristics. Aged between 45 and 89, half of the participants were women who were engaged in 20.2±12.5 activities per month. About 2.5% of the variance in social participation was attributable to municipalities. Better outdoor spaces and buildings (p < 0.001), worse communication and information (p < 0.01), and lower material deprivation (p < 0.001) were associated with higher social participation. Age was the only individual-level variable to have a significant random effect, indicating that municipal contexts may mediate its impact with social participation. This study provides insights to help facilitate social participation and promote age-friendliness, by maintaining safe indoor and outdoor mobility, and informing older adults of available activities.


Introduction
World population aging creates new challenges for health and well-being, and calls for a shift toward promoting active aging and equitable access to social care [1].By 2038, 23% of Canadians are expected to be 65 and over, even exceeding one-third in the Atlantic provinces [2].This warrants the need for adapted, effective, and innovative interventions to maintain and improve health, equity, and well-being, particularly by promoting social participation.Social participation is defined as a person's involvement in activities providing interactions with others [3] in community life and in important shared spaces.It evolves according to available time and resources, the societal context, and what individuals want and is meaningful to them [4].Social participation of older adults is influenced by multiple individual characteristics, such as age, gender [5], functional capacities [6], socioeconomic status [7], and having a driver's license [6].Higher social participation has also been found to be associated with several health outcomes such as lower risk of premature mortality [8], greater self-rated overall health [9], and better functional autonomy [10].
Social participation can be fostered by municipalities through age-friendly policies, services, and structures related to the communities' material and social environments, which optimize the interactions between the abilities of the person and their environment.In fact, supporting older adults in staying active by enabling social participation and inclusion in community life is an essential component of public policies promoted by the World Health Organization (WHO) [11].In Canada, municipalities from all provinces and territories recognize the need for supportive environments [13] according to an age-friendly approach [12].
Although increasing the age-friendliness of municipalities is important, little is known about which characteristics of a municipality best foster social participation.According to a Chinese study of 1383 aging adults (≥ 60 years old), clean and safe outdoor spaces, effective and helpful communication, and diverse and accessible community and health services were associated with more social support and activities [13].
In a smaller study of 171 older adults living in four Canadian cities (Ottawa, Thunder Bay, Toronto, and Victoria), greater age-friendliness was correlated with enhanced social participation and greater satisfaction with social relationships and support [14].More specifically, safety and information and advocacy were associated with the latter.Canadian municipalities where aging adults (≥ 45 years old) report greater positive health, social participation, and health equity had higher population, dwelling, and intersection density averages, and residents felt safer, more included, and supported, than in the other municipalities [Levasseur M, Naud D, Dubois MF, et al., unpublished data, June 2023].According to another Canadian study, the user-friendliness of the walking environment and proximity to key resources (e.g., groceries, banks, pharmacies, or community centers) were correlated with a higher frequency of social activities but were moderated by disability in older men (74.5 ± 4.0 years old) [10].Semi-structured focus groups were also conducted with older adults, professionals, and local actors from two Quebec municipalities (Saint-Brunode-Montarville and Sainte-Julie) where older residents age more actively [15].According to the focus groups, age-friendliness promotes active aging by favoring proximity and human-scale environments, as well as transversality (e.g., opportunities for intergenerational interactions in accessible public spaces).Since older Vol:.(1234567890) adults with restricted mobility reported lower satisfaction with social participation [16], universal accessibility may also be important.
In addition to age-friendly components, other neighborhood characteristics were found to be associated with higher social participation of older adults, such as the urban and rural classification, street connectivity, population density, and affluence [17].Moreover, a greater proportion of older residents was correlated with higher age-friendliness in Manitoban rural municipalities, especially in the domains of social environment, opportunities for participation, and communication and information [18].
Despite these studies, age-friendly key components that best foster social participation of aging adults remain mostly undetermined.A better understanding of the components of Canadian municipalities that are most associated with individual social participation is essential to support health professionals and local actors in crafting holistic interventions that enhance the social participation of aging adults living in the community.This study thus aimed to identify key agefriendly components of municipalities that best promote the social participation of older Canadians.

