Bem-Estar e Envelhecimento ( SABE ) Social network and older people ’ s functionality : Health , Well-being , and Aging ( SABE ) study evidences

REV BRAS EPIDEMIOL 2018; 21(SUPPL 2): E180003.SUPL.2 RESUMO: Introdução: Possuir redes sociais ativas parece influenciar positivamente o desempenho funcional de idosos. Objetivo: Verificar a associação entre as características das redes sociais de idosos e o surgimento de comprometimento funcional. Métodos: Estudo longitudinal de base populacional que utilizou as coortes de 2006 (n = 1.413) e 2010 (n = 990) do Estudo Saúde, Bem-Estar e Envelhecimento (SABE). Para caracterização das redes sociais utilizou-se as seguintes variáveis: número de integrantes da rede; arranjo domiciliar; sexo e idade dos integrantes; co-residência com criança ou apenas com idosos; satisfação com a relação; recebimento e oferecimento de apoio social (financeiro, material, emocional, realização de tarefas dentro e fora de casa, companhia e cuidados pessoais). Utilizou-se regressão logística para a análise dos dados. Todos os cuidados éticos foram observados. Resultados: As redes sociais dos idosos possuem, em média, 8,15 integrantes e são constituídas predominantemente por familiares com idade entre 15 e 59 anos. Idosos dependentes recebem mais apoio material, para realização de tarefas domésticas, fora de casa e cuidados pessoais, enquanto os idosos independentes recebem mais apoio emocional e companhia. Oferecer apoio social (OR = 0,32; IC95% 0,14 – 0,71) diminuiu as chances de desenvolver dependência, independente de condições sociodemográficas e de saúde. Conclusão: Deve-se estimular o fortalecimento das redes sociais na velhice, uma vez que a confiança no cuidado informal, oferecido, principalmente pelas famílias, pode não ser a melhor opção para lidar com a demanda de cuidado crescente que acompanha o envelhecimento da população brasileira.

Social network and older people's functionality: Health, Well-being, and Aging (SABE) study evidences

INTRODUCTION
With aging, people are more susceptible to chronic diseases, which may be followed by disabling sequelae, creating situations of dependency and greater need for care 1 .This makes the health conditions of elderly people closely related to functionality, which involves independence and autonomy for the performing of their daily activities 2 .
These activities, also called activities of daily living (ADL) are didactically divided into basic (BADL) and instrumental ones (IADL), the first one being related to self-care (showering, dressing, walking, transferring, feeding, and maintaining continence) and the later, to the maintenance of an independent community life and to the ability of living alone (using transportation, going shopping, managing their own money, using the phone, controlling and taking their own medication, preparing hot meals, and developing light and heavy housework) 3,4 .
Regardless of the instrument used, the functional evaluation results in the identification of the level of functional impairment of the elderly people in terms of independence or dependence (partial or total), determining the care needed, as well as by whom and how these care should be carried out.With the populational aging, it is expected an increase in the proportion of people who will need care, situation that tends to last for a longer time than the one lived so far demanding adequacy of the existing care policies 1 .
• expressive drop in fertility rates (lower than 1.8 children/woman) 5 ; • progressive entry of women in the labor market reducing the care potential once that, traditionally, the care for the ill people, children, and elderly people has always been the role of women 6 ; • modifications in the structure of the unions and family configuration.Nowadays we live, for example, with situations as artificial insemination in women without a life partner; homoaffective marriages, remarriages with the constitution of different families, etc.Such changes may have an impact, in the near future, on the provision of care in the family context 6 .
In many developed countries, the formal care offered by the social and health systems are overlaid to the informal care offered by the families 7 .In Brazil, this care, organized and efficient for more dependent elderly people, practically never existed.
The assistance to a dependent elderly person involves time, personal and family organization, and resources (emotional, structural, and financial) which, in addition to other demands, may result in an overload of tasks negatively affecting the care given and, consequently, the quality of life of the elderly person.One of the resources that may contribute positively to the organization of the necessary care is the existence of a social and effective support network, where the elderly people and their family find the support they need to fulfill their everyday demands and the complication there might be 8 .
People who have active social networks live longer and with better health than the ones who do not have them 9 .Effects such as improvement in the quality of life, the subjective well-being, functionality, and mortality have been reported in the literature [9][10][11][12][13][14][15] .
With regard to the functionality, keeping few social contacts seems to negatively influence the functional performance, besides worsening the perception of the quality of life.On the other hand, taking part in social activities and maintaining active social relations influence positively the functional performance, especially among the oldest ones.Emotional and instrumental support are the most associated to the reduction of disabilities [10][11][12][15][16][17][18] .
The absence of involvement in social networks, as a risk factor to health, has been considered as damaging as smoking, hypertension, obesity, and the absence of physical activities, which suggests that health deterioration may be caused not only by diseases and life habits considered inappropriate but also by the reduction of both quantity and quality of social relationships 18 .
Thus, the objective of this study is to verify the association between the characteristics of the social network of elderly people living in the municipality of São Paulo and the development of functional impairment in a period of 4 years (2006-2010).

