Background
The number of people living with dementia in 2015 was estimated at 47.5 million globally, and that number is expected to double by 2030, reaching 135.5 million by 2050 [
1]. The prevalence of dementia increases with age, with a risk of the disease doubling every five years after 65 [
2].
Both cognitive impairment and social disengagement are prevalent concomitants of old age [
3]. Cognitive impairment has been associated with adverse physical and psychological changes including self-rated health, disability, quality of life, and mortality [
3,
4]. Most epidemiological studies on cognitive function have shown that health factors, including chronic diseases (e.g., hypertension, diabetes, and cardiovascular disease), lower physical function, and depressive symptoms predict risk of CD [
5,
6]. Social factors including old age, sex, marital status, education, socioeconomic position, and social isolation are also associated with CD [
3,
7]. Cognitive impairment can cause severe physical, psychological, and socioeconomic burdens not only for the individuals affected, but also their families and society [
4,
8]. Finding a good population-based strategy for preventing or delaying CD is a challenging public health priority for the aging society [
9]. Numerous studies have shown that social activities are beneficial to various health outcomes of elderly people including self-rated health, physical functioning, depressive symptoms, and quality of life [
10‐
12]. Previous research has also shown that greater participation in social activities can reduce CD in the elderly by providing intellectual and emotional stimulation [
13‐
17]. Reportedly, intellectual stimulation through social activities can maintain cognitive function by increasing cognitive reserve capacity [
13,
18]. Participation in social activities can also provide meaningful social roles, self-esteem, and social competence, which might protect against neuropathology including reduction of the stress response [
19,
20]. Although social activity is widely believed to guard against CD in later life, knowledge regarding the types of protective social activity is limited.
Social activities can be classified into two categories (formal and informal) based on intimacy and intensity [
21]. Formal activity describes an activity with formal organizations such as an alumni society, based on specific objectives and focusing on the achievement of a goal. Conversely, informal activity includes interactions with family, friends, and neighbors. It has universal character in objectives and focuses on emotional function [
19,
21]. There may be some differences among activities classified as informal. In later life, there may be some differences in relationships between close friends and children [
22]. Despite the intimacy on which both types of relationships are based, the children-parent relationship is dominated by the kinship model and family norms, while the relationship with friends is based on the egalitarian relationship model and voluntarism [
22]. Furthermore, due to the differences in the norms on which these relationships are based, the mental health of the elderly may be affected differently [
12,
22‐
24].
A growing body of literature has shown that type of social activity can differentially affect morale, mental health, and survival in old age [
10,
12,
19,
24]. For example, a previous study reported that informal social activities such as ‘face-to-face talks’ and ‘phone conversations’ with friends were significantly related to decreased mortality risk in later life after controlling for relevant confounders [
25]. Several researchers have reported that risk of late-life depression showed an inverse association with frequent participation in informal activities such as contact with offspring by phone or letters; however, no significant association with participation in formal activities such as volunteering and attending an alumni society was noted [
12]. Another study among older Chinese people indicated that playing Mahjong, chess or card games; interacting with friends; and having an activity center in the community are more beneficial for maintaining episodic memory than are other social activities such as volunteering [
26]. A 20-year follow-up study of Swedish samples showed that prior political activities or mental activities were associated with late-life cognition after adjusting for covariates, but activities in organization were not [
27]. However, knowledge regarding the effects of different types of social activities on CD in older people is limited. Due to the different natures of various social activities, all types of social activities might not have a beneficial effect on cognitive function among older people. For example, in Asian countries, because of traditional filial piety, familism, and lack of a welfare system, the types of social activity beneficial to cognitive function in older age might be different than those in the Western world.
In the industrialized nations, most people 65–74 years of age are still living independently and are in relatively good health [
28]. However, the status of adults 75 years of age and older becomes increasingly worse in terms of mental and physical health [
28]. Many gerontologists have also suggested that older adults are not a homogeneous group [
28,
29]. Based on the physical and mental differences in later life depending on age group, numerous studies on CD have divided older adults into two or more age groups [
29‐
31].
The aim of this study is to investigate which types of social activity reduce cognitive decline 4 years later among young-old (Y-O) (age 65–74 years) and old-old (O-O) (age ≥ 75 years) adults using data from a longitudinal study of a community-dwelling Korean population.