Design and Participants
Secondary data analyses were carried out using (1) the Canadian Longitudinal Study on Aging (CLSA), a nationwide cohort study on adults living in private dwellings and aged 45-85 at baseline (2012-2015) [19] which collected individual health, sociodemographic, and neighborhood data, (2) the Canadian census (2016) [20], for which all municipal demographic and income data were released, and (3) the Canada-wide Age-Friendly Survey (AFS) documenting key age-friendly components of municipalities (2017) [21].For the study's objective, individual and municipal data had to be considered.The individual data were based on the CLSA's respondents recruited in two cohorts, using a stratified random sampling strategy based on age, gender, and province.Potential respondents were excluded if they could not communicate in English or French, lived in one of the three Canadian territories, were full-time members of the Canadian Armed Forces, or lived on First Nations reserves or settlements.A comprehensive cohort (n = 30,111) was recruited among aging adults living within 25 km of a data collection site (or within 50 km in lower-density cities).A tracking cohort (n = 21,241) recruited respondents from the ten provinces.In the present study, both cohorts were considered and respondents (n = 41,400) were included if they lived in municipalities (m = 170) that had at least 30 CLSA participants, a group size found sufficient to produce reliable coefficient estimates for linear multilevel models [22].For the municipal data, the 2016 Canadian census was used to gather data about municipal demographic characteristics [20], specifically (a) the material and social deprivation indices, as compiled by the National Institute of Public Health of Quebec [23], (b) the urban/rural typology, and (c) demographic data.Moreover, the AFS questionnaire, documenting the key age-friendly components, was sent to all Canadian municipalities (M = 3555) [24], and was completed by the mayor or a person appointed by the mayor.A total of 921 municipalities (27.0%) completed the AFS questionnaire, mostly from Quebec, the Prairie provinces (Alberta, Saskatchewan, and Manitoba), and Ontario.For the current study, the available sample comprises a total of 25,411 respondents living in 110 municipalities that completed the AFS, with at least 30 CLSA participants.The Research Ethics Committee of the Health and Social Services Centre, University Institute of Geriatrics of Sherbrooke, approves this secondary analysis of the CLSA data (MP-31-2017-656).

Individual Variables
From the CLSA, the dependent variable, social participation, reports the frequency of participating in eight activities outside the home with others: family or friends outside the household, church or religious, sports or physical, educational and cultural, service club or fraternal organization, neighborhood, community or professional association, volunteer or charity work, and other recreational activities.
Responses are converted into the number of social activities monthly: "at least one a day" = 20; "at least one a week" = 6; "at least one a month" = 2; "at least one a year" = 1; and "never" = 0 [25].The number of monthly activities for each category is summed (theoretical range = 0-160; observed range = 0-146), and a Vol.: (0123456789) higher score represents a higher frequency of participation.The instrument has good internal consistency (α = 0.85) [6].
Five continuous variables from the CLSA were used to describe and compare characteristics of respondents: age, household size (excluding self), number of chronic conditions, depressive symptoms score (10-item Center for Epidemiologic Studies Depression scale [CES-D-10]) [26], and social support score (Medical Outcomes Study [MOS] Social Support Survey) [27].The CES-D-10 is a widely used instrument that estimates the weekly frequency of 10 items associated with depressive symptoms (e.g., being bothered, feeling fearful) and has shown good internal consistency (α = 0.72) [28].The MOS Social Support Survey includes 19 items evaluating emotional, informational, tangible and affectionate support, and positive social interaction and presented good reliability (α = 0.78) and a minimal detectable change of 1.57 among older adults [26].
Thirteen other categorical variables from the CLSA were also considered: gender, immigration status, education level, marital status, household income class, retirement status, dwelling ownership, driver's license possession and use, absence of pain and discomfort and having had at least one fall in the last year, and functional limitations (basic and instrumental activities of daily living [ADL and IADL]).Both ADL and IADL measures were dichotomized (some limitations/no limitations) according to whether participants could perform all activities unassisted.

Municipal Variables
All municipal data were measured or aggregated within municipal boundaries, which facilitated the merging with the CLSA data.From the 2016 Canadian census, the material deprivation index considers the proportion of persons without a high school diploma and employed, as well as the average personal income.The social deprivation index is based on the proportion of persons living alone, separated, divorced or widowed, and single-parent families [29].Measured at the scale of the dissemination area (DA; a sub-municipal statistical unit embedded within the municipality's boundaries) and grouped by quintiles, higher categories are associated with enhanced deprivation.Because the deprivation index was only available at the DA level, it was rescaled at the municipality level through a weighted average of the quintile of each DA encompassed by a municipality [24].
Based on Statistics Canada's typology [30], municipalities were classified as urban or rural.The former includes adjacent municipalities centered on a city core, within census metropolitan areas (CMAs; ≥ 100,000 inhabitants) and census agglomerations (CAs; ≥ 50,000), and non-adjacent municipalities (≥ 10,000, known as fringe areas).The latter (rural) refers to non-urban areas within and outside CMAs and CAs.Finally, demographic data (all measured at the municipal level) included total population density (inhabitants/km 2 ) and the percentage of the population aged 65 and over, within the municipality's boundaries.
The Age-Friendly Survey (AFS) is adapted [24] from the original survey designed by Menec and colleagues [18].The AFS contains 56 items for which municipal age-friendliness is coded as (1) yes, (2) no, or (3) I don't know.The items are categorized within nine key components (Appendix Table 4).Items answered yes were summed for each component, and then standardized on a 0-100 scale for comparison, higher scores representing greater age-friendliness.

