Association of living arrangements with depressive symptoms among older adults in China: a cross-sectional study

Depression is regarded as a major public health concern and may become the second most common cause of disability by 2020, trailing only heart disease [1]. According to the World Health Organization (WHO) report, nearly 350 million people were affected by depression worldwide [2]. Depression causes great suffering, decreases physical and social functioning, and even increases the risk for suicide among the elderly [3, 4]. As the population is rapidly aging in China, it appeals to aware the importance of depression as an public health issue across the nation. According to the China Health and Retirement Longitudinal Study (CHARLS), nearly 40% of older adults aged 60 and over have reported depressive symptoms. [4]

Depression, the major chronic disease currently, has been proved to be associated with genetic [5], behavioral, physical activity [6], quality of sleep [7] and the health condition like chronic diseases [5]. In addition, evidence suggests that social support is an important contributor to depression [3]. Living arrangements as a structural factor of social support may contribute to older adults’ depression.

The association between depression and living arrangements had been studied previously all over the world. And these results may differed across societies and cultures. It’s noteworthy that the studies in western countries focused more on the living arrangements whether living alone or not had different effect on depressive symptoms. An American study showed that older people living alone had more depressive symptoms than those living with others [8]. Another study in Finland also found the same evidence that persons living alone and living with others were more likely to have depressive disorder than people living with spouse [9]. However, there are more different structure of living arrangements in Asian. Most previous studies have suggested that elderly living with child are associated with higher risk of depression overseas: in Singapore, people living alone and living with children are associated with higher depressive symptom scores [10]. In Korea, older adults living alone, living with an unmarried child, living with grandchildren are more likely to have depressive symptoms [11]. In Thailand, having child living in the district predicted a higher odds of depression in the elderly [12]. Another study has found an inverse relationship that living with child is a protective factor to the prevalence of depressive symptoms [13].

However, the evidence for the relationship between living arrangements and depressive symptoms in China were very limited. And in Silverstein et al. study, it had drawn conflicting conclusions: older adults living with three-generation or living with grandchild were less likely to have depressive symptoms [14].

Based on traditional Chinese culture and the existence of only one child, parental depression might be more serious in China [14]. Since the implementation of the universal two-child policy put forward by the Chinese government in 2015 [15], parents bear a heavier burden to care for their children. Therefore, we researched whether older adults living with children may be associated with negative psychological states.

Previous studies have shown the association between depression and gender, marital status, physical health status in older adults [5, 16‐18]. However, the association between the living arrangements of older adults and depression is conflict in different studies.

What’s more, it has been recognized in many studies that depressive symptoms vary by gender among different living arrangements. In Taiwan, women living alone were more detrimental to be depressive symptoms [19]. In Vietnam, older male living with child had more positive effect in psychological wellbeing than female [20]. While in Singapore, male living alone with weak social networks had higher depressive symptom scores [10]. However, little was known in China about the association between living arrangements and depressive symptoms by different gender. Many studies in China just identified that female seemed to have higher risk of depressive symptoms than male [21‐23], and we cannot find the gender effect of association between living arrangements and depressive symptoms.

Therefore, our study aimed to investigate the association between living arrangements (including whether living with a spouse and whether cohabitating with a child) and depressive symptoms by gender, exploring whether older adults living with children may have more depressive symptoms, verifying whether the influence of such symptoms differs by gender, and considering what our society should do to improve mental health in older adults.

Table 1 presents the descriptive characteristics of our sample. Of the total number of participants, 3,050 were men (57.02%) and 2,951 were women (42.98%). Nearly 35.94% were measured as having depressive symptoms (28.13% for men and 44.02% for women). 41.29% lived with a spouse but not a child, while 35.14% lived with both a spouse and a child. Only 12.36 and 11.20% lived without a spouse with/without a child, respectively. Sociodemographic variables revealed that a large majority (66.89%) was 60–69 years old. Nearly 80% had low education levels, and more than four-fifths of participants lived in rural areas. Without any adjustment, we found that living arrangements was associated with depressive symptoms (p < 0.05). The covariates (sociodemographic variables, health behavior and health condition) included in the baseline were all associated with depressive symptoms respectively (p < 0.05) expect for the age group.

