Abstract

To estimate the prevalence of social participation (exposure) and its association with positive self-assessment of overall health status (SAH) (outcome) among 7,712 Brazilian elderly interviewed in the National Health Survey 2013. A cross-sectional study that used Propensity Score (PS) to improve comparability between the group exposed and no exposed to social participation. Poisson regression was performed to determine the prevalence and association of interest using crude and adjusted by inverse probability of selection of PS. Social participation was reported by 25.1% (CI95%: 23.4-26.9) and was lower among poor older people, who depend on public transportation and live in more precarious contexts. Most did not SAH positively, but the proportion was higher when they had social participation (48.0%; CI95%: 46.0-51.0). There was a positive association of social participation with SAH positive. The association using the adjusted model (PR: 1.15; CI95%: 1.08-1.22) attenuated the estimated in the crude model. Elderly exposed were 15% more likely to provide a positive SAH. Despite low levels in Brazil, there was a positive association between of social participation and SAH, confirming that engagement in such activities provides important gains for the health and quality of life.

Key words:
The elderly; Social participation; Health status; Health Surveys; Propensity score

Resumo

Estimar a prevalência da participação social (exposição) e sua associação com a autoavaliação positiva do estado de saúde (AAS) (desfecho) de 7.712 idosos entrevistados na Pesquisa Nacional de Saúde 2013. Trata-se de estudo transversal que usou o Escore de Propensão (EP) para melhorar a comparação entre o grupo exposto e não exposto a participação social. Realizou-se regressão de Poisson para determinar a prevalência e associação em estudo com o modelo bruto e ajustado pelo inverso da probabilidade de seleção do EP. A participação social foi referida por 25,1% (IC:95%: 23,4-26,9) e foi menor entre os idosos pobres, que dependem de transporte público e vivem em contextos mais precários. A maioria não avaliou positivamente sua saúde, mas a proporção foi maior quando tem participação social (48,0%; IC95%: 46,0-51,0). Houve associação positiva da participação social com a AAS positiva. A magnitude de associação usando o modelo ajustado (RP: 1,15; IC95%: 1,08-1,22) atenuou a estimada no modelo bruto. Idosos expostos tiveram 15% a mais de chance de referir AAS positiva. Apesar de baixa prevalência no Brasil, houve positiva influência da participação social na avaliação de saúde, confirmando que o engajamento social permite ganhos importantes para a saúde e qualidade de vida.

Palavras-chave:
Idosos; Participação social; Estado de saúde; Inquéritos de Saúde; Escore de propensão

Introduction

Although population ageing is a global phenomenon, considerable heterogeneity exists across countries at different stages of development¹1 United Nations. Department of Economic and Social Affairs. Population Division. World Population Ageing 2017: Highlights [Internet]. 2017. Disponível em: http://www.un.org/en/development/desa/population/publications/pdf/ageing/WPA2017_Highlights.pdf
http://www.un.org/en/development/desa/po... . In Brazil, population aging occurs in a context of deep socioeconomic inequalities in conjunction with pronounced health risks and social and health needs²2 Oliveira BLCA, Thomaz EBAF, Silva RA. The association between skin color/race and health indicators in elderly Brazilians: a study based on the Brazilian National Household Sample Survey (2008). Cad Saúde Pública 2014; 30(7):1438-1452.

3 Pinto JM, Neri AL. Trajetórias da participação social na velhice: uma revisão sistemática da literatura. Rev Bras Geriatr Gerontol 2017; 20(2):260-273.

4 Nunes APN, Barreto SM, Gonçalves LG. Relações sociais e autopercepção da saúde: Projeto Envelhecimento e Saúde. Rev Bras Epidemiol 2012; 15(2):415-428.^-55 Lima-Costa MF, Peixoto SV, Malta DC, Szwarcwald CL, Mambrini JVM. Cuidado informal e remunerado aos idosos no Brasil (Pesquisa Nacional de Saúde, 2013). Rev Saúde Pública 2017; 51(Supl. 1):6s.. Aging is characterized by a significant decline in social, physical and emotional functioning²2 Oliveira BLCA, Thomaz EBAF, Silva RA. The association between skin color/race and health indicators in elderly Brazilians: a study based on the Brazilian National Household Sample Survey (2008). Cad Saúde Pública 2014; 30(7):1438-1452.

3 Pinto JM, Neri AL. Trajetórias da participação social na velhice: uma revisão sistemática da literatura. Rev Bras Geriatr Gerontol 2017; 20(2):260-273.

4 Nunes APN, Barreto SM, Gonçalves LG. Relações sociais e autopercepção da saúde: Projeto Envelhecimento e Saúde. Rev Bras Epidemiol 2012; 15(2):415-428.

5 Lima-Costa MF, Peixoto SV, Malta DC, Szwarcwald CL, Mambrini JVM. Cuidado informal e remunerado aos idosos no Brasil (Pesquisa Nacional de Saúde, 2013). Rev Saúde Pública 2017; 51(Supl. 1):6s.^-66 Bélanger E, Ahmed T, Vafaei A, Curcio CL, Phillips SP, Zunzunegui MV. Sources of social support associated with health and quality of life: a cross-sectional study among Canadian and Latin American older adults. BMJ Open 2016; 6:e011503., which negatively affects social participation and interaction leading to the risk of social isolation and deterioration in quality of life and health status⁴4 Nunes APN, Barreto SM, Gonçalves LG. Relações sociais e autopercepção da saúde: Projeto Envelhecimento e Saúde. Rev Bras Epidemiol 2012; 15(2):415-428.^,55 Lima-Costa MF, Peixoto SV, Malta DC, Szwarcwald CL, Mambrini JVM. Cuidado informal e remunerado aos idosos no Brasil (Pesquisa Nacional de Saúde, 2013). Rev Saúde Pública 2017; 51(Supl. 1):6s.^,77 Amoah PA. Social participation, health literacy, and health and well-being: A crosssectional study in Ghana. SSM - Population Health 2018; 4:263-270.^,88 Gilmour H. Social participation and the health and well-being of Canadian seniors. Statistics Canada, Catalogue Health Reports 2012; 23(4):1-12..

The recognition of the importance of living together in society for the health of the individual in the aging process has been highlighted in several studies and indicated that social participation (social sharing of individual resources through active involvement in collective social actions such as religious activities, hobby clubs, sports groups, cultural and political events) and maintaining interpersonal relationships are social determinants of physical and emotional health among older people⁶6 Bélanger E, Ahmed T, Vafaei A, Curcio CL, Phillips SP, Zunzunegui MV. Sources of social support associated with health and quality of life: a cross-sectional study among Canadian and Latin American older adults. BMJ Open 2016; 6:e011503.

7 Amoah PA. Social participation, health literacy, and health and well-being: A crosssectional study in Ghana. SSM - Population Health 2018; 4:263-270.

8 Gilmour H. Social participation and the health and well-being of Canadian seniors. Statistics Canada, Catalogue Health Reports 2012; 23(4):1-12.

9 White AM, Philogere S, Fine L, Sinha S. Social support and self-reported health status of older adults in the United States. Am J Public Health 2009; 99(10):1872-1878.

10 Hermingway A, Jack E. Reducing social isolation and promoting well being in older people. Qually Ageing Older Adults 2013; 14(1):25-35.