METHOD
This study is part of the Health, Well-Being, and Aging Study (Estudo Saúde Bem-Estar e Envelhecimento -SABE) and is characterized as longitudinal, exploratory, and descriptive, REV BRAS EPIDEMIOL 2018; 21(SUPPL 2): E180003.SUPL.2 with a quantitative approach and will use the data from the cohorts A 06 and B 06 in 2006 and their follow-up in 2010.In 2006, from the 2,143 elderly people in the cohort A interviewed in 2000, 1,115 elderly people were located and reinterviewed.Additionally to those 1,115 elderly, 298 elderly people in the new cohort (cohort B 06 ) were added, which amounted a total of 1,413 elderly people interviewed.In the year of 2010, cohorts A and B were once more located and reinterviewed, totalizing 990 elderly people.The difference observed corresponds to deaths, changes of address, institutionalizations, refusals, and impossibility of locating the elderly people.The interviews were carried out by trained interviewers, at elderly's home, with a standardized instrument.When the elderly person was unable to answer (owing to physical and/or cognitive problems), the information was obtained with a qualified informant (proxy).
The variables used for the characterization of elderly people were gender, age, marital status, education (years of schooling), self-perception of sufficient income, cognitive decline (adapted Mini-Mental State Exam), depressive symptoms (short Geriatric Depression Scale), self-evaluation of health; self-report of chronic diseases (hypertension, diabetes, chronic obstructive pulmonary disease, cardiovascular disease, osteoarticular disease, and osteoporosis), number of self-reported diseases, and falls within the 12 months prior to the interview.For the characterization of the social networks, the following were used: the total number of members in the network; number of members in each one of the different configurations (living in the elderly's home, children living in the same house as the elderly people, and other family and friends); type of living arrangement; gender and age of the members of the network; living with infants/children or only with elderly people; satisfaction with the relationship; and receiving and offering social support.The following kinds of support were considered: financial, material, emotional, performing of tasks inside and outside home, company and personal care.
In order to identify the level of dependency of the elderly people, the following categories were considered: independent, mild dependence, moderate dependence, and high dependence.The different levels of dependency were obtained by the hierarchical scale of difficulties reported in the functional performance with the elderly population in the SABE Study in 2006, built by the Guttman scaling method.The creation of the variable categories "level of dependency" was made by the grouping of activities according to the hierarchy of difficulty and the need for the presence of a caregiver to assist in their performance.
In the descriptive analysis of the data, we estimated the distribution of frequencies, means, and standard deviations for the continuous variables of the study.For the categorical variables, the proportions were estimated.The differences between the groups were estimated using Pearson χ 2 test, with the Rao-Scott correction, which takes into consideration the sample weights for the estimates with population weighting.For the continuous variables, the differences between the groups were evaluated by the Student's t-test.In all the analysis, the significance level of 5% was used.The analysis of the associations between the characteristics of the network of social support and the development of dependence was made by multiple logistic regression.The independent elderly people were selected in 2006 and the ones who developed some level of dependence in 2010 (mild, moderate, and high dependence) (n = 459).The magnitude of the association was estimated by the simple and adjusted odds ratio (OR).The model including the variables regarding the social network was adjusted by gender, age, income sufficiency, marital status, education, cognitive decline, number of disease, and falls.The statistical package Stata ® version 11.0 was used.

