In this study, we examined the association between social capital and loneliness among the older population and explored the interactive effect of social capital and demographic factors and health-related factors on loneliness in Anhui Province, China. Our results confirmed the association of social capital and loneliness and the combined influence of social capital and some other factors on the development of loneliness in later life, suggesting the relevance of social capital in preserving the emotional health of older people.
The results of the current study demonstrated the relationship between loneliness and social capital. Specifically, older people who lacked social capital concerning social participation, social connection, and reciprocity were more likely to experience loneliness, which was echoed by findings from previous studies [10, 20, 25, 41]. Similar to our results, Nyqvist et al.  found that infrequent social connection with neighbours had an increased likelihood of being lonely among older people aged 65-80 years in Western Finland. Results from the fifth wave of the Survey of Health, Ageing and Retirement in Europe (SHARE) also concluded that more social participation was a protective factor for loneliness and mitigated the impact of unfavorable socioeconomic status among older people . Some studies also suggested that forming and building reciprocal connections or relationships with others could alleviate the impact of some mental health issues, including loneliness [20, 42].
According to the results of our adjusted logistic regression analysis, a non-significant association was observed between a lower level of social capital regarding social support, trust, and cohesion, and loneliness. Different from our results, a prior study revealed that less trust and a weak sense of belonging to the community were significantly linked to a higher risk for developing loneliness . Meanwhile, a study also found that an increase of trust at the community-level also contributed to a reduction in loneliness among community-dwelling older people . Also, a previous study showed that insufficient social support is significantly related to the onset of loneliness [44-46]. There are two possible reasons for this inconsistency. First, this may due to different measurements that were used to assess social capital in the current study and earlier works. For example, in this paper, social support was measured by asking participants “how often can they get mental or material support when they are in need”. But Chen et al.  measured social support by asking surveyed subjects the quantity of social support offered by friends. The lack of common and widely used social capital measurement tool has been recognized as a repeated issue, which contributes to some inconclusive results in this research field. Another reason may be that our adjusted model included several variables such as demographic and health-related as covariates, which were associated with social capital and loneliness. As a result, the association between social support, trust, and cohesion with loneliness was attenuated or even became non-significant. However, more research is still warranted to further verify our results in the future.
Previous studies demonstrated that limitation in functional ability, being single, and multimorbidity was related to a higher risk for developing loneliness among older people [19, 47]. In the present study, most importantly, an interacting relationship between social capital and functional ability, marital status, and the number of diseases was observed. That is, older people who reported limitations in functional ability and had a lower level of trust and social participation were the most likely to experience loneliness. This indicates that trust and social participation as a social capital dimension may be of relevance in loneliness prevention [10, 25]. Besides, single older people, who had less social connection were more prone to loneliness as compared to those who had a higher-level social connection, which further highlights the protective role of social connection in preventing the incidence of loneliness . Interestingly, the significance of trust was not found in the adjusted logistic regression model, however, it was observed in the CART model. We suggest that the role of trust may depend on the appearance of other social capital dimensions and variables, which emphasizes the effectiveness of the CART model in examining the complex interactions among multiple variables that may be overlooked in the conventional analytical approach . The importance of this finding lies in adding a scientific explanation of using the CART model to help examine the association between social capital and loneliness while revealing how social capital interacts with other factors and produces an effect on the development of loneliness. Additionally, the CART model was used as a predictive model to estimate the subsets of older people that are more likely to become lonely.
Regarding the assessment of loneliness, different tools and methods were employed to identify loneliness as well. For instance, the UCLA loneliness scale was the most common and validated scale used to collect loneliness data in previous research [49, 50]. However, a single item by asking participants how often they feel lonely, which potential responses included often, sometimes, and never, etc., also had been validated a good predictive validity in many studies [10, 11, 30]. Moreover, the use of a single question to assess loneliness has been proved to have relevant edges such as succinctness, easily understood, and well accepted by subjects . Besides, in the process of data analysis, many studies dichotomized the status of loneliness into two groups (lonely, not lonely) by combining often and sometimes as the presence of loneliness, while rarely or never as the absence of loneliness [10, 11, 30, 37]. Despite such dichotomization could result in some loss of information about the outcome, to obtain statistical power during statistical process and comparisons with other research. Likewise, we also categorized loneliness into two kinds in this paper.
Loneliness is a subjective and negative feeling that may be the consequence of dissatisfaction with an individual’s social relations and unsatisfied social needs . Therefore, by looking at these findings from the present study, some relevant suggestions for the interventions of social capital to reduce the level and alleviate the impact of loneliness among community-dwelling older adults can be offered. First, we suggest older people should engage in more social activities, for instance, older adults are encouraged to participate the formal/informal groups (political/non-political parties, elections, hobby groups, etc.), voluntary activities, and services (heath lecture, culture, and physical education activities) within the community, which provide opportunities for them to meet the social needs, share life experiences and exchange interests. Second, the children, relatives, friends, and neighbors are encouraged to have more frequent interactions and communications with the older people and care about them, which is beneficial to maintain a good social relationship, promote and improve the quality of social connections, in turn, reduce the chance of developing loneliness among older adults. Third, to have an ideal reciprocal relationship, programs, and activities designed to cultivate and escalate the willingness to actively provide help to each other, including their relatives, friends or neighbors, and strangers, should be introduced. Lastly, to make good use of the role of social capital in the prevention of loneliness and maintain good emotional health, more attention should be paid to those who reported limitations in functional ability, had a lower level of trust and insufficient social participation, and single older people who had a less social connection.
This study has several limitations. First, since it was a cross-sectional study, which limited to conclude the causal relations between social capital and loneliness. Future studies using a longitudinal or randomized control trial design are warranted. Second, data in our study were based on self-report and might be subject to a recall or reporting bias. Nevertheless, to improve the data accuracy, we formulated clear and precise questions and carried out a pilot study before investigation. During data collection, forward or backward recall techniques were also used. Third, since our study was only conducted in Anhui Province, this might constrain the generalization of our findings to other regions or countries. Future studies that include extensive sites and larger samples are needed. Fourth, in the present study, we did not focus on the concept of emotional and social loneliness yet took loneliness as an umbrella concept instead. Besides, variables included in this study were not broad enough; some other variables such as depression and other mental health data were not well considered. More attention should be given to these mental health factors in the future.
Despite the limitations, this study also has some strength. Findings from this study are reliable because the sample was representative and had high response rates from participants. We also used a validated and standardized social capital scale, which may facilitate the development of social capital theory. Moreover, to our knowledge, this study is the first using the CART model allowed for further exploring multi-variable interactions and yielded a straightforward and visible tree, which is suitable in devising more specific and accurate strategies to counteract the impact of loneliness.