11 Thanakwang K. Social relationships influencing positive perceived health among Thai older persons: A secondary data analysis using the National Elderly Survey. Nursing Health Sci 2009; 11:144-149.^-1212 Douglas H, Georgiou A, Westrook J. Social participation as an indicador of successful aging: an overview of concepts and their associations with health. Aust Health Rev 2017; 41(4):455-462., key indicators of successful aging¹³13 Norstrand J, Chan KT. The relationship between health and community across aging cohorts. J Aging Research 2014; 626097.^,1414 Sousa NFS, Lima MG, Cesar CLG, Barros MBA. Envelhecimento ativo: prevalência e diferenças de gênero e idade em estudo de base populacional. Cad Saúde Pública 2018; 34(11):e00173317. and predictor of mortality (all-cause and cause-specific)¹⁵15 Gontijo CF, Firmo JOA, Lima-Costa MF, Loyola Filho AI. Um estudo longitudinal da associação do capital social e mortalidade entre idosos brasileiros residentes em comunidade. Cad Saúde Pública 2019; 35(2):e00056418.^,1616 Kawachi I, Kennedy BP, Lochner K, Prothrow-Stith D. Social capital, income inequality, and mortality. Am J Public Health 1997; 87(9):1491-1498.. Strong social support networks and high levels of community life engagement are protective factors for maintaining good health and quality of life and have been shown to have a positive effect on various health indicators and wide-ranging benefits for active aging and the well-being of elderly people⁶6 Bélanger E, Ahmed T, Vafaei A, Curcio CL, Phillips SP, Zunzunegui MV. Sources of social support associated with health and quality of life: a cross-sectional study among Canadian and Latin American older adults. BMJ Open 2016; 6:e011503.

7 Amoah PA. Social participation, health literacy, and health and well-being: A crosssectional study in Ghana. SSM - Population Health 2018; 4:263-270.

8 Gilmour H. Social participation and the health and well-being of Canadian seniors. Statistics Canada, Catalogue Health Reports 2012; 23(4):1-12.

9 White AM, Philogere S, Fine L, Sinha S. Social support and self-reported health status of older adults in the United States. Am J Public Health 2009; 99(10):1872-1878.^-1010 Hermingway A, Jack E. Reducing social isolation and promoting well being in older people. Qually Ageing Older Adults 2013; 14(1):25-35.^,1414 Sousa NFS, Lima MG, Cesar CLG, Barros MBA. Envelhecimento ativo: prevalência e diferenças de gênero e idade em estudo de base populacional. Cad Saúde Pública 2018; 34(11):e00173317.^,1717 Warren-Findlow J, Laditka JN, Thompsom ME, Laditka SB. Effects of social ties on self-rated physical health among African American adults. J Natl Med Assoc 2013; 105(1):23-32.

18 Dong X, Li Yu, Simon MA. Social engagement among U. S. Chinese older adults - Findings from the PINE Study. J Gerontol A Biol Med Sci 2014; 69(Supl. 2):S82-S89.

19 Siegler E, Lama SD, Knight MG, Laureano E, Reid, MC. Community-based supports and services for older adults: a primer for clinicians. J Geriatrics 2015; 678625.^-2020 Rawal I, Kaur M, Khanna P, Lakshmi PVN, Kaur N. Association between Social Capital and Self-rated Health of Older People in Chandigarh, India. Indian J Community Health 2017; 29(2):198-202..

The mechanisms explaining the relationship between social participation and health indicators have been described as capable of occurring in multiple pathways⁷7 Amoah PA. Social participation, health literacy, and health and well-being: A crosssectional study in Ghana. SSM - Population Health 2018; 4:263-270.^,88 Gilmour H. Social participation and the health and well-being of Canadian seniors. Statistics Canada, Catalogue Health Reports 2012; 23(4):1-12.^,1616 Kawachi I, Kennedy BP, Lochner K, Prothrow-Stith D. Social capital, income inequality, and mortality. Am J Public Health 1997; 87(9):1491-1498.^,2121 Yamaoka, K. Social capital and health and well-being in East Asia: A population based study. Soc Sci Med 2008; 66:885-899.. For example, the physiological impact of social isolation is hypothesized to influence the neuroendocrine and immune systems. Also, social ties may encourage individuals to engage in health-promoting behaviours such as physical activity, seeking medical care, or to refrain from damaging ones such as smoking. The psychological effects of social connectedness may include feelings of self-efficacy, a sense of meaning and purpose, and better mental health. In particular, interactions that provide social support are thought to be facilitators of physical and emotional health and well-being.

Self-assessment of overall health status (SAH) is one of the most widely used indicators for measuring the health status of elderly people and social participation in community life⁴4 Nunes APN, Barreto SM, Gonçalves LG. Relações sociais e autopercepção da saúde: Projeto Envelhecimento e Saúde. Rev Bras Epidemiol 2012; 15(2):415-428.^,66 Bélanger E, Ahmed T, Vafaei A, Curcio CL, Phillips SP, Zunzunegui MV. Sources of social support associated with health and quality of life: a cross-sectional study among Canadian and Latin American older adults. BMJ Open 2016; 6:e011503.

7 Amoah PA. Social participation, health literacy, and health and well-being: A crosssectional study in Ghana. SSM - Population Health 2018; 4:263-270.

8 Gilmour H. Social participation and the health and well-being of Canadian seniors. Statistics Canada, Catalogue Health Reports 2012; 23(4):1-12.^-99 White AM, Philogere S, Fine L, Sinha S. Social support and self-reported health status of older adults in the United States. Am J Public Health 2009; 99(10):1872-1878.^,1111 Thanakwang K. Social relationships influencing positive perceived health among Thai older persons: A secondary data analysis using the National Elderly Survey. Nursing Health Sci 2009; 11:144-149.

12 Douglas H, Georgiou A, Westrook J. Social participation as an indicador of successful aging: an overview of concepts and their associations with health. Aust Health Rev 2017; 41(4):455-462.

13 Norstrand J, Chan KT. The relationship between health and community across aging cohorts. J Aging Research 2014; 626097.

14 Sousa NFS, Lima MG, Cesar CLG, Barros MBA. Envelhecimento ativo: prevalência e diferenças de gênero e idade em estudo de base populacional. Cad Saúde Pública 2018; 34(11):e00173317.

15 Gontijo CF, Firmo JOA, Lima-Costa MF, Loyola Filho AI. Um estudo longitudinal da associação do capital social e mortalidade entre idosos brasileiros residentes em comunidade. Cad Saúde Pública 2019; 35(2):e00056418.

16 Kawachi I, Kennedy BP, Lochner K, Prothrow-Stith D. Social capital, income inequality, and mortality. Am J Public Health 1997; 87(9):1491-1498.

17 Warren-Findlow J, Laditka JN, Thompsom ME, Laditka SB. Effects of social ties on self-rated physical health among African American adults. J Natl Med Assoc 2013; 105(1):23-32.

18 Dong X, Li Yu, Simon MA. Social engagement among U. S. Chinese older adults - Findings from the PINE Study. J Gerontol A Biol Med Sci 2014; 69(Supl. 2):S82-S89.