RESULTS
The sample of this study consists of 1,413 elderly people assessed in 2006.From these elderly people, 59.4% of them were female subjects, 57.4% were married, 38.6% studied from 4 to 7 years, 55.1% reported not having enough income for their needs, and the mean age was 69.6 years old [Standard Error (SE) = 0.6] (Table 1 12.2% of them exhibited cognitive decline, 14.3% with depressive symptoms, and 28.5% of them reported falls within the 12 months prior to the interview.Despite that, most elderly (53.4%) people were independent regarding their self-care (Table 1).
From the total of elderly people interviewed, 0.45% did not have social networks, that is, they did not report knowing people they could count on.This represents 4,585 elderly people in the population of São Paulo.Of these, 78.34% were women, and 50.83% were aged between 60 and 69 years of age.None of these elderly had indications of cognitive alterations, and 11.43% of them showed depressive symptoms; 20.07%were dependent, being 15.87% of them with mild dependence and 4.20% with high dependence.
The social networks of the elderly people consisted of, on average, 8.15 (SE = 0.21) members.The males' networks showed on average 8.55 (SE = 0.27) members and the females' 7.87(SE = 0.22) members (p = 0.009).Elderly people aged between 60 and 69 years of age had networks consisting of, on average, 8.70 (SE = 0.24) members, while elderly people aged between 70 and 79 years of age had an average of 7.65 (SE = 0.29) members, and those aged 80 years or older showed significantly smaller networks (6.61 members, SE = 0.20: p < 0.001).The mean number of members living in the same household with the elderly was 2.05 (SE = 0,08), whereas the mean number of children living in different households was 2.66 (SE = 0.12), and other family members and friends was 2.90 (SE = 0.75).
The number of members in the network consisting of "other family members and friends" was higher than of "children living in different households" which was larger than their home network.However, even with the lower number of members, the household network stood out in the offer of support to the elderly people, where there was more expressive reciprocity of support (Table 2).
As for the characteristics of the network according to the level of dependence of the elderly, both in the network composed of residents of the household and the one composed of sons and daughters who live in another house, it was observed the most dependent elderly people received more material support for the execution of house chores, outside their houses, and personal care, while independent elderly people received more emotional support and company.In the network composed by other family members and friends, in addition to the types of support mentioned earlier, the dependent elderly people received and offered more company.
There was no statistically significant difference observed between the number of members in the network (as a whole or in different configurations) and level of dependency.However, a difference regarding the household arrangements was observed.Among the independent elderly people, with mild dependence or with moderate dependence, there was a predominance of bigenerational arrangements (elderly people who lived with their children), whereas among the elderly with high dependence there was a predominance of trigenerational arrangements (elderly people who lived with their children and grandchildren) (p < 0.05).
In 2006, from the 1,413 elderly people evaluated, 615 of them were independent and were selected for the analysis of the characteristics of the social networks that determine the development of the dependence.Of these, 459 elderly were reinterviewed in 2010, 51 of them died and 105 were follow-up losses.There was a statistically significant difference between the accompanied elderly (n = 459) and the losses (refusals, moves to another municipality, institutionalization, and unable to be located) or deaths (n = 156), regarding gender and age.Higher proportions of death among elderly were observed among males and older elderly.
From the independent elderly people in 2006, 70.51% of them remained in this condition in 2010, and 29.49% of them became dependent at some level: 16.98% of them developed mild dependence, 5.43% moderate dependence, and 7.08% high dependence.
A statistically significant association was found between the increased dependence and age, marital status, education and self-evaluation of health, and the number of diseases.From the characteristics of the social network, "offering support only" showed statistical significance, considering that the elderly who offered support developed dependence in lower proportions (Table 3).
Table 4 presents the model of final logistic regression showing the offer of social support as a protective factor against the development of dependence.Elderly who offered support had about three times less chances of developing dependence.