Analysis
Participants and municipalities were described using means and standard errors for continuous variables, or frequencies and percentages for categorical ones [32].
Multilevel models were used to identify key agefriendly components (level 2) of municipalities best associated with individual social participation (level 1).First, a null model calculated the intra-class correlation and proportion of the total variation explained at the municipal level.Second, individual covariables (age, sex, and functional limitations in activities of daily living) were added and tested for fixed and random effects of their relationship with social participation.Third, the nine key age-friendly components were added.Fourth, interactions were tested between significant variables of the two levels.Lastly, other municipal variables were added (material and social deprivation indices, urban/rural classification, total population density, and the percentage of the population aged 65 and over).At all steps, a backward elimination strategy with a p-value < 0.05 was used to select the variables.In addition, as recommended, the CLSA Sample Weights [31] version 1.2 were applied Vol:.(1234567890) for all analyses, which were carried out using HLM 8 [32].

Individual Characteristics
The available sample comprises a total of 25,411 respondents living in 110 municipalities that completed the AFS.Aged between 45 and 89 years, most respondents engaged monthly in an average of 20 social activities with others and lived with at least another person (Table 1).Respondents also reported an average of three chronic conditions, few depressive symptoms, and good social support.Equally distributed according to sex, most respondents had a post-secondary education and household income below 100,000 CAD, were living in a couple, were free of discomfort, were owners of their dwelling, had a driver's license and used it, and could accomplish their ADL.Almost one respondent out of five was born outside of Canada, a third were retired, and one in ten had fallen during the previous year Table 1 Description of respondents (n = 25,411) SE, standard error; unweighted n and weighted %. a 10 items answered on a 4-level scale (0 = rarely or never; 3 = all of the time), then summed (0 to 30), with higher scores indicating greater depressive symptoms [26].b 19 items answered on 5-level scale (1 = none of the time; 5 = all of the time), then summed (0-100), with higher scores indicating more support [27]

Continuous variables
Total Vol.: (0123456789) (Table 1).When compared to the total CLSA sample (n = 51,352), those in the present study had had one less activity per month on average (p < 0.001; data not shown) and fewer chronic conditions (p < 0.001).
Respondents in the current study were also statistically less likely to be female (p < 0.001), be born outside Canada (p < 0.001), have a post-secondary education (p < 0.001) or a household income above 50 K CAD (p < 0.001), live in a couple (p < 0.001), feel pain or discomfort (p < 0.001), and not drive a car (p < 0.001).

Municipal Characteristics
Municipalities that had at least 30 CLSA respondents but did not complete the AFS questionnaire (m = 60) did not differ from those that did (data not presented).
Municipalities were mainly urban with a density less than 1000 inhabitants per km 2 , about a fifth of their total population was aged 65 and older, and their material and social deprivation were above the Canadian median, indicating greater deprivation (Table 2).Most municipalities reported high levels of agefriendliness overall but varied according to the key components: respect and social inclusion, outdoor spaces and buildings, and community and workforce participation (which includes opportunities for volunteering).Municipalities reported lower age-friendliness for communication and information adapted for an older population and adapted and affordable housing.
According to the null model, 2.5% of the social participation variance was explained at the municipal level, regardless of their characteristics or key age-friendly components (data not shown).Accounting for individual and municipal characteristics, the final model is presented in Table 3. Regarding the individual-level variables, higher social participation was associated with older age, being female, having a post-secondary education, living alone, having fewer chronic conditions and depressive symptoms, being retired, having a higher household size, more social support, a driver's license, and no functional limitations.The only random slope was for the relationship between social participation and age, which indicates that this relationship varies across municipalities.Municipal variables had a significant effect on the intercept, with the final model showing that higher social participation was associated with lower material deprivation, higher age-friendliness of outdoor spaces and buildings, and lower communication and information (Table 3).No interaction was found between age and key AFC, which means that the varying relationship between age and social participation across municipalities could not be explained by age-friendly components.