19 Siegler E, Lama SD, Knight MG, Laureano E, Reid, MC. Community-based supports and services for older adults: a primer for clinicians. J Geriatrics 2015; 678625.^-2020 Rawal I, Kaur M, Khanna P, Lakshmi PVN, Kaur N. Association between Social Capital and Self-rated Health of Older People in Chandigarh, India. Indian J Community Health 2017; 29(2):198-202.. This indicator reveals the integrated and subjective perceptions of the elderly in relation to their health⁶6 Bélanger E, Ahmed T, Vafaei A, Curcio CL, Phillips SP, Zunzunegui MV. Sources of social support associated with health and quality of life: a cross-sectional study among Canadian and Latin American older adults. BMJ Open 2016; 6:e011503.^,22 Oliveira BLCA, Thomaz EBAF, Silva RA. The association between skin color/race and health indicators in elderly Brazilians: a study based on the Brazilian National Household Sample Survey (2008). Cad Saúde Pública 2014; 30(7):1438-1452.^,2222 Nayak S, Hubbard A, Sidney S, Syme SL. Characteristics associated with self-rated health in the CARDIA study: Contextualising health determinants by income group. Preventive Med Reports 2016; 4:199-208.^,2323 Santos SM, Chor D, Werneck GL, Coutinho ESF. Associação entre fatores contextuais e auto-avaliação de saúde: uma revisão sistemática de estudos multinível. Cad Saúde Pública 2007; 23(11):2533-2554., is an independent and strong predictor of morbidity and mortality²²22 Nayak S, Hubbard A, Sidney S, Syme SL. Characteristics associated with self-rated health in the CARDIA study: Contextualising health determinants by income group. Preventive Med Reports 2016; 4:199-208.^,2323 Santos SM, Chor D, Werneck GL, Coutinho ESF. Associação entre fatores contextuais e auto-avaliação de saúde: uma revisão sistemática de estudos multinível. Cad Saúde Pública 2007; 23(11):2533-2554. and a multifactorial marker of quality of life²²22 Nayak S, Hubbard A, Sidney S, Syme SL. Characteristics associated with self-rated health in the CARDIA study: Contextualising health determinants by income group. Preventive Med Reports 2016; 4:199-208.. However, until recently there was no nationwide data on this theme and studies of the prevalence of the social participation of old people in community activities and its association with SAH using nationally representative samples are therefore still rare in Brazil. Moreover, the data provided by the few local level studies that have been conducted³3 Pinto JM, Neri AL. Trajetórias da participação social na velhice: uma revisão sistemática da literatura. Rev Bras Geriatr Gerontol 2017; 20(2):260-273.^,44 Nunes APN, Barreto SM, Gonçalves LG. Relações sociais e autopercepção da saúde: Projeto Envelhecimento e Saúde. Rev Bras Epidemiol 2012; 15(2):415-428.^,66 Bélanger E, Ahmed T, Vafaei A, Curcio CL, Phillips SP, Zunzunegui MV. Sources of social support associated with health and quality of life: a cross-sectional study among Canadian and Latin American older adults. BMJ Open 2016; 6:e011503.^,2424 Rebellato C, Hayashi MCPI. Participação social do idoso: estudo bibliométrico da produção científica recente (2010-2013). Rev Eletron Comun Inf Inov Saúde 2014; 8(3):264-287.^,2525 Miranda LC, Banhado EFC. Qualidade de vida na terceira idade: a influência da participação em grupos. Psicol. Pesq 2008; 2(1):69-80. is generally inconsistent, partially due to the different methods and indicators used to explore social interaction and interpersonal relationships³3 Pinto JM, Neri AL. Trajetórias da participação social na velhice: uma revisão sistemática da literatura. Rev Bras Geriatr Gerontol 2017; 20(2):260-273.^,44 Nunes APN, Barreto SM, Gonçalves LG. Relações sociais e autopercepção da saúde: Projeto Envelhecimento e Saúde. Rev Bras Epidemiol 2012; 15(2):415-428.. Also, social participation among the elderly is not a random phenomenon, as it depends on a set of individual and contextual socioeconomic, demographic and health characteristics. Thus, controlling the systematic differences in the distribution of attributes among the elderly groups and focusing directly on the determinants of social participation are still methodological issues that research on the relationship between social participation and health states need to address⁶6 Bélanger E, Ahmed T, Vafaei A, Curcio CL, Phillips SP, Zunzunegui MV. Sources of social support associated with health and quality of life: a cross-sectional study among Canadian and Latin American older adults. BMJ Open 2016; 6:e011503.^,77 Amoah PA. Social participation, health literacy, and health and well-being: A crosssectional study in Ghana. SSM - Population Health 2018; 4:263-270.^,2222 Nayak S, Hubbard A, Sidney S, Syme SL. Characteristics associated with self-rated health in the CARDIA study: Contextualising health determinants by income group. Preventive Med Reports 2016; 4:199-208.. Thus, it is important to adopt more robust methods of analysis, such as propensity score weighting, which provides a more accurate estimate of prevalence and the association between given variables in comparison groups with different individual and contextual characteristics²⁶26 Andrade ACS, Mingoli SA, Fernandes AP, Andrade RG, Friche AAL, Xavier CC, Proietti FA, Diez-Roux AV, Caiaffa WT. Neighborhood-based physical activity differences: Evaluation of the effect of health promotion program. Plos One 2018; 13(2):e0192115.^,2727 Sousa-Júnior PRB, Freitas MPS, Antonaci GA, Szwarcwald CL. Desenho da amostra da Pesquisa Nacional de Saúde 2013. Epidemiol Serv Saúde 2015; 24(2):207-216..

This because, it is possible that there are important differences between the profiles of the elderly who have social participation in activities in the community of those who do not participate, and that this participation can positively favor the engagement of the elderly to healthier lifestyles and the self-assessment health.

This present study used data from the 2013 National Health Survey (NHS) to explore the prevalence of social participation in community activities (exposed) and its association with positive self-assessment of overall health (outcome) status among older people in Brazil.

Methods

Data Source

The 2013 National Health Survey (NHS) was conducted by the Brazilian Institute of Geography and Statistics (IBGE, acronym in Portuguese) in partnership with the Ministry of Health²⁷27 Sousa-Júnior PRB, Freitas MPS, Antonaci GA, Szwarcwald CL. Desenho da amostra da Pesquisa Nacional de Saúde 2013. Epidemiol Serv Saúde 2015; 24(2):207-216.^,2828 Szwarcwald CL, Malta DC, Pereira CA, Vieira MLFP, Conde WL, Souza Júnior PRB, Damacena GN, Azevedo LO, Silva GA, Theme Filha MM, Lopes CS, Romero DE, Almeida WS, Monteiro CA. Pesquisa Nacional de Saúde no Brasil: concepção e metodologia de aplicação. Cien Saude Colet 2014; 19(2):333-342..

NHS sampling plan

NHS sampling plan was developed by conglomerates in three stages of selection. In the frst stage, Primary Care Units (UPA) were selected by simple random sampling, consisting of censos tracts or sets of census tracts (when the tracts counted with few households). In the second stage, a fixed number of households were selected by simple random sampling for each UPA (from 10 to 14). In each household sampled, one resident with ≥18 years was selected, also by simple random sampling, to take part in the third stage of selection²⁷27 Sousa-Júnior PRB, Freitas MPS, Antonaci GA, Szwarcwald CL. Desenho da amostra da Pesquisa Nacional de Saúde 2013. Epidemiol Serv Saúde 2015; 24(2):207-216.^,2828 Szwarcwald CL, Malta DC, Pereira CA, Vieira MLFP, Conde WL, Souza Júnior PRB, Damacena GN, Azevedo LO, Silva GA, Theme Filha MM, Lopes CS, Romero DE, Almeida WS, Monteiro CA. Pesquisa Nacional de Saúde no Brasil: concepção e metodologia de aplicação. Cien Saude Colet 2014; 19(2):333-342..

Sample was calculated at approximately 80,000 households. Information were collected on 62,986 households. The calculation took into account average values, variances, and the effects of the sampling plan, assuming a nonresponse rate of 20%²⁷27 Sousa-Júnior PRB, Freitas MPS, Antonaci GA, Szwarcwald CL. Desenho da amostra da Pesquisa Nacional de Saúde 2013. Epidemiol Serv Saúde 2015; 24(2):207-216.^,2828 Szwarcwald CL, Malta DC, Pereira CA, Vieira MLFP, Conde WL, Souza Júnior PRB, Damacena GN, Azevedo LO, Silva GA, Theme Filha MM, Lopes CS, Romero DE, Almeida WS, Monteiro CA. Pesquisa Nacional de Saúde no Brasil: concepção e metodologia de aplicação. Cien Saude Colet 2014; 19(2):333-342.. The weights of the households and all its residents were calculated by the product of the weight of the UPA in question and the inverse of the probability of selection of the household within the UPA. The weights were adjusted to correct nonresponses and to calibrate the estimates according to population totals known from other sources. The selection of the resident who answered to the individual interview was done by simple random sampling. Thus, the weight of the selected resident was calculated by the product of the weight of the household by the number of eligible residents (equivalent to the inverse of the probability of selection). More details can be consulted in another study²⁷27 Sousa-Júnior PRB, Freitas MPS, Antonaci GA, Szwarcwald CL. Desenho da amostra da Pesquisa Nacional de Saúde 2013. Epidemiol Serv Saúde 2015; 24(2):207-216.^,2828 Szwarcwald CL, Malta DC, Pereira CA, Vieira MLFP, Conde WL, Souza Júnior PRB, Damacena GN, Azevedo LO, Silva GA, Theme Filha MM, Lopes CS, Romero DE, Almeida WS, Monteiro CA. Pesquisa Nacional de Saúde no Brasil: concepção e metodologia de aplicação. Cien Saude Colet 2014; 19(2):333-342..