Results
The study population consisted of 1568 non-demented community-dwelling subjects who had good cognition at baseline (MMSE score ≥ 24). Additional file
1: Table S1 shows the characteristics, social activities, and cognitive function levels of the subjects at baseline and CD 4 years later. O-O adults were more likely to be bereaved, less educated, more dependent in IADL, have lower household incomes, and have more comorbidities than Y-O adults. O-O adults had lower MMSE scores in 2008 and 2012 than Y-O adults. The average CD 4 years later was more pronounced in O-O adults than Y-O adults (Y-O adults, mean = 2.18; O-O adults, mean = 4.59;
P < 0.001). CD in MMSE sub-scores was also more pronounced in O-O adults than Y-O adults (Additional file
1: Table S2 and Additional file
1: Table S3).
Overall, the subjects were more involved in informal social activities (e.g., face-to-face contact with close friends) than formal social activities (e.g., participating in alumni societies or family councils). The percentages of subjects who were involved in informal social activities ranged from 99.6 to 100.0 % (contact with close friends, 100.0 %; contact with one’s children, 99.6 %; contact with one’s children by phone or letter, 99.7 %). The levels of participation in informal social activities were different between Y-O adults and O-O adults. O-O adults had significantly less frequent phone and face-to-face contacts with their children compared with Y-O adults. The percentages of subjects who were involved in formal social activities ranged from 13.7 to 59.8 % (church or other religious groups, 24.9 %; senior citizen clubs or senior centers, 59.8 %; alumni societies or family councils, 13.7 %). Subjects who participated in senior citizen clubs were more likely to be men, younger and married, they had better ADL compared with those who did not. The majority of subjects who were involved in religious activity were women (Additional file
1: Table S4).
Additional file
1: Table S7 shows the results of the multivariate linear regression analysis with social activities at baseline as the independent variable and CD 4 years later as the dependent variable, controlling for sociodemographics, health-related characteristics, mental well-being (quality of life and depressive symptom), and cognitive function at baseline. We investigated the types of social activity that reduce the risk of CD over time. In formal social activities, subjects who participated in senior citizen clubs or senior centers at baseline had a lower risk of CD 4 years later than those who did not in Y-O adults and total sample (Y-O adults, B = − 0.80, SE = 0.40,
p = 0.045; total sample, B = − 0.95, SE = 0.37,
p = 0.012). In informal social activities, more frequent contact with children by phone or letters was associated with reduced CD in O-O adults (B = − 0.95, SE = 0.45,
p = 0.038) but not in Y-O adults (B = − 0.14, SE = 0.15,
p = 0.343). More frequent face-to-face contact with children was positively associated with CD in O-O adults (B = 1.54, SE = 0.45,
p = 0.001) but not in Y-O adults (B = 0.08, SE = 0.15,
p = 0.589). After adjustment for covariates, participating in two or more formal social activities was significantly associated with lower CD compared with nonparticipation in O-O adults (B = −3.17, SE = 1.26,
p = 0.012) (Table
1).
Table 1
Multivariate linear regression analysis of the associations between the number of formal social activities and cognitive decline 4 years later
The number of formal social activities |
Zero | Ref. | Ref. | Ref. |
One | −0.64(0.42) | −0.18(0.43) | −1.96(1.05) |
Two or more | −0.94(0.49) | −0.25(0.50) | −3.17(1.26)* |
Adjusted R-square | 0.07 | 0.06 | 0.13 |
F-value | 4.33*** | 3.13*** | 2.26** |
As expected, in the study sample, older age and being dependent in ADL was positively associated with CD. These results were similar when the analyses were repeated separately for Y-O adults and O-O adults. Subjects with elevated depressive symptom over time or high levels of depressive symptom at baseline had higher risk of CD after adjusting for related covariates. Subjects with higher level of quality of life had a lower risk of CD. Subjects with low cognitive level at baseline had higher risk of CD.
Additionally, we analyzed which type of social activity can mediate associations between CD and depressive symptom using sobel tests. The depressive symptom at baseline did not significantly predict the level of social activities (B = −0.01, SE = 0.01, p = 0.490, for level of face to face contact with one’s children; B = −0.01, SE = 0.02, p = 0.136, for level of contact with one’s children by phone or letter; B = 0.01, SE = 0.02, p = 0.706, for participation in senior citizen clubs or senior centers) and so the conditions for mediation were not met.