Discussion
This study aimed to identify key age-friendly components best associated with the social participation of older Canadians.Although social participation has been essentially explained by individual factors, the present multilevel study showed that higher social participation at the individual level was associated with two key age-friendly components of municipalities: outdoor spaces and buildings, and communication and information, as well as lower material deprivation.In addition, material deprivation was the only other municipal characteristic also associated with the social participation of aging adults.The relationship between age and social participation was also different across municipalities.In line with other studies [13,33], age-friendly outdoor spaces and buildings facilitated the social participation of aging Canadians.Well-maintained, outdoor and walking environments enable older adults to move around in their community in a safe manner.For example, timely snow clearing was the most reported item from the outdoor spaces and buildings component [24], an essential factor for mobility during Canadian winters.Similarly, an age-friendly environment (e.g., accessible public restrooms, buildings, and stores; and active transportation infrastructures) has been found by other studies to be a known social participation facilitator [33], which increases the likelihood that a person will go outside and enjoy getting to their activity and participating.Contrary to previous studies, which highlighted its importance for active aging [14], this study reported that a less age-friendly communication and information component was associated with slightly higher social participation.Although this result should be further investigated, this association might be explained by the fact that informal networks compensated for and enhanced social participation in the absence of official documentation in languages other than French and English, or services aimed at helping them complete official forms.Additionally, communication and information was a key age-friendly component within the lowest average scores, possibly because providing elder advocacy services and documentation in non-official languages may only be applicable to larger municipalities with a greater share of immigrants, where smaller municipalities may be less agefriendly, but still facilitate social participation.The importance of a digital divide and informal network of communication should also be further investigated.the Eastern Townships, Quebec, Canada [34], showed how communities rely on newsletters, directories, resource booklets, and door-to-door outreach to increase awareness of available activities and promote social participation and health equity.Exploring the individual items in the communication and information component could be insightful.

Recent research with older adults and local actors in
The study confirms the association between social participation and material deprivation, suggesting that thriving and wealthier communities would have policies and resources in place to facilitate mobility and opportunities for social activities [17].A British longitudinal study found that materially deprived neighborhoods have limited a wide range of social activities for older adults, including civic and leisure participation, as well as activities with family and friends [35].The authors also found that the participation of aging adults decreased at a faster rate in more-deprived versus more-affluent neighborhoods, making them vulnerable to social isolation.Nonetheless, reducing material deprivation of municipalities with aging populations can foster social participation, notably by providing the older population with more opportunities for social interactions and options for adapted housing.
Age was the only individual-level variable to have a significant random effect, indicating that municipal contexts may mediate the impact on the association between social participation and age.Although previous research showed that age is a significant predictor of social participation [8], the present results reveal that municipal characteristics not included in the models may facilitate or impede how older Canadians of various ages engage in social activities.For example, mixed-use zoning could facilitate getting outside of one's home to accomplish activities of daily living and (spontaneously or intentionally) meeting with other people, which could be relevant when mobility is restricted.In addition to using the AFS to assess the municipal age-friendly components, a qualitative assessment of the neighborhood may explain how municipal contexts interact with age to foster or impede social participation, while taking into account the heterogeneity of aging populations [14].

Strengths and Limitations
To our knowledge, this study is the first to use extensive data nested in municipalities to explain social participation.The large sample size of participants and municipalities, applied with the weighting matrix, enabled population-level inferences.This study also has some limitations.Discarding municipalities with fewer than 30 respondents for the purpose of multilevel modeling resulted in excluding most rural areas, limiting generalization to all municipalities.Initially calculated at the DA level, material deprivation was rescaled using weighted averages at the larger municipal level, similar to other works [24], which reduced the variability of deprivation (F = 0.395 l; p < 0.001) and possibly prevented observing a stronger effect.Lastly, municipalities might be too large to model the association between social participation and the environment.Although representing a diversity of neighborhoods, some key components may nevertheless be attributable to the larger regional scale (e.g., transportation).

Conclusion
Using population-based representative data and multilevel regression modeling, this study aimed to identify key age-friendly components of Canadian municipalities best associated with social participation of aging adults.Considering other individual and municipal variables, higher age-friendliness of outdoor spaces and buildings, and lower communication and information were associated with higher social participation.Additionally, the study confirmed the association of material deprivation with lower social participation of aging adults.Although social participation was primarily explained by individual factors, a multilevel approach exploring the municipal context provides grounds for testing theories and designing policies and interventions that could change behaviors and attitudes associated with age-friendly components.Because correlates of social participation may differ according to the types of activities respondents engaged in, future studies should consider analyzing finer spatial units, in addition to the profiles of social participation, rather than the total monthly participation.Overall, this study provides valuable insights to facilitate social participation and promote age-friendliness, by maintaining safe indoor and outdoor environments, and inclusively informing older adults of activities that are available.

Table 3
Final model from multilevel regression for the social participation of aging adults in Canadian municipalities (n respondents = 25,411; n municipalities = 110)