Data collection was carried out with the use of handheld computers (personal digital assistant), programmed to critique the received values. This household survey consists of three questionnaires designed to obtain information on the living conditions and health of the Brazilian population: a household questionnaire; an individual questionnaire, answered by all household members; and a second individual questionnaire answered by a randomly selected sample of the adult household members (≥18 years)²⁷27 Sousa-Júnior PRB, Freitas MPS, Antonaci GA, Szwarcwald CL. Desenho da amostra da Pesquisa Nacional de Saúde 2013. Epidemiol Serv Saúde 2015; 24(2):207-216.. The present study uses the data generated by the latter questionnaire applied to the randomly selected adult household members who were aged ≥65 years (n=7,712).

Exposure and outcome variables

To measure the effect of social participation, two comparison groups were created: (a) the exposed group, comprising older people (n=1,906) who reported that they participate in community activities; and (b) the unexposed group, comprising those who reported that they do not participate in community activities (n=5,806). Social participation in community activities was determined based on the response to the following question in the NHS questionnaire: “Do you participate in organized social activities, such as clubs, community or religious groups, social centers for the elderly, etc.?” (Yes=1 or No=0).

The health outcome used for the purpose of this study was self-assessment of overall health status (SAH) based on the following response categories: very good, good, normal, bad, and very bad. For the purposes of this study the indicator was dichotomized into negative self-assessment (normal/bad/very bad) and positive self-assessment (very good/good), in accordance with conventional practice⁶6 Bélanger E, Ahmed T, Vafaei A, Curcio CL, Phillips SP, Zunzunegui MV. Sources of social support associated with health and quality of life: a cross-sectional study among Canadian and Latin American older adults. BMJ Open 2016; 6:e011503.^,2222 Nayak S, Hubbard A, Sidney S, Syme SL. Characteristics associated with self-rated health in the CARDIA study: Contextualising health determinants by income group. Preventive Med Reports 2016; 4:199-208.^,2323 Santos SM, Chor D, Werneck GL, Coutinho ESF. Associação entre fatores contextuais e auto-avaliação de saúde: uma revisão sistemática de estudos multinível. Cad Saúde Pública 2007; 23(11):2533-2554..

Control Covariates

Control covariates included sex (male, female); age (years); color/race (white, nonwhite); marital status (married, single); household reference person (yes, no); level of schooling (no schooling/primary incomplete, primary completed/higher incomplete, higher complete, ignored); currently working (yes, no); total household income in R$ (in quintiles); physical, intellectual, hearing, or visual disability (yes, no); number of chronic diseases (physical or mental); health plan (yes, no); household registered in the Family Health Strategy (yes, no); household location (urban, rural); macro region (North, Northeast, Center-West, Southeast, and South); number of residents in the household; number of household appliances in the household; number of vehicles in the household. Two interaction terms were also added: household registered in the FHS (Family Health Strategy) with urban house hold, and age with person with a chronic disease more health plan. These terms were tested considering theoretical conceptions about the joint influences of these variables in the exposure to participation in community activities by the elderly. As well, interaction terms have been used to increase the estimated probability in propensity score and so the common support area²⁶26 Andrade ACS, Mingoli SA, Fernandes AP, Andrade RG, Friche AAL, Xavier CC, Proietti FA, Diez-Roux AV, Caiaffa WT. Neighborhood-based physical activity differences: Evaluation of the effect of health promotion program. Plos One 2018; 13(2):e0192115.^,2929 Pan W, Bai H. Propensity Score Analysis. In: Pan W, Bai H, organizadores. Propensity Score Analysis: fundamentals and developments. Nova York: The Guilford press; 2015. p. 3-19..

Statiscal Analysis

Two-stage propensity score weighting was used to control for lack of group homogeneity in relation to individual and contextual socioeconomic, demographic, and health covariates. Initially, the propensity score (PS) was taken as the conditional probability of an individual being exposed (receiving treatment), or for the purposes of this study participating in community activities, based on the observed covariates²⁶26 Andrade ACS, Mingoli SA, Fernandes AP, Andrade RG, Friche AAL, Xavier CC, Proietti FA, Diez-Roux AV, Caiaffa WT. Neighborhood-based physical activity differences: Evaluation of the effect of health promotion program. Plos One 2018; 13(2):e0192115.^,2929 Pan W, Bai H. Propensity Score Analysis. In: Pan W, Bai H, organizadores. Propensity Score Analysis: fundamentals and developments. Nova York: The Guilford press; 2015. p. 3-19.. It is represented by a single variable that simultaneously considers all the potential confounding variables. Thus, individuals with the same propensity score have the same distribution of observed covariates, regardless of their exposure conditions²⁹29 Pan W, Bai H. Propensity Score Analysis. In: Pan W, Bai H, organizadores. Propensity Score Analysis: fundamentals and developments. Nova York: The Guilford press; 2015. p. 3-19.

30 Oakes JM, Johnson P. Propensity score matching for social epidemiology. In: Oakes JM, Johnson P, organizadores. Methods in social epidemiology. Nova Jersey: Jossey-Bass, A Wiley Imprint; 2006. p. 370-392.

31 Hahs-Vaughn DL. Propensity Score Analysis with complex survey samples. In: Pan W, Bai H, organizadores. Propensity score analysis: fundamentals and developments. Nova York: The Guilford press; 2015. p. 237-264.

32 Patorno E, Grotta A, Bellocco R, Schneeweiss S. Propensity score methoology for confouding control in health care utilization databases. Epidemiol Biostat Public Health 2013; 10(3):8940-8916.^-3333 Hernan MA, Robins JM. Causal Inference. Boca Raton: Chapman & Hall/CRC; 2017..

The PS was estimated using binary logistic regression with the maximum likelihood method. This probability ranges from 0 to 1. Thus, each sample member had a conditional probability of (a propensity for) being exposed (social participation in community activities) based on the set of covariates tested in the proposed model. Thus, it is sought to reduce the dimensionality of a set of confounders to a single measure, and it is allowed that the units of analysis with similar EP have, on average, similar probabilities of receiving the treatment and distribution of the covariates. In the logistic regression that estimated the PS, it was decided for using the set of variables based on the theoretical conception of their importance. This method of estimation of the propensity score has been used by other authors²⁶26 Andrade ACS, Mingoli SA, Fernandes AP, Andrade RG, Friche AAL, Xavier CC, Proietti FA, Diez-Roux AV, Caiaffa WT. Neighborhood-based physical activity differences: Evaluation of the effect of health promotion program. Plos One 2018; 13(2):e0192115.^,2929 Pan W, Bai H. Propensity Score Analysis. In: Pan W, Bai H, organizadores. Propensity Score Analysis: fundamentals and developments. Nova York: The Guilford press; 2015. p. 3-19.

30 Oakes JM, Johnson P. Propensity score matching for social epidemiology. In: Oakes JM, Johnson P, organizadores. Methods in social epidemiology. Nova Jersey: Jossey-Bass, A Wiley Imprint; 2006. p. 370-392.

31 Hahs-Vaughn DL. Propensity Score Analysis with complex survey samples. In: Pan W, Bai H, organizadores. Propensity score analysis: fundamentals and developments. Nova York: The Guilford press; 2015. p. 237-264.^-3232 Patorno E, Grotta A, Bellocco R, Schneeweiss S. Propensity score methoology for confouding control in health care utilization databases. Epidemiol Biostat Public Health 2013; 10(3):8940-8916..