Discussion
Using community-based longitudinal data of people whose initial cognition level was good, this study examined the types of social activities that can reduce cognitive decline among Y-O and O-O adults 4 years later. Several important findings emerged from our study. We found that older adults who participated more frequently in senior citizen clubs or senior centers at baseline had lower risk of CD 4 years later than those who did not. This association was independent of the influence of age, sex, education, income, marital status, ADL, IADL, chronic diseases, depressive symptoms, elevated depressive symptoms, quality of life, and cognitive functioning at baseline. A possible explanation for this is that senior center participants are more involved in cognitively and emotionally stimulating activities than non-participants. In many countries, the senior citizen center is an organization that provides various services, including physical/mental health, social and educational services, and recreational activities for older people [
26,
41]. Studies has indicated that intellectually challenging activities and active interpersonal exchanges can increase or maintain cognitive reserve, allowing individuals to cope longer before cognitive impairment is manifested [
42], and produce beneficial effects on cognition even in old age [
13]. Studies of cognitive reserve suggest that there are individual differences in the ability to cope with the brain pathologic changes [
43,
44] and cognitive reserve may be improved by exposure to an enriched environment such as social interaction or mentally challenging activity, physical activity, education [
45], and may buffer against age-related CD [
46]. Biological mechanisms related to the vascular hypothesis could also partially explain our findings. Reportedly, an active social activity might enhance cardiopulmonary fitness and cerebral oxygenation, which could protect against neuropathology in older people [
47]. Other research on senior center participation and health has shown that senior center participants had higher levels of social interaction and better mental health than non-participants [
48]. In South Korea, older people participating in senior citizen clubs can spend their leisure time with other people playing hwatu (Korean card game) or chess, singing songs, or dancing [
49]. The senior centers provide diverse group activities for the elderly including painting, calligraphy, origami, gardening, playing musical instruments, acting, and exercise programs [
41,
49]. Our result suggests that participation in senior citizen clubs may protect against CD in elderly people.
The interesting finding in our study was that frequent face-to-face contact with one’s children was positively associated with CD 4 years later in O-O adults but not in Y-O adults after controlling for potential confounders. According to the social exchange theory, excessive giving or receiving might be harmful for mental and physical health [
50]. Excessive giving may result in exhaustion of resources of giver. On the contrary, excessive receiving can be distressing to the recipient because it may results in loss of independence [
51]. Both situations might be harmful for mental and physical health among older adults [
52]. In Korea, participation of women in the labor market has grown speedily as a result of industrialization. High female employment rate has resulted in increased need of assistance for household affairs such as taking care of young children in the family [
19]. In Asian cultures where family interdependence is culturally valued, when adult children request, most of older adults meet their adult children and provide instrumental support [
19,
53]. The burdens of helping their adult children might generate chronic unremitting stress, which may result in neuronal degeneration and have a negative impact on cognitive function in older adults [
44,
54]. Several existing studies in Asian countries have reported that the care-giving burdens for their adult children might lead to stress and depressive symptoms among older adults [
19,
53]. A study among older Nepalese adults found that close family relationship was related with greater stress and likely to have a negative impact on mental well-being in older women [
23].
Receiving support that cannot be returned, meanwhile, can be distressing to the recipient [
50]. With advancing years, older adults exchange support unequally with their children. The inability to reciprocate support can cause negative effects and stress in older people. Evidence has indicated that this stress can negatively impact cognitive functions [
55]. An elderly person’s dependency generally increases with age. Age and poor health can restrict the physical functioning of older people and therefore increase the possibility of receiving aid from their adult children [
51]. Empirical studies have found that O-O adults tend to have worse physical health and everyday functioning than Y-O adults [
56]. While Y-O adults generally reciprocate their children’s support fairly equally, O-O adults are less likely to maintain reciprocal exchanges and become dependent on their adult children [
22,
56]. O-O adults might experience this situation as a hassle or burden, which might impair cognitive functions. Studies have indicated that greater dependence can cause a devaluing of self and a lowering of morale [
51,
57]. Our findings showing that face-to-face contact with one’s children had a positive association with CD 4 years later only in O-O adults can be interpreted as mentioned above. Our results were in agreement with a study on social support and cognitive function among American elderly that showed that lack of reciprocity in the parent-children relationship can have deleterious effects on cognitive function in older adults [
58]. In Korea, there are strong instrumental interdependences between children and older parents because of insufficient social welfare programs (e.g., elder care support and child-caring) [
59]. Thus, plenty of social welfare programs for older parents and adult children should be developed to reduce excessive support exchange of parent-children.
Conversely, another study reported that co-residence with children or close interaction with children was positively associated with emotional well-being and self-rated health in the Japanese elderly [
60]. A study in Europe also showed that few contacts with children were associated with an increased number of depressive symptoms in older people [
61]. Because impacts of contacts with children on older adults’ mental health are inconsistent in the studies, more research is needed to identify plausible relationships between them.
