The data used in this study were from the 2011, 2013, 2015, and 2018 waves of the China Health and Retirement Longitudinal Study (CHARLS). It is a nationally representative longitudinal study that surveys Chinese residents aged 45 years or older since 2011 (wave 1). It covers not only personal information and environmental information, but also factual information and attitude information, such as sociodemographic characteristics, socioeconomic status, health status, and psychological conditions. In the sampling method, a stratified (by per capita GDP of urban districts and rural counties) multistage (county/district-village/community household) random sampling strategy was adopted, and finally a total of 150 counties in 28 provinces of China were sampled [24]. The CHARLS baseline survey in 2011 included 17,708 respondents aged 45 years or older. Up to wave 4, 5587 respondents were lost to follow-up. For this study, we excluded older adults aged less than 60 years (n=7217) at baseline. Then, older adults with missing data (n=987) on main variables (SIs, loneliness, and social isolation) and who moved between urban and rural areas during follow-up (n=848) were excluded. Finally, a total of 3069 older adults who participated in all follow-up waves were included in our study, with 530 urban residents and 2539 rural residents (Multimedia Appendix 1).

Descriptive statistics were used to describe the baseline characteristics of the sample. Continuous variables were summarized using means and SDs. Categorical variables were reported using numbers and percentages. The generalized estimating equation model assuming an independent working correlation structure was used to examine the associations of SIs with loneliness and social isolation among older adults in rural and urban areas during the follow-up period. We calculated the estimate odds ratios (ORs) and 95% CIs while adjusting for all identified confounders. Furthermore, in sensitivity analysis, we treated social isolation as a continuous variable to assess the robustness of the relationship between SIs and social isolation. A 2-sided P<.05 was considered statistically significant. Stata version 14.2 (StataCorp) was used for the data analyses.

This study protocol was approved and organized by Peking University’s Institutional Review Board (IRB00001052-11015). All procedures were in accordance with the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards. This survey was anonymous, and the answers were protected by privacy law. Written informed consents clarifying the study purposes were obtained from each participant before the survey.

Using longitudinal data from a nationally representative sample, this study provides an insight into the effect of DSI on loneliness and social isolation among Chinese older adults living in urban and rural areas. Results show that HI only, VI only, and DSI are each independently associated with a higher risk of loneliness in both urban and rural areas. HI and VI have a synergistic relationship with loneliness in rural areas, but we did not observe such an effect in urban areas. Our findings supported the view that SIs did not affect social isolation.

There are multiple reasons why older adults in rural areas have higher prevalence of DSI, loneliness, and social isolation than their urban counterparts. Older adults in rural areas may rarely receive appropriate treatment measures when HI or VI appears, as they have fewer health care resources, lower access to several health information sources, and poorer socioeconomic status compared with their urban counterparts [32-34]. With rapid industrialization and urbanization, there is massive migration of younger and middle-aged people from rural to urban areas, leaving older adults to either live alone or with their spouse in rural areas. Older adults with SIs who lived in rural areas had poorer perceived tangible support than their urban counterparts [35]. Furthermore, public services and voluntary organizations in rural areas are less developed than those in urban areas; consequently, older adults in rural areas have less choices for participating in social activities [36]. Therefore, older adults in rural areas might be more likely to experience loneliness and social isolation.

Our study found that individuals with SIs experience more loneliness compared with those without SIs, which was consistent with other studies [19,37]. Older adults with SIs are often lonely but they are not necessarily socially isolated. One possible reason is that SIs influence loneliness by affecting communication with their closest relatives or friends, as loneliness is not caused by being alone but rather by the unmet affective gain of their closest relationship [38]. Previous studies have indicated that older adults with DSI were more prone to experiencing a breakdown in communication compared with those without SIs [39,40]. Older adults with HI may feel frustrated or embarrassed over their difficulty communicating, resulting in loneliness [41]. Another reason may be neural changes associated with SIs. HI may contribute to changes in the frontal lobe, which alter the regulation of emotion [42], that may contribute to the likelihood of loneliness.