With this, after estimating the PS, this was then used to calculate the inverse probability of selection (IPS), applying a probability of selection to each sample member (1/IPS those the exposed group and 1/(1 - IPS) for unexposed group to make the two groups homogenous and comparable³⁰30 Oakes JM, Johnson P. Propensity score matching for social epidemiology. In: Oakes JM, Johnson P, organizadores. Methods in social epidemiology. Nova Jersey: Jossey-Bass, A Wiley Imprint; 2006. p. 370-392.

31 Hahs-Vaughn DL. Propensity Score Analysis with complex survey samples. In: Pan W, Bai H, organizadores. Propensity score analysis: fundamentals and developments. Nova York: The Guilford press; 2015. p. 237-264.

32 Patorno E, Grotta A, Bellocco R, Schneeweiss S. Propensity score methoology for confouding control in health care utilization databases. Epidemiol Biostat Public Health 2013; 10(3):8940-8916.

33 Hernan MA, Robins JM. Causal Inference. Boca Raton: Chapman & Hall/CRC; 2017.^-3434 D'Agostino-Júnior RB. Tutorial in biostatistics: propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Statist Med 1998; 17:2265-2281.. As it was verified extreme weights derived from the PS, there was considered the use of stabilized weights in the construction of the IPS. For both groups we estimated the proportion (mean of numeric variables) and standard error of the selected covariates to make up the model for estimating the PS before and after inverse probability weighting in order to determine the pattern of distribution of covariates across the groups. In order to verify the imbalance of the covariate before and after weighting²⁹29 Pan W, Bai H. Propensity Score Analysis. In: Pan W, Bai H, organizadores. Propensity Score Analysis: fundamentals and developments. Nova York: The Guilford press; 2015. p. 3-19.^,3131 Hahs-Vaughn DL. Propensity Score Analysis with complex survey samples. In: Pan W, Bai H, organizadores. Propensity score analysis: fundamentals and developments. Nova York: The Guilford press; 2015. p. 237-264. it was estimated the standardized mean diferences, and variance analysis was performed using the F-test, where a p-value of >0.05 is considered to show homogeneity of variance³⁰30 Oakes JM, Johnson P. Propensity score matching for social epidemiology. In: Oakes JM, Johnson P, organizadores. Methods in social epidemiology. Nova Jersey: Jossey-Bass, A Wiley Imprint; 2006. p. 370-392.^,3131 Hahs-Vaughn DL. Propensity Score Analysis with complex survey samples. In: Pan W, Bai H, organizadores. Propensity score analysis: fundamentals and developments. Nova York: The Guilford press; 2015. p. 237-264.. Box plots were elaborated to illustrate the pattern of distribution of estimated probabilities before and after weighting³⁰30 Oakes JM, Johnson P. Propensity score matching for social epidemiology. In: Oakes JM, Johnson P, organizadores. Methods in social epidemiology. Nova Jersey: Jossey-Bass, A Wiley Imprint; 2006. p. 370-392.^,3131 Hahs-Vaughn DL. Propensity Score Analysis with complex survey samples. In: Pan W, Bai H, organizadores. Propensity score analysis: fundamentals and developments. Nova York: The Guilford press; 2015. p. 237-264.. The F-test statistic has been widely used as a way to verify the differences in the variances between the comparison groups, and to measure the loss and the extent of homogeneity between them²⁹29 Pan W, Bai H. Propensity Score Analysis. In: Pan W, Bai H, organizadores. Propensity Score Analysis: fundamentals and developments. Nova York: The Guilford press; 2015. p. 3-19.

30 Oakes JM, Johnson P. Propensity score matching for social epidemiology. In: Oakes JM, Johnson P, organizadores. Methods in social epidemiology. Nova Jersey: Jossey-Bass, A Wiley Imprint; 2006. p. 370-392.^-3131 Hahs-Vaughn DL. Propensity Score Analysis with complex survey samples. In: Pan W, Bai H, organizadores. Propensity score analysis: fundamentals and developments. Nova York: The Guilford press; 2015. p. 237-264.^,3434 D'Agostino-Júnior RB. Tutorial in biostatistics: propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Statist Med 1998; 17:2265-2281..

Prevalence and 95% confidence intervals (CI95%) were estimated for positive self-assessment of overall health status according to variable exposure before and after weighting. Differences in the distribution of these frequencies were estimated using Pearson’s chi-squared test with a significance level of p<0,05.

The association between social participation and positive self-assessment of overall health status (SAH) was estimated based prevalence ratio (PR) and respective 95% confidence intervals (CI95%) calculated using Poisson regression (crude and weighted by PS).

The analyses were performed using the SPSS® software package (version 23, SPSS Inc, Chicago, Illinois), which incorporates the effects of complex sample designs such as that of the 2013 NHS into all stages of analysis, including the estimation of the propensity score.

Ethical Aspects

The NHS was approved by the National Research Ethics Committee and all participants signed an informed consent form²⁷27 Sousa-Júnior PRB, Freitas MPS, Antonaci GA, Szwarcwald CL. Desenho da amostra da Pesquisa Nacional de Saúde 2013. Epidemiol Serv Saúde 2015; 24(2):207-216..

Results

The average age of the sample (n=7,712 older people ≥65 years) was 72 years (68-78), while the prevalence of social participation was 25.1% (CI95%: 23.4-26.9). In relation to the group of elderly people exposed to social participation, those not exposed presented a higher proportion of men, older people (≥80 years), nonwhite, no married, and household references, and were more likely to have primary completed/higher incomplete schooling, no work, live with other people, be in the lowest income quintiles, have a lower average number of chronic diseases, not have a health plan, live in rural areas and the Northeast of the country, and not own vehicles, most likely to be registered in the FHS and lower ownership of health plans (Table 1). Significant differences were observed in the standardized means differences before PS weighting in most categories of covariates, but differences were strongly reduced after PS weighting, reaching values close to zero. As well, the results of the F-test showed a reduction in the magnitude and loss of statistical significance of covariance across the two groups after weighting. Homogeneity of variance was shown for the majority of covariates that were imbalanced before weighting. But, since the sample size is large, the possibility of obtaining a significant F-test should be considered (Table 1).

Less than half of the sample gave a positive assessment of their overall health status. However, the prevalence of positive assessment was significantly higher in the exposed group. After weighting, the prevalence of positive self-assessment changed from 47.7% (CI95%: 44.0-51.5; p=0.002) for 48.0% (CI95%: 46.0-51.0; p<0.001), without considering the differences in the distribution of covariates between the two groups (Table 2).

Figure 1 demonstrates that the PS has a similar distribution across both groups, confirming that propensity score weighting balanced the socioeconomic, demographic and health variables across the two groups.

Finally, the results obtained from both the crude and weighted logistic regression models show that there was a positive association between participation in community activities and positive health assessment. However, the magnitude of the association obtained using the weighted model (PR: 1.15; CI95%: 1.08-1.22) indicates that it is necessary an adjustment of measure of the association obtained by the crude model (PR: 1.21; CI95%: 1.14-1.28). Thus, considering the homogeneity of variance across groups after weighting, the exposed group were 15% more likely to provide a positive assessment of their overall health status (Table 3).

