Social network types and functional dependency in older adults in Mexico

From a health systems perspective, the consequences of the demographic and epidemiological changes of the population have been extensively analyzed. Ageing means an increase in life expectancy, prevalence of chronic diseases, and need for health and social care services [1]. The aging of the population, particularly in developing countries, imposes new challenges and is a field of study that still requires answers from health and social care institutions.

In developed countries, older adults have access to retirement or pension funds, live in nuclear families, receive government and/or community support and social and health care [2]. In contrast in developing countries, older adults (up to 60%) do not receive income from retirement or pension funds, live in extended family groups, rely on family support and receive little support from governmental programs [3].

A topic that is gaining increasing attention is the effect of social networks on the health and wellbeing of older adults. A social network is the collection of interpersonal and communal bonds that people have throughout their lives to establish social relations that satisfy certain needs, and maintain their wellbeing [4]. Social networks can be characterized by their structure, function, and quality. Structure is the number of members, frequency of contacts, variety of links and proximity to the individual. Function comprises the type and frequency of support provided to an individual, and quality is how the individual perceives the structure and function of the network itself [4–6].

The concept of social network type provides a useful framework to derive groups of people with a common set of social ties. There are two prevalent approaches to determine the network type: a) The structural approach [5], which is based on structural components, such as the number of ties and composition, and interactional components, such as the frequency of contacts; b) The multidimensional approach which [6] takes into consideration structural, functional and quality components.

Studies in Europe, Israel, Japan, and the United States have reported that there are different types of social networks (TSNs). The most common are: a) diverse, with distinct sources of potential support (family, friends, neighbours, community groups) and with frequent contact; b) focused on family; c) focused on friends, and; d) restricted in terms of potential sources of support and frequency of contacts [4–7].

The social network patterns and their social relations vary among and within countries, due to cultural and socio-economic conditions [7–10]. For example, Japan and the United States have some unique TSNs, like "married and distal" in the former and structurally and functionally restricted in the latter [7]. TSNs have a differential effect on the health of older adults. Older adults within a restricted network decline in their physical function and overall health status [11–13] and show increased rates of mortality [14–16] and depression [17], whereas individuals related to a diverse network have longer life expectancy and better health status.

The potential effect of social networks is crucial in older adults with high degree of functional, mental or economic dependency. Dependency means that people require social, family or institutional support, due to temporary or definitive loss of their abilities [18]. The prevalence of functional dependency among older adults is high in both, developed and developing countries [19, 20]. In México the rates of functional dependency in basic and instrumental activities of daily living (BADL and IADL) is 24% and 23%, respectively [21].

The main risk factors for functional dependency are older age, female sex, low literacy, disease burden, such as cerebrovascular and other chronic diseases, depression, vision and cognitive impairment, lower extremities functional limitation, poor self-perceived health, low level of physical activity, smoking and low frequency of social contacts [19, 20, 22, 23]. Dependency is evaluated and classified in different ways, although in general terms, the greater the dependency, the greater the negative consequences for the older adult, his/her family and for the society.

In developing countries, such as Mexico, most of care for functionally dependent older adults relies on their family and friends, and they can or cannot receive limited institutional support, which comes from social security benefits or social programs. Such support is not universal and it does not necessarily fulfil their social and health needs. To our knowledge, there is information about the characteristics of social networks but there is no information about the types of networks that exist in such countries.

Mexico has 106 million inhabitants and a significant proportion of its population (approximately 48 million) are affiliated with the Mexican Institute of Social Security (IMSS for its acronym in Spanish). Within the population that IMSS covers, 12% are above 60 years of age and 40% of IMSS health care expenditures are allocated to this age group. While a large number of older adults require medical care [24], and do not have information about the types of networks that this age group is part of.

This study has three objectives: 1) To derive the types of social networks among IMSS affiliated older adults; 2) To describe the main characteristics of the older adults in each network type, including the instrumental and economic support they receive and their satisfaction with the network; 3) To determine the association between functional dependency and the type of social network.

This study is a secondary data analysis of the 2006 Survey of Autonomy and Dependency (EDAD-IMSS), which was performed on a sample of IMSS affiliated older adults (≥ 60 years old) who are residents of urban areas, without disorders of dementia, according to the Caban Minimental psychometric scale [25].

