According to the activity theory, participation in social activities is likely to increase life satisfaction and contribute to promoting health for older adults [1]. In Rowe and Kahn's successful aging model, active participation in social activities is emphasized as a key component for successful aging in addition to preventing diseases disability and maintaining physical cognitive function [2]. As the population rapidly ages, many cohort studies in Europe and the United States (US) have reported that participating in social activities has positive health outcomes for older adults [3, 4, 5]. Specifically, a 17-year longitudinal study reported that collective social activities reduced the risk of mortality and institutionalization among older adults [3]. In addition, participating in social activities is likely to decrease the risk of physical/cognitive function decline and the risk of dementia [4, 5]. In Korea, positive health outcomes from participating in social activities were found to be relatively large in a group of older adults compared to other age groups [6]. Older adults who participate in more social activities including religious and leisure activities tend to have better cognitive function [7], and those who participated in more alumni and public activities were likely to report better subjective well-being [8]. These study findings suggest that participating in social activities could enhance older adults' health status [7, 8].

Regarding the mechanisms underlying the association between social activities and health, social activities are known to play a role in protecting individuals from psycho-social stress or buffering the effects of stress on the human body. Cobb reported that supportive social relationships allow individuals to perceive that they are cared for, respected, and members of an entity of mutual obligations [9]. Positive perception and experiences from their social relationships serve as a moderator of life stress, and ultimately protect individuals from stress-related negative health outcomes such as alcoholism, depression, and death [9]. An integrative review shows that supportive social relationships interact with individuals' health behaviors and mental health. These behavioral and mental changes are linked to cardiovascular, neuroendocrine, and immune systems, which, in turn, contribute to decreasing morbidity and mortality [10].

Conversely, there is a stream of research claiming that better health status influences older adults' levels of participation in activities. A cross-sectional study found that limited locomotor activities and chronic diseases (e.g., arthritis, cardiovascular diseases) are associated with lower levels of social participation [11]. A longitudinal study also demonstrated that the level of mobility-related functional ability at 75 years old was significantly related to the level of participation in social activities after five years [12]. Moreover, the better the subjective health status of the elderly, the higher the level of participation in social activities; and the greater the depressive symptoms, the lower the level of participation in social activities [13]. In summary, previous literature has shown that there is a significant relationship between social activity participation and health among older adults. Specifically, two groups of studies exist in terms of the relationship directions of two variables, 1) increased social activities influence positive health outcomes; and 2) better health status influences the level of activity participation.

However, most studies on social activities and health have primarily focused on elderly participants who are cognitively intact or the studies have not considered the cognitive function in the study design. Cognitive impairment refers to a declined state in cognitive domains such as memory, orientation, language, and executive function [14]. According to a national survey on the prevalence of dementia in Korea, the number of older adults with dementia was about 540,000 (9.2%) in 2012, and is expected to be about 2,710,000 (15.1%) by 2050 [15]. In addition, 27.8% of older adults are reported to have mild cognitive impairment, which may increase the risk of later progression to dementia [15]. Older adults with cognitive impairment are physically and socially vulnerable and have been marginalized due to age and cognitive impairment; however, few studies have examined the meaning of daily and social activities in their lives [16].

Research on the relationship between cognitive decline and participation in social activities has only recently been conducted. In a US national study, social activities decreased by 8% among the elderly with mild cognitive impairment, and by 19% in the elderly with dementia compared to those who are cognitively intact [17]. A longitudinal study consistently reported that the levels of social activity participation decreased as their cognitive function declined among Korean older adults [18]. These findings indicate that older adults who are cognitively impaired are less likely to participate in social activities, but the social activities are still valuable for the older adults who are cognitively impaired because they experience the meaning of pleasure and a sense of belonging through social activities [19]. However, there is a lack of research on what differences exist in the relationship between social activities and health outcomes with reference to the level of cognitive function for older adults.

The purpose of this study is to examine the relationship between social activities and health for community-dwelling older adults with reference to their level of cognitive function. Furthermore, given the two streams of research about social activities and health in terms of the relationship direction, we examined whether higher participation in social activities predicts better health status, whether better health status predicts more active participation in social activities, or whether there is a reciprocal causal relationship in both directions at the same time using national-level longitudinal data.

In this study, we examined the reciprocal causal relationships between social activities and health by level of cognitive function of community-dwelling older adults using CLPA. A major finding for all three groups was that higher levels of social activities are related to better perception of subjective health after two years, which means that participating in more social activities is effective in predicting higher health levels after two years for not only the older adults without cognitive impairment but also for those with mild and moderate/severe cognitive impairment. In the analysis of the opposite path, better health status predicted higher levels of social activities after two years only in the group of older adults who were cognitively intact. Our results suggest that social activities and health have a reciprocal causal relationship in the group of older adults with normal cognitive function, and a causal relationship was found in social activities as a predictor of health for older adults with mild and moderate/severe cognitive impairment.