Discussion

Around a quarter of the elderly people mentioned that they participated in community activities and those that participated these activities were more likely to provide a positive assessment of their overall health status than those who were unexposed to these activities. In concordance with the results of other studies, the present study observed important differences in the individual and contextual socioeconomic, demographic, and health characteristics of older people who participate in community activities and those who do not³3 Pinto JM, Neri AL. Trajetórias da participação social na velhice: uma revisão sistemática da literatura. Rev Bras Geriatr Gerontol 2017; 20(2):260-273.^,66 Bélanger E, Ahmed T, Vafaei A, Curcio CL, Phillips SP, Zunzunegui MV. Sources of social support associated with health and quality of life: a cross-sectional study among Canadian and Latin American older adults. BMJ Open 2016; 6:e011503.^,88 Gilmour H. Social participation and the health and well-being of Canadian seniors. Statistics Canada, Catalogue Health Reports 2012; 23(4):1-12.^,99 White AM, Philogere S, Fine L, Sinha S. Social support and self-reported health status of older adults in the United States. Am J Public Health 2009; 99(10):1872-1878.^,1212 Douglas H, Georgiou A, Westrook J. Social participation as an indicador of successful aging: an overview of concepts and their associations with health. Aust Health Rev 2017; 41(4):455-462.^,1414 Sousa NFS, Lima MG, Cesar CLG, Barros MBA. Envelhecimento ativo: prevalência e diferenças de gênero e idade em estudo de base populacional. Cad Saúde Pública 2018; 34(11):e00173317.^,1818 Dong X, Li Yu, Simon MA. Social engagement among U. S. Chinese older adults - Findings from the PINE Study. J Gerontol A Biol Med Sci 2014; 69(Supl. 2):S82-S89.^,1919 Siegler E, Lama SD, Knight MG, Laureano E, Reid, MC. Community-based supports and services for older adults: a primer for clinicians. J Geriatrics 2015; 678625.^,3535 Ferreira, FR, César CC, Andrade FB, Souza Junior PRB, Lima-Costa MF, Proietti FA. Aspects of social participation and neighborhood perception: ELSI-Brazil. Rev Saude Publica 2018; 52(Supl. 2):18s., suggesting common characteristics associated with this practice. In general, it was observed that unexposed older people had worse socioeconomic conditions, depended more on public transport and health services, and lived in settings associated with greater material deprivation and social infrastructure. Studies show that older people who are more vulnerable from socioeconomic and health point of view are more likely to live in areas where the access to such services and opportunities is irregular and precarious⁹9 White AM, Philogere S, Fine L, Sinha S. Social support and self-reported health status of older adults in the United States. Am J Public Health 2009; 99(10):1872-1878.^,1212 Douglas H, Georgiou A, Westrook J. Social participation as an indicador of successful aging: an overview of concepts and their associations with health. Aust Health Rev 2017; 41(4):455-462.^,1313 Norstrand J, Chan KT. The relationship between health and community across aging cohorts. J Aging Research 2014; 626097.^,3535 Ferreira, FR, César CC, Andrade FB, Souza Junior PRB, Lima-Costa MF, Proietti FA. Aspects of social participation and neighborhood perception: ELSI-Brazil. Rev Saude Publica 2018; 52(Supl. 2):18s.^,3636 Diez-Roux AV, Mair C. Neighborhoods and health. Ann N Y Acad Sci 2010; 1186:125-145.. Furthermore, cumulative disadvantage over the life course can hinder adhesion to or maintenance of social participation in community activities in old age⁶6 Bélanger E, Ahmed T, Vafaei A, Curcio CL, Phillips SP, Zunzunegui MV. Sources of social support associated with health and quality of life: a cross-sectional study among Canadian and Latin American older adults. BMJ Open 2016; 6:e011503.^,1111 Thanakwang K. Social relationships influencing positive perceived health among Thai older persons: A secondary data analysis using the National Elderly Survey. Nursing Health Sci 2009; 11:144-149..

The prevalence of social participation in community activities still varies widely between countries. In the present study, this prevalence was lower than that found by a study of elderly Chinese people living in the United States¹⁸18 Dong X, Li Yu, Simon MA. Social engagement among U. S. Chinese older adults - Findings from the PINE Study. J Gerontol A Biol Med Sci 2014; 69(Supl. 2):S82-S89., elderly from Ghana⁷7 Amoah PA. Social participation, health literacy, and health and well-being: A crosssectional study in Ghana. SSM - Population Health 2018; 4:263-270., Canada⁶6 Bélanger E, Ahmed T, Vafaei A, Curcio CL, Phillips SP, Zunzunegui MV. Sources of social support associated with health and quality of life: a cross-sectional study among Canadian and Latin American older adults. BMJ Open 2016; 6:e011503.^,88 Gilmour H. Social participation and the health and well-being of Canadian seniors. Statistics Canada, Catalogue Health Reports 2012; 23(4):1-12., United States¹³13 Norstrand J, Chan KT. The relationship between health and community across aging cohorts. J Aging Research 2014; 626097., Japan, Australia, Germany, the United Kingdom, Switzerland and Denmark, albeit showing that social participation declined over time and with age³3 Pinto JM, Neri AL. Trajetórias da participação social na velhice: uma revisão sistemática da literatura. Rev Bras Geriatr Gerontol 2017; 20(2):260-273.^,66 Bélanger E, Ahmed T, Vafaei A, Curcio CL, Phillips SP, Zunzunegui MV. Sources of social support associated with health and quality of life: a cross-sectional study among Canadian and Latin American older adults. BMJ Open 2016; 6:e011503.

7 Amoah PA. Social participation, health literacy, and health and well-being: A crosssectional study in Ghana. SSM - Population Health 2018; 4:263-270.^-88 Gilmour H. Social participation and the health and well-being of Canadian seniors. Statistics Canada, Catalogue Health Reports 2012; 23(4):1-12.^,2424 Rebellato C, Hayashi MCPI. Participação social do idoso: estudo bibliométrico da produção científica recente (2010-2013). Rev Eletron Comun Inf Inov Saúde 2014; 8(3):264-287. and varies by sex⁸8 Gilmour H. Social participation and the health and well-being of Canadian seniors. Statistics Canada, Catalogue Health Reports 2012; 23(4):1-12..

In Brazil, few studies have investigated social engagement and participation among the elderly. A study conducted in 2007 with 361 older people (≥60 years) in a health district in the southeast of the city of Belo Horizonte observed that 26% reported participating in group recreation or artistic activities⁴4 Nunes APN, Barreto SM, Gonçalves LG. Relações sociais e autopercepção da saúde: Projeto Envelhecimento e Saúde. Rev Bras Epidemiol 2012; 15(2):415-428., while a study carried out in 2008 with 89 older persons living in Juiz de Fora in the State of Minas Gerais (MG) showed that 52.8% were active members of community groups²⁵25 Miranda LC, Banhado EFC. Qualidade de vida na terceira idade: a influência da participação em grupos. Psicol. Pesq 2008; 2(1):69-80.. More recently, a study included 935 surviving elderly from the cohort of the city of Bambuí-MG, presented a prevalence of 17.4%¹⁵15 Gontijo CF, Firmo JOA, Lima-Costa MF, Loyola Filho AI. Um estudo longitudinal da associação do capital social e mortalidade entre idosos brasileiros residentes em comunidade. Cad Saúde Pública 2019; 35(2):e00056418.. Other study analyzed prevalence rates and gender and age differences in indicators of active aging in elders participating in the Campinas Municipal Health Survey in Campinas, São Paulo State (2014-2015). The prevalence varied from 23.3% and 89.0% depending on the type of activity, but it only differed statistically between the sexes in participation in religious activity (p<0.001), between age groups no differences¹⁴14 Sousa NFS, Lima MG, Cesar CLG, Barros MBA. Envelhecimento ativo: prevalência e diferenças de gênero e idade em estudo de base populacional. Cad Saúde Pública 2018; 34(11):e00173317.. Lastly, study conducted in a representative sample of the Brazilian population aged 50 years and older and who lived in urban areas (n=7,935) observed prevalence of social participation 81.8% for women and 82.0% for men³⁵35 Ferreira, FR, César CC, Andrade FB, Souza Junior PRB, Lima-Costa MF, Proietti FA. Aspects of social participation and neighborhood perception: ELSI-Brazil. Rev Saude Publica 2018; 52(Supl. 2):18s..