The survey was conducted to analyze the needs and demands of social services for dependent older adults. The sampling design included four states of the country located in four different geographical regions (North-Durango, West-Nayarit, Central-Mexico City and South Veracruz). These states were selected because of the high proportion of older adults living there.

The sample was stratified by age group and sex since these variables are thought to be related to dependency. A multi-stage sampling approach was performed. In each state, the main city was identified and then the IMSS primary care clinics were located. Next, within the area of coverage of the clinics, a random selection of households in which an IMSS affiliated older adult lived was selected.

To collect the information for the EDAD-IMSS survey, previously trained interviewers carried out direct, structured interviews with the older adults at their homes. The interviewers used a questionnaire on social networks that was previously constructed and applied by the 2004 Guanajuato's State Survey for Older Adults. The EDAD-IMSS survey was approved by the IMSS Research Review Board and IMSS Research Ethical Committee, and older adults provided verbal informed consent.

The survey included 3,426 older adults, from which 64 (1.9%) were excluded because of dementia disorders and 14 (0.4%) answered incompletely. The final analysis included 3, 348 older adults.

The diverse network with community participation corresponded to a large family network (mean size 6.3 members) in which almost half of the older adults had a life partner, all of them had children and friends and belonged to a community group. The frequency of contact with the family (7.1; SD 2.4) and friends was high (3.9; SD 1.2).

The diverse network without community participation was the most frequent. It was composed of the life partner, children or other relatives (mean size 6.8 members) and friends, with whom the older adult had frequent contact. Compared with other TSNs the older adults embedded in this network had the highest average of frequency contact with their family; although they did not belong to a community group.

The widowed type of network was the second common network. In this type of network the widowed individuals had children (average family size 6.8 members), with whom they had below average frequency of contacts; furthermore, they had above average frequent contact with their friends. The members of this network did not belong to a community group.

The nonfriends-restricted network was characterized by a large family (mean size 7.3 members). In this cluster all older adults had children, half of them had a life partner and they had frequent contacts with their family. The older adults of this group did not have friends and a low percentage (9.2%), were in a community group.

The nonfamily-restricted network was characterized by a small number of family members (on average 1.2). Only one third had a life partner and none had children; the frequency of contacts with their family was below average. Most of them had friends (93.3%) with whom they had relatively frequent contact.

This study derived a typology of social networks by applying the structural approach that has been used in developed countries [4, 5, 17]. To our knowledge, previous research of social relationships in México and in other developing countries has been focused on isolated aspects of social relationships [31, 32]. Therefore, deriving social network typologies could provide valuable information to better understand the current social conditions of older adults.

Our analysis derived five TSNs among urban dwelling older adults, affiliated with the social security in Mexico: diverse network with community participation, diverse network without community participation, widowed, nonfriends-restricted network and nonfamily-restricted network. These TSNs share similarities, yet they have differences from derived social networks in developed countries.

The diverse network with community participation was characterized by a large network, with many social ties, frequent contacts among members and participation in community groups. The characteristics of this network make it comparable to the diverse networks identified in Israel [4] and the United States [17]. The Mexican diverse network with community participation was larger in size and had more frequent contacts with its family members, than those of the above mentioned countries. In contrast, we derived a type of diverse network without community participation. This network was the most frequent in the present study.

Older adults within diverse social networks (with/without community participation) showed lower proportions of negative self-rated health, depression and dependency. This finding is similar to those from other studies that show that having several sources of support is associated with an improved state of health among older adults [6, 13, 33, 34]. Conversely, people in good states of health tend to have several sources of support.

Belonging to a community group does not replace the support that the family and friends give; although, it has a positive effect. In Mexico, given the results of this study, it seems advisable to create or promote insertion and participation of older adults in community programs. Current social support programs for older adults comprise transfer of economic resources, discounts in public services, and free health services, but these programs do not encourage participation in community groups.

The cluster analysis allowed deriving a widowed network type that has not been previously reported in other studies. The members of this network were predominantly women, they had lost their life partner and they did not belong to any community group. However, all the members of this network had family and friends with whom they had frequent contacts. It seems that the lack of both, life partner and community participation negatively affects older adults. When compared with the diverse networks, the proportion of older adults within the widowed network with negative self-rated health, depression and dependency was higher. Our interpretation is that in Mexico, older adults belonging to the widowed group, who are predominately women, are most vulnerable. Several reasons may account for this situation, mostly due to flawed social policies. For example, in the past women had very low opportunity for social development, such as access to education or employment. The family and eventually the husband were their main providers. Gender equity policies are relatively new and we believe that our results can be a valuable input to design comprehensive interventions for dependent older adults. Further information is needed to ascertain whether widows within this network have the conditions that ease them to have an active relationship with their family and friends and to be involved in community groups.