A major finding of this study that older adults' participation in social activities is a predictor of better health is consistent with previous studies indicating that social activities have positive effects on physical [3, 4], cognitive [5, 7] and subjective well-being [8] for older adults. Even for those who are cognitively impaired, participation in social activities is a major requirement for maintaining and promoting health. In Korea, public care services for older adults with cognitive impairment or dementia in Korea are primarily treatment services in hospitals or long-term care services in nursing homes, but there are relatively few community-based services such as daytime care centers [28]. Moreover, community-dwelling older adults with a low level of cognitive function need various levels of support to help them participate in social activities and maintain interpersonal relationships. It is necessary to provide opportunities to participate in social activities with a focus on community resources such as public health centers, elderly welfare centers, day care centers, and dementia support centers. In addition, safe mobile and commuting services are necessary to improve accessibility to the resources for those with a low level of cognitive function.

Subjective health status was also found to be a determinant of participation in social activities after two years in the group of older adults who had normal cognitive function, which is consistent with the results of previous studies [22]. However, this causal relationship was not supported in the group of older adults who were cognitively impaired (both mild and moderate/severe impairment). Even if older adults who are cognitively impaired perceive their own health condition as good, there are many barriers to participating in social activities due to limited mobility support from the family or community. In other words, for older adults with cognitive impairment, the level of health perceived by the individual is insufficient to play an independent role as a predictor of social activity. In terms of policy, the Framework Act on Low Birth Rate in an Aging Society, article 14 (Encouragement of Leisure, Culture & Social Activities) prescribes that “the state and local governments should promote the leisure and cultural activities for older adults and establish a social basis to promote the participation of the older adults in social activities.” However, the current implementation system and road map are still insufficient for community practices related to this policy. In academia, the major factors and mechanisms contributing to the participation of older adults' social activities with reference to the cognitive level are yet to be identified. Future studies are needed to examine what and how support of the family and community acts as a mediator of participation in social activities in order to provide evidence to develop relevant policies.

The descriptive analyses also showed significant differences in the participation of social activities according to the level of cognitive function in the elderly. Specifically, the number of social activities in which they participated was higher in the following order: normal cognitive function, mild cognitive impairment, and moderate/severe cognitive impairment. This is consistent with a previous study in which older adults who experienced severe cognitive function impairment and mild cognitive function impairment had a low level of participation in social activities compared to those who experienced no cognitive decline (B=-1.09, p<.05; B=-0.92, p<.05, respectively)[18]. Therefore, it is necessary to endeavor to improve the whole community environment such as the physical environment (e.g., traffic, roads and homes) and psycho-social environment (e.g., a sense of belonging) so that elderly people with cognitive impairment can maintain their participation in social activities. In western countries, the concept of “dementia-friendly communities” has been popular since the early 2000s, with the recognition of the need for a safe and accommodating community for older adults with dementia [29]. Since 2010, there has been active implementation of dementia-friendly communities and efforts have continued to enable older adults with dementia to safely and conveniently participate in community activities in a supportive atmosphere. In Korea, the Sosu Village in Buan-gun, Jeonbuk Province was selected as a pilot project for a rural type of “dementia-safe village” in August 2016.

To create a dementia-friendly environment, the level of awareness of dementia in the community is the most important factor. However, the Ministry of Health and Welfare reported that the level of dementia awareness was 64.7 out of 100, on average, nationwide, and 88.2% of the subjects did not have any education about dementia [15]. In particular, given that the general population has a negative social perception of cognitive impairment and dementia and there is a stigma of the disease compared to other diseases, it is necessary to systematically educate the public and raise fair awareness of dementia. This awareness could be important to create a supportive environment where older adults with cognitive impairment can be socially respected and treated like other person.

Although this longitudinal panel data analysis study is superior to the other cross-sectional studies, there are some limitations. The two-year period in this study might not be sufficient to explain the behavior and to adequately observe health changes in older adults. Second, participation in social activities was calculated using the sum of seven types of activity items coded as dichotomous (yes or no) variables, so the quality of activity participation (i.e., frequency, commitment) was not considered in the measure of social activity.

This study aimed to examine the reciprocal causal relationship between social activities and health with reference to the cognitive function level among community-dwelling older adults. This results showed that a reciprocal causal relationship between social activities and health in all three groups regardless of the cognitive function level. However, current health status was not a predictor of future social activities in older adults with mild and moderate/severe cognitive impairment. Based on the study results, particularly the results related to the longitudinal effects of social activities on older adults' health in the group of those who have cognitive impairment, community interventions to promote social activities for older adults with cognitive impairment should increase. This could ultimately improve older adults' overall health status, which may in turn decrease the national-level social cost to care for older adults in the future.

Encouraging older adults to participate in social activities can support the concept of the “Aging in place (AIP).” AIP emphasizes that older adults should attempt to continue to stay in their homes and communities where they have lived while receiving adequate support or services despite disabilities or limitations [30]. AIP also highlights the preservation of self-reliant living in a familiar community with the individual's identity and intimate interpersonal relationships. To accomplish this goal., it is imperative to develop policy and institutional efforts for older adults with cognitive impairment to 1) provide more opportunities in the community and secure available community resources, 2) improve accessibility to social activities for the elderly by providing mobile and commuting services, and 3) create an aging-friendly community environment.