Such differences could be explained not only by differences in values, beliefs and social norms across countries, regions, states and cities but also by access to community services, elderly care and support networks, which can vary considerably by context. Another explanation may be differences in type of activities, in the criteria and research methods used to determine social participation, which can fragment or aggregate the dimensions of social life and result in differing prevalence rates³3 Pinto JM, Neri AL. Trajetórias da participação social na velhice: uma revisão sistemática da literatura. Rev Bras Geriatr Gerontol 2017; 20(2):260-273.^,1212 Douglas H, Georgiou A, Westrook J. Social participation as an indicador of successful aging: an overview of concepts and their associations with health. Aust Health Rev 2017; 41(4):455-462.^,3535 Ferreira, FR, César CC, Andrade FB, Souza Junior PRB, Lima-Costa MF, Proietti FA. Aspects of social participation and neighborhood perception: ELSI-Brazil. Rev Saude Publica 2018; 52(Supl. 2):18s.. However, despite variations between countries that may affect the interpretation of our findings, the present study was the first of its kind in Brazil: to estimate the prevalence of social participation using a nationally representative sample of urban and rural elderly; to use propensity score weighting to overcome observed differences in homogeneity across study groups; and to capture the association between social participation and positive SAH.

SAH status has been used by different studies as a key indicator for assessing and monitoring old people’s health²2 Oliveira BLCA, Thomaz EBAF, Silva RA. The association between skin color/race and health indicators in elderly Brazilians: a study based on the Brazilian National Household Sample Survey (2008). Cad Saúde Pública 2014; 30(7):1438-1452.^,66 Bélanger E, Ahmed T, Vafaei A, Curcio CL, Phillips SP, Zunzunegui MV. Sources of social support associated with health and quality of life: a cross-sectional study among Canadian and Latin American older adults. BMJ Open 2016; 6:e011503.^,1616 Kawachi I, Kennedy BP, Lochner K, Prothrow-Stith D. Social capital, income inequality, and mortality. Am J Public Health 1997; 87(9):1491-1498.. Studies conducted in other countries have also shown the association between SAH and social participation³3 Pinto JM, Neri AL. Trajetórias da participação social na velhice: uma revisão sistemática da literatura. Rev Bras Geriatr Gerontol 2017; 20(2):260-273.^,66 Bélanger E, Ahmed T, Vafaei A, Curcio CL, Phillips SP, Zunzunegui MV. Sources of social support associated with health and quality of life: a cross-sectional study among Canadian and Latin American older adults. BMJ Open 2016; 6:e011503.

7 Amoah PA. Social participation, health literacy, and health and well-being: A crosssectional study in Ghana. SSM - Population Health 2018; 4:263-270.

8 Gilmour H. Social participation and the health and well-being of Canadian seniors. Statistics Canada, Catalogue Health Reports 2012; 23(4):1-12.^-99 White AM, Philogere S, Fine L, Sinha S. Social support and self-reported health status of older adults in the United States. Am J Public Health 2009; 99(10):1872-1878.^,1313 Norstrand J, Chan KT. The relationship between health and community across aging cohorts. J Aging Research 2014; 626097.^,1717 Warren-Findlow J, Laditka JN, Thompsom ME, Laditka SB. Effects of social ties on self-rated physical health among African American adults. J Natl Med Assoc 2013; 105(1):23-32.

18 Dong X, Li Yu, Simon MA. Social engagement among U. S. Chinese older adults - Findings from the PINE Study. J Gerontol A Biol Med Sci 2014; 69(Supl. 2):S82-S89.

19 Siegler E, Lama SD, Knight MG, Laureano E, Reid, MC. Community-based supports and services for older adults: a primer for clinicians. J Geriatrics 2015; 678625.^-2020 Rawal I, Kaur M, Khanna P, Lakshmi PVN, Kaur N. Association between Social Capital and Self-rated Health of Older People in Chandigarh, India. Indian J Community Health 2017; 29(2):198-202.. The prevalence of positive SAH in this study was similar the other research²2 Oliveira BLCA, Thomaz EBAF, Silva RA. The association between skin color/race and health indicators in elderly Brazilians: a study based on the Brazilian National Household Sample Survey (2008). Cad Saúde Pública 2014; 30(7):1438-1452.^,44 Nunes APN, Barreto SM, Gonçalves LG. Relações sociais e autopercepção da saúde: Projeto Envelhecimento e Saúde. Rev Bras Epidemiol 2012; 15(2):415-428.^,66 Bélanger E, Ahmed T, Vafaei A, Curcio CL, Phillips SP, Zunzunegui MV. Sources of social support associated with health and quality of life: a cross-sectional study among Canadian and Latin American older adults. BMJ Open 2016; 6:e011503., but our study points similar the international studies that the prevalence of positive SAH was higher among the elderly who participate in social activities in the community⁶6 Bélanger E, Ahmed T, Vafaei A, Curcio CL, Phillips SP, Zunzunegui MV. Sources of social support associated with health and quality of life: a cross-sectional study among Canadian and Latin American older adults. BMJ Open 2016; 6:e011503.^,77 Amoah PA. Social participation, health literacy, and health and well-being: A crosssectional study in Ghana. SSM - Population Health 2018; 4:263-270.^,88 Gilmour H. Social participation and the health and well-being of Canadian seniors. Statistics Canada, Catalogue Health Reports 2012; 23(4):1-12.^,1111 Thanakwang K. Social relationships influencing positive perceived health among Thai older persons: A secondary data analysis using the National Elderly Survey. Nursing Health Sci 2009; 11:144-149.. These results are also consistent with researches that the social engagement and the nature of social relations positively affect the overall health status and well-being of the elderly independent of individual and contextual socioeconomic, demographic and health factors.

Between the explanations for this association are that high levels of social support (interactions with friends and relatives and frequenting social centers and community groups, religions, organizations or civil associations) can act to improve quality of life and are protective factor for health⁶6 Bélanger E, Ahmed T, Vafaei A, Curcio CL, Phillips SP, Zunzunegui MV. Sources of social support associated with health and quality of life: a cross-sectional study among Canadian and Latin American older adults. BMJ Open 2016; 6:e011503.^,33 Pinto JM, Neri AL. Trajetórias da participação social na velhice: uma revisão sistemática da literatura. Rev Bras Geriatr Gerontol 2017; 20(2):260-273.^,1111 Thanakwang K. Social relationships influencing positive perceived health among Thai older persons: A secondary data analysis using the National Elderly Survey. Nursing Health Sci 2009; 11:144-149.. It is argued that active social participation in associational activities may promote access to vital health-related information, which may enhance health-promoting choices and SAH⁷7 Amoah PA. Social participation, health literacy, and health and well-being: A crosssectional study in Ghana. SSM - Population Health 2018; 4:263-270.. Socially isolated individuals have at increased risk for poor health outcomes because of their limited access to resources such as medical care, information, and emotional support²¹21 Yamaoka, K. Social capital and health and well-being in East Asia: A population based study. Soc Sci Med 2008; 66:885-899..