Frequent contact with children by phone or letters was associated with reduced risk of CD 4 years later in O-O adults. As a result of Westernized lifestyle and urbanization in recent decades, many Koreans no longer live with their parents. The proportion of empty nest families and elderly living alone is steadily increasing in Korea [
62]. Talking with their adult children on the phone is likely to provide older adults emotional support and intimacy. This finding showing the beneficial effect of contact with children by phone or letters on CD is in agreement with previous study results reporting that frequent contact with children by phone or letters was significantly protective against depression among older people [
12]. Studies have indicated that perceived emotional support protects against CD and operates as an antidote to stress, thus delaying neurodegenerative processes, whereas loneliness or isolation can worsen CD in older adults [
20,
55]. Other studies have also shown that talking on the phone with children can provide older people with emotional support and be an important social activity in late life, confirming the concept that intimacy is needed for psychological well-being [
12,
25].
In this study, older adults tended to more frequently participate in informal rather than formal social activities. Empirical evidence showed that Asians are likely to have a family-oriented culture, and older Asians tend to participate in fewer formal social activities compared to older Western people [
12]. Our result also showed that there was a significant association between participating in two or more formal social activities and lower CD for O-O adults. This finding is line with a prior longitudinal study that greater social group participation can prevent CD [
54]. Other studies among older adults also have reported that participation in a number of different organization reduce the onset of long-term care [
63] and participation in a variety of social groups is effective for prevention of CD [
64]. Our result indicates that encouraging older adults to participate in various formal social activities may help preserve the cognitive function in community-dwelling elderly population.
In our study, a higher level of quality of life showed a lower risk of CD. Quality of life has been associated with decreased risk of depressive symptoms and cognitive impairment in later life [
65]. Our finding is consistent with a prior study reporting that quality of life was inversely associated with CD [
66].
Both elevated depressive symptom over time and high levels of depressive symptom at baseline were positively associated with CD. Depressive symptoms have been associated with increased risk of cognitive impairment in later life [
17,
38]. Consistent with our results, prior longitudinal studies have reported that cognitive function significantly declined over time in elderly women with elevated depressive symptom [
6] and depressive symptoms at baseline predicted CD independently of age, gender, duration of follow up and baseline cognitive status [
39].
Generally, formal and informal social activities have been considered to be more relevant for psychosocial well-being such as depressive symptom than for cognition. Ample evidence has shown the associations between social activities and depressive symptoms [
12,
23]. Recently, however, many works have reported the significant relationship between social activities and CD [
27,
54,
64]. Although age has been believed to be the strongest predictor of cognitive decline, our study showed that social activities may reduce CD in later life independently of the influence of age.
The possibility that social activities can mediate the relationship between depressive symptoms and CD might be raised because of associations between depressive symptoms and social activities. However, in our analyses, the social activities did not show mediating effect between depressive symptom and CD. Thus, it is reasonable to assume that social activities have direct relationship with CD in this sample.
This study had several strengths. First, using national longitudinal data, we examined a causal effect of social activity on delaying or preventing CD in late life. Second, we adjusted for various potential confounders of sociodemographics, health-related variables, mental well-being variables, and cognitive functioning at baseline. Third, subjects with cognitive impairment at baseline were excluded because impaired cognition can reduce social activity.
The present study had several limitations. First, although our analyses were restricted to study participants who had an MMSE score ≥ 24 at baseline, and adjusted for a wide variety of potential confounders, there still may be a possible reverse causation between cognitive function and social activity. Second, our findings are based a global measure of cognitive function. Thus, we cannot know which domains of cognitive function are specifically affected by social activities. Other research has measured specific cognitive domains such as episodic memory, semantic memory, perceptual speed, and visuospatial ability [
67]. Further research using fuller batteries of cognitive function will be able to address the relationships between specific domains of cognitive function and social activity. Third, social activity measures relied on self-reports, which are subject to errors of recall [
15]. Fourth, empirical evidence has indicated that quality rather than quantity of social interactions is more important to predict health outcomes [
68]. However, in this study, social activity was measured only based on frequency. Our scale did not involve a potentially important dimension of social activity such as specific content, period of time, satisfaction, troublesome aspects, or burden. Future research should use detailed instruments including quantity and quality of social activity in order to obtain more accurate information on how social activities affect CD in late life. Finally, social activities and covariates were only measured at baseline. Previous studies have indicated that changes in social activities may affect CD in old age [
64] and changes in other covariates such as health condition can influence the risk of CD [
69]. Thus, future research needs to take into account the change over time in social activities and covariates.
Despite these limitations, this study supports the notion that more participation in social activity may help to prevent or postpone CD in old age and provides important information on cognitive function, offering convincing evidence for the types of social activity that reduce CD over time using a large, nationally representative sample of the Korean elderly.