DISCUSSION
The results of this study showed that the social networks of the elderly people consist of, mainly, family members, corroborating the results of other researches in Brazil, in Latin America in general, and in Asia [19][20][21][22] .Such fact may represent a concern owing to the changes in the constitution of families, which become increasingly smaller.Lower rates of fertility and birth and the changes in the social roles (characterized by the entrance of women in the labor market) were some of the factors which contributed for the decreased size of families and, as a consequence, a decrease in the possibilities of offering support 23,24 .
As for the size of the network, the average found (8.15 members) was below the average of 12.4 members found in a study carried out with elderly people with cognitive alterations in a city in the countryside of São Paulo, and also below the average of 13 members found in a population-based study carried out with an American elderly population.A result below 8.15 members was found in a populational study with Australian elderly (mean of 6.4 members) [25][26][27] .However, there are important differences regarding the selection of the sample in the studies mentioned earlier that should be considered.Additionally, the sociability pattern in a big metropolis such as São Paulo is characterized by greater distance between the individuals 28 .
Despite being a small percentage, the fact that 0.45% of the elderly reported not having any members in their social networks should be highlighted, as it reveals that a portion of the elderly, which is numerically significant in terms of population (n = 4,585), cannot count on anyone in case of need.Even more serious is the fact that 20% of these elderly (about 900 elderly people) have some kind of functional limitation, tending to be completely unassisted.Despite the social networks have 8.15 members on average, most part of the elderly people did not perform support reciprocity, that is, the proportion of elderly people who did not receive or did not offer support was always higher than the number of elderly people who received or offered support, corroborating the findings of a study carried out in a city located in the countryside of the state of São Paulo 25 .In other studies, it was observed that larger social networks are more beneficial, as they offer greater possibility of support reciprocity 29,30 .
With regard to the reciprocity of support, receiving help on personal care and domestic chores is essential, but the emotional support may work as a motivation for social integration, in addition to being essential for the physical and psychological well-being 31 .Therefore, in addition to practical support, receiving emotional support may lead to greater satisfaction, consequently improving well-being, which may positively affect elderly's health.
Among the dependent elderly people a difference between the support received from the coresidents and the children who lived in different households or other family members and friends was observed.Residents in the household of the elderly and children who do not live in the same household would offer more material support for performing house chores, outside the house, and personal care, whereas other family members and friends offered more company.
It is known that elderly people feel better when their practical needs (help with domestic chores and personal care) are met through family relations and their emotional needs are met through friendships 32 .Thus, the type of relationship with the elderly should not be analyzed by the benefits generated, but rather should seek what relation is more complementary 33 .
Considering that the social network of the evaluated elderly is mostly from family relations and that reciprocity of support with other family members and friends is scarce, stimulating the relationship with the latter may enhance the positive effects of the network.The beneficial effect of the relationship with friend is owing to the fact that they usually are more available in problematic situations.Besides that, friends tend to having the same interests, sharing experiences, meeting voluntarily, and developing pleasant social activities, unlike what happens in the family relationships, which are frequently focused on personal dependence and encumbered by the roles, expectations, and limitations imposed by family relationships 34 .
Offering social support was the only variables related to the networks which remained in the adjusted model by gender, age, perception of income sufficiency, marital status, education, cognitive decline, number of diseases, occurrence of falls within the last 12 months prior to the interview, number of members of the network, receiving support, and household arrangement.Previous studies already demonstrated the relation between the social networks and functionality.In these studies, the dissatisfaction with the social support received, scarce social participation, poor diversity in relationships, and the perception of low social support were risk factors for the development of disabilities 11,12,26,27,35 .
The possible explanations for the clear relation between the social networks and functionality may be based on behavioral, psychosocial, and physiological aspects.With regard to the behavioral aspects, the greater involvement with the social networks is related to the adoption of more positive behaviors for health over time, once that the members of the network may monitor, inhibit, and regulate such behaviors in a way to stimulate the adoption of healthier habits.In the psychosocial context, the greater involvement in social networks favors the feeling of belonging and of purpose of life, which, in turn, may improve mental health and, consequently, the physiological processes and physical health.Physiologically, the involvement in social networks may improve the immunologic, endocrine, and cardiovascular function, which improves, consequently, the organic response in stressing situations [36][37][38] .

CONCLUSION
The social networks of elderly people have, on average, 8.15 members and are formed predominantly by family members aged between 15 and 59 years of age.Dependent elderly receive more material support to perform their house chores, outside home, and personal care, whereas independent elderly receive more emotional support and company.Offering social support by the elderly was a protective factor against the development of dependence in a period of four years.
Considering the positive effects on health of the elderly, strengthening the networks at old age should be implemented as a habit to be developed and spread.Health professionals should stimulate a network formation where the elderly may, effectively, provide mutual support.In addition to favoring the relations that may satisfy the needs of the elderly in situations of dependence, intermediating relations where the elderly people may, also, offer support, may protect the elderly from dependency.More than using the family to establish these networks, the importance of the bonds of friendship should be highlighted, once that few elderly people reported having reciprocity of support among friends.
The trust in informal care, offered mainly by the household networks, might be the best option to deal with the growing demands for care that comes with the aging of the Brazilian population.Political approaches are necessary in order to support the families and to offer different forms of assistance, especially to dependent elderly population.

Table 4 .
Logistic Regression Model for association between network characteristics of social support and level of dependency, São Paulo, 2006-2010. Continue...