Formal or informal local community support provides access to resources, good, services and assistance that can complement or meet the needs of the elderly and their families³3 Pinto JM, Neri AL. Trajetórias da participação social na velhice: uma revisão sistemática da literatura. Rev Bras Geriatr Gerontol 2017; 20(2):260-273.^,1313 Norstrand J, Chan KT. The relationship between health and community across aging cohorts. J Aging Research 2014; 626097.^,3636 Diez-Roux AV, Mair C. Neighborhoods and health. Ann N Y Acad Sci 2010; 1186:125-145.. Limited social support can therefore mean that elderly people have less capacity to cope with adversity and lead to negative health behaviors (for example, social isolation, physical inactivity, functional disability, and depression)³3 Pinto JM, Neri AL. Trajetórias da participação social na velhice: uma revisão sistemática da literatura. Rev Bras Geriatr Gerontol 2017; 20(2):260-273.^,66 Bélanger E, Ahmed T, Vafaei A, Curcio CL, Phillips SP, Zunzunegui MV. Sources of social support associated with health and quality of life: a cross-sectional study among Canadian and Latin American older adults. BMJ Open 2016; 6:e011503.^,1313 Norstrand J, Chan KT. The relationship between health and community across aging cohorts. J Aging Research 2014; 626097.^,3636 Diez-Roux AV, Mair C. Neighborhoods and health. Ann N Y Acad Sci 2010; 1186:125-145. until all-cause and cause-specific mortality¹⁵15 Gontijo CF, Firmo JOA, Lima-Costa MF, Loyola Filho AI. Um estudo longitudinal da associação do capital social e mortalidade entre idosos brasileiros residentes em comunidade. Cad Saúde Pública 2019; 35(2):e00056418.^,1616 Kawachi I, Kennedy BP, Lochner K, Prothrow-Stith D. Social capital, income inequality, and mortality. Am J Public Health 1997; 87(9):1491-1498.. Social participation in community activities can strengthen social support networks, even among people from more underprivileged backgrounds, mitigating hardship and acting as a protective factor against health¹²12 Douglas H, Georgiou A, Westrook J. Social participation as an indicador of successful aging: an overview of concepts and their associations with health. Aust Health Rev 2017; 41(4):455-462.^,1313 Norstrand J, Chan KT. The relationship between health and community across aging cohorts. J Aging Research 2014; 626097.^,2020 Rawal I, Kaur M, Khanna P, Lakshmi PVN, Kaur N. Association between Social Capital and Self-rated Health of Older People in Chandigarh, India. Indian J Community Health 2017; 29(2):198-202.^,3636 Diez-Roux AV, Mair C. Neighborhoods and health. Ann N Y Acad Sci 2010; 1186:125-145.. However, the majority of studies remain based on stratified analysis and multivariate regression⁶6 Bélanger E, Ahmed T, Vafaei A, Curcio CL, Phillips SP, Zunzunegui MV. Sources of social support associated with health and quality of life: a cross-sectional study among Canadian and Latin American older adults. BMJ Open 2016; 6:e011503.^,99 White AM, Philogere S, Fine L, Sinha S. Social support and self-reported health status of older adults in the United States. Am J Public Health 2009; 99(10):1872-1878.^,1212 Douglas H, Georgiou A, Westrook J. Social participation as an indicador of successful aging: an overview of concepts and their associations with health. Aust Health Rev 2017; 41(4):455-462.^,1313 Norstrand J, Chan KT. The relationship between health and community across aging cohorts. J Aging Research 2014; 626097.^,2020 Rawal I, Kaur M, Khanna P, Lakshmi PVN, Kaur N. Association between Social Capital and Self-rated Health of Older People in Chandigarh, India. Indian J Community Health 2017; 29(2):198-202.^,3131 Hahs-Vaughn DL. Propensity Score Analysis with complex survey samples. In: Pan W, Bai H, organizadores. Propensity score analysis: fundamentals and developments. Nova York: The Guilford press; 2015. p. 237-264.. One of the disadvantages of this approach is that these methods have inherent limitations when it comes to determining homogeneity across exposed and unexposed groups. We therefore believe that the findings of this study can help to improve the understanding of the relationship between participation in community activities and positive SAH status of older people in Brazil.

On the other hand, it is important to highlight some of the study’s limitations. First, social participation in community activities can vary according to the number, duration and type of activity or service, interpersonal relations and social setting in which they occur. This diversity may partially reflect differing needs among the older persons who participate in the activities, as well as the size and source of their resources. But the dichotomous classification of this variable assumes that variations in exposure to this variable are irrelevant or too small to alter the observed effects. Furthermore, cross-sectional studies are limited in their ability to determine direction of causality, because the interaction between social relations and health is bidirectional and may change over the life course: a decline in health status restricts social relations, while a reduction in the latter is a repetitive and prospective predictor of mortality and acute morbidity. However, in the case of this study it seems unlikely that older people first make a positive assessment of their health and then become interested in frequenting clubs, community or religious groups, or social centers for the elderly. Rather, we believe that the health benefits result from social participation. Our results are consistent with some longitudinal studies have found similar results⁸8 Gilmour H. Social participation and the health and well-being of Canadian seniors. Statistics Canada, Catalogue Health Reports 2012; 23(4):1-12.^,99 White AM, Philogere S, Fine L, Sinha S. Social support and self-reported health status of older adults in the United States. Am J Public Health 2009; 99(10):1872-1878.^,3737 Kim B, Park S, Antonucc TC. Longitudinal changes in social networks, health and wellbeing among older Koreans. Ageing Soc 2016; 36(9):1915-1936.. One of the limitations of the NHS is that it did not select older people living in temporary or makeshift accommodation, who may be more vulnerable (not participating in social activities), and long-term care facilities (where the elderly may be exposed to good or poor quality activities depending on the facility)²⁷27 Sousa-Júnior PRB, Freitas MPS, Antonaci GA, Szwarcwald CL. Desenho da amostra da Pesquisa Nacional de Saúde 2013. Epidemiol Serv Saúde 2015; 24(2):207-216..

Despite these limitations, the weighted model used by this study provides a useful alternative method for guaranteeing a higher degree of internal validity and ensure that results are representative of the vast majority of the country’s elderly population, which has similar characteristics to the variables observed by this study. Population-based studies have also used propensity score weighting to ensure greater compatibility between exposed and no exposed groups²⁶26 Andrade ACS, Mingoli SA, Fernandes AP, Andrade RG, Friche AAL, Xavier CC, Proietti FA, Diez-Roux AV, Caiaffa WT. Neighborhood-based physical activity differences: Evaluation of the effect of health promotion program. Plos One 2018; 13(2):e0192115.^,3030 Oakes JM, Johnson P. Propensity score matching for social epidemiology. In: Oakes JM, Johnson P, organizadores. Methods in social epidemiology. Nova Jersey: Jossey-Bass, A Wiley Imprint; 2006. p. 370-392.. This method provides an unbiased estimate of the exposure effect close to that which would be obtained from the random assignment of individuals to study groups, as in randomized studies²⁷27 Sousa-Júnior PRB, Freitas MPS, Antonaci GA, Szwarcwald CL. Desenho da amostra da Pesquisa Nacional de Saúde 2013. Epidemiol Serv Saúde 2015; 24(2):207-216.^,3030 Oakes JM, Johnson P. Propensity score matching for social epidemiology. In: Oakes JM, Johnson P, organizadores. Methods in social epidemiology. Nova Jersey: Jossey-Bass, A Wiley Imprint; 2006. p. 370-392.^,3333 Hernan MA, Robins JM. Causal Inference. Boca Raton: Chapman & Hall/CRC; 2017.^,3434 D'Agostino-Júnior RB. Tutorial in biostatistics: propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Statist Med 1998; 17:2265-2281.. Indicated that it is necessary an adjustment of measure of the association obtained by the crude model, which can be overestimated because of the lack of homogeneity between the comparison groups.

This is important, because without propensity score weighting the group that participated in social community activities would not be comparable to the group that did not participate, given the distinct individual and contextual opportunities and propensities related to social and community engagement. Therefore, health studies need to consider these differences in the profiles of elderly people who engage in these activities.

Thus, these limitations do not detract from the importance of the findings of this study in relation to social and community engagement among the elderly, since they were obtained using probability sampling to produce a sample that is representative of the Brazilian population across all regions of the country, enabling an accurate estimate of the phenomenon under study and showing that population aging in Brazil is a dynamic, complex, heterogeneous, and inequitable process.

Conclusion

The findings show that, despite generally low levels of social participation in community activities among older people in Brazil, there is a positive association between participation and positive self-assessment of overall health status, confirming that engagement in these activities provides important gains for the health, well-being and quality of life of this population group. The government and civil society organizations should therefore work together to promote the participation of elderly people in clubs, community and religious groups, and social centers for the elderly in order to increase social interaction, strengthen social and emotional support, and extend the benefits of social participation to a greater proportion of Brazil’s growing elderly population.

Acknowledgements

We are grateful to the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brazil (CAPES) - Financing Code 001, and to the Graduate Program in Public Health at UFMA.

References

Publication Dates

History