Two additional types of social networks were derived in the study. The nonfriends-restricted (the older adults have family but no friends) and the nonfamily-restricted (the older adults have friends and a very small number of family members). These types of networks are similar to those reported in the United States, where they are more frequent [17] than in our study. Studies in Europe and Israel have reported only one type of restricted network, which is characterized by limitations in both, family and friend's ties and contacts [4, 35].

In accordance with other studies we found that the nonfriends-restricted network had high proportions of negative self-rated health, depression and dependency [17]. The presence of friends in the social network is related to longer life expectancy [36], better self-rated health, lower risk of disability [37] and depression [17]; it has been suggested that friendship may be more influential than family relations on well being [12, 38].

The structure and function of social networks are not always correlated; the diverse and restricted networks have both, high and low support [6, 7]. Older Mexican adults in diverse networks without community participation and in nonfriends-restricted network received instrumental and economic support more often than those belonging to a diverse with community participations network or to a nonfamily-restricted network. Furthermore, we found that older adults in all TSNs were satisfied with the contacts they had with their family and/or friends. This finding is in accordance with other studies that have reported "a tendency for older adults to report less relationship negativity" [6, 39].

In this study, the distribution of the TSNs was different in the states where the survey was conducted. This is probably due to cultural and socio-economic variations, a situation that other countries have reported [8]. Mexico is a mosaic that reflects asymmetry in economic development and cultural differences among the states [40]. The results of the study would serve to develop interventions tailored to the local conditions and considering the predominant type of social network. This would increase the probability of satisfying the needs of specific groups of older adults.

Functionally dependent older adults more likely belonged to a widowed social network. This finding is consistent with previous studies that have stressed that the inability to carry out BADL and IADL increases as the diversity of the social network decreases [13] and greater instrumental support is related to greater risk of functional disability [37, 41]. It is important to offer support that stimulates the older adults, while respecting their initiative to carry out activities that they are capable of performing without help, and to identify those for which they need assistance.

In our study, the sex of older adults was not related to dependency for BADL and IADL and to social networks. The results of previous studies are controversial; some have described differences in the relationship between social networks and functionality of older adults depending on sex, while other studies have not confirmed such relationships. Possibly, this is attributable to the convergence of gender roles at these ages in certain cultures [13, 42].

Among the limitations of this study, it is worth mentioning that it is a secondary data analysis, which reduces the possibility of an in depth exploration of other aspects of social networks. For example, we did not have information about emotional support. Due to the multidimensional nature of the social networks, having information about such emotional support would enrich the analysis. Also, measuring satisfaction needs further development. The questionnaire addressed satisfaction in general terms and that could overestimate the results. The available information limits the possibility to use a multidimensional approach for the analysis; however, the structured approach provides valuable data to derive the above mentioned types of social networks in the context of Mexican population affiliated to social security. Also, our results are limited to the population of older adults affiliated with the social security and residents of urban areas, whose social networks might be different than those of the rural population or the uninsured. In social security, as well as in urban areas, there are usually more retirees; the labour background might play a role in strengthening the network.

The survey offers cross-sectional information thus it is not possible to make inferences about causal relationships or the direction of the association between the type of network and dependency. The conditions of the TSNs and functional dependency have a dynamic nature, due to changes in different aspects, such as the composition of the social network, in which its members enter and exit continuously; also, the health status of the older adults shows variable periods of dependency and independence. These circumstances create a complex scenario. The scope of this work is limited to the analysis of social networks and functional dependency, although there are other kinds of dependency in the older adult, such as psychological or economic dependency where the relationships between dependency and TSNs might be different.

We can conclude that this study shows that within the population affiliated with social security, some types of the social networks are similar than those described in developed countries. However, we identify the existence of a diverse network without community participation and of a widowed network that have not been previously described and are frequent. These networks represent a potential unmet need of the social security affiliates.