Surprisingly, older adults with DSI are at a compounded risk of loneliness compared with those with HI only or VI only in rural areas, but not in urban areas. Previous studies reported similar results, but there was no difference in mental health between older adults with DSI and single SI [43-45]; however, these studies did not consider the regional difference. A typical cultural and personal value among Chinese older adults was that they do not tend to bother others or even their adult children as they perceived that people are very busy with their own lives; therefore, they scarcely bother to seek help even though they may be sick [46]. In other words, older adults almost do not depend on their children until they are unable to take care of themselves, although children are the main caregivers for their older parents in China. Older adults with HI only or VI only were able to compensate for their impairment in one sense with the other so that they could take care of themselves to some extent [37]. As a particularly vulnerable group with challenges, older adults with DSI probably have to count on their children [47]. In urban areas, older adults with DSI could get more attention and care from their children in a timely manner, which might help them relieve loneliness. Likewise, older adults with DSI also deserve more attention and care in rural areas. However, most of the rural adults out-migrate to the cities for work and cannot take care of their parents by their side. Phone calls appear to be the main way of communication between older adults and their children, but this may be limited by HI unfortunately. Moreover, older parents with DSI have to reduce social activities due to the lack of their caregivers’ help. Consequently, older parents with DSI who lived in rural areas may be at a higher risk to be lonely.

Previous cross-sectional studies have yielded conflicting results on the relationship between single SI and social isolation. A systematic review indicated that most studies found a relationship between HI and social isolation [41]. Kotwal et al [48] also found that HI rather than VI was associated with social isolation among older adults. However, significant correlations were found between VI and social isolation in another study [20]. Interestingly, in this longitudinal study, social isolation of older adults with SIs were not fundamentally different from those without SIs. According to the Socioemotional Selectivity Theory, older adults may selectively narrow their range of social partners and focus more on their closest relationships [38]. Reduction in social networks is likely normal in older adults, regardless of their health status including SIs and ADL limitation. Notably, undesired social isolation represents a low quality of social relationships and is very closely related to loneliness [49,50]. This led us to speculate that SIs may affect undesired social isolation. Future research is needed to shed more light on this issue. In addition, given the high prevalence of social isolation found in this study, observation should continuously be made for any signs of negative health outcomes in older adults with social isolation. Besides, active steps should be taken to prevent older adults from being socially isolated. Maintaining ties with family members and friends is important for preventing social isolation among American older adults [50]. In the context of Chinese culture, social ties to children may be an important priority in older adults.

Our findings provide some new inspiration for older adults to relieve loneliness. Many prior interventions were conducted to mitigate loneliness by stimulating socialization [51], but they rarely considered that SIs may prevent older adults from enrolling or adhering to an intervention. Furthermore, an individual may be lonely without being socially isolated [49]. Future interventions considering SIs might be more accessible and effective for reducing loneliness among older adults. Recently, efforts to improve hearing have already shown beneficial effects on loneliness [52,53]. However, 74.39% (2283/3069) of older adults reported HI in 2018, but only 0.72% (22/3069) were treated with a hearing aid in this study. This study emphasized that preventing HI among older adults is of high priority to reduce their risk of loneliness. In addition, we propose that parents with single SI in urban areas and those with DSI in rural areas should be get more attention from their children.

A particular strength of this study is the longitudinal examination of the effects of DSI on loneliness and social isolation by region of residence in Chinese older adults. However, several limitations of this study should be noted. First, SIs were self-reported, which might result in some bias. Participants may overestimate or underestimate their abilities to see and hear. Further research should use clinical diagnostic measurements to verify the data. Second, due to data limitation, the severity of SIs was not measured. Although the prevalence of SIs in urban areas was lower than that in rural areas, whether the severity of SIs in urban areas is higher than that in rural areas remains unknown, as urban people are more likely to experience occupational noise exposure. Future studies are therefore needed to consider these issues as the more serious SIs older adults have, the more likely they are to have physical and mental health problems [37,54]. Finally, loneliness was assessed with only 1 question. Although this measure was widely used in the literature [25,26], it might be less reliable than a composite measure.

Overall, this study found that SIs were significantly associated with loneliness rather than social isolation among older adults living in both urban and rural China. A synergistic effect of HI and VI on loneliness was observed in rural areas, but such an effect was not found in urban areas. A better understanding of the longitudinal effect of SIs on loneliness by region of residence could help policy makers to allocate health resources and conduct targeted interventions accordingly.