Table 2 shows the results of the multivariate logistic regression analysis between living arrangements and depressive symptoms. We found that different living arrangements had statistic difference in depressive symptoms. Crude OR’s shows that living without either spouse or child, and living without spouse but with child had higher risk of depressive symptoms than only living with spouse (OR 1.61, 95% CI 1.32–1.96 / OR 1.59, 95% CI 1.30–1.94). After controlling for the sociodemographic variables in model 2, we found that only older adults living with a spouse and a child had positive significantly effect in depressive symptoms, compared with those living only with a spouse (OR 1.15, 95% CI 1.00–1.33). However, model 3 showed that there was no difference between living arrangements and depressive symptoms adjusting for both sociodemographic variables and health behavior (including smoking, drinking and social activity). When controlling for all confounding variables (containing sociodemographic, health behavior and health conditions), the association between people living with both spouse and child, living without neither spouse nor child and depressive symptoms improved in magnitude. Model 4 showed that older adults living with only a spouse had the lowest risk of having depression tested by the multivariate logistic regression analysis. Compared with living with a spouse only, people living without neither a spouse nor a child and living with both a spouse and a child were more likely to have depressive symptoms (OR 1.40, 95% CI 1.03–1.92; OR 1.23, 95% CI 1.06–1.42). Even though individuals living only a child were 33% more likely to have depressive symptoms than people living with only a spouse, the proportion was nearly not statistically significant. In addition, the following variables were associated with stronger depressive symptoms: being female, younger age group, lower education levels, living in rural areas, current smoker, little social activity, lower BMI, having ADL disability, worse self-related health, and more chronic diseases shown in Table 3.

Considering the prevalence of depressive symptoms was 2.13 times higher than male, Table 4 reveals a gender-stratified analysis regarding the association between living arrangements and depression. Adjusted for all covariates, depressive symptoms were strongest in men living with a spouse and a child (OR 1.37, 95% CI 1.12–1.68) compared with those living with a spouse only. But there was no significant association between living arrangements and depressive symptoms for women.

In this study, we aimed to investigate the association between depressive symptoms and living arrangements in Chinese older adults and to distinguish the difference between men and women. Our results showed that elderly people living with a spouse and a child, or living without a spouse nor a child had higher odds for depressive symptoms compared with people living with only a spouse. We also confirmed women were more likely to have depressive symptoms but they had no significant association between living arrangements and depressive symptoms. Men living with a spouse and child were more likely to have depressive symptoms.

The results of the present study demonstrated that only living with spouse was least likely to have depressive symptoms, which is consistent with results of earlier studies [9, 36, 37]. As Chappell et al. claimed having a spouse is “greatest guarantee of support in old age” in 1991, having children in households does not add too much health benefits [38].

We examined living with/without a spouse and living with/without a child as independent variables. We found a direct relationship between such living arrangements and depressive symptoms. This association was most apparent in older adults who lived without a spouse nor a child or lived without a spouse but with a child. Since older adults were special group depending more on social support [10], living without a spouse nor a child (means living alone), was more likely to lead to depression because there was minimum social interaction. Owing to lower levels of social support for those living without spouses [39], they cannot share their emotions with another person, which may lead to mental problems. Although the odds ratio of people living alone decreased from 1.61 to 1.40 after adjusting all covariate, such living arrangement still had the highest risks of having depressive symptoms.

In addition, living with a spouse and a child had statistically significant on the risk of depressive symptoms in our study. This is a contradictory phenomenon compared to previous studies. Previous studies had reported that older adults living with a child [14, 40] were less likely to have depression. Probably, cohabitating with children may give older people a sense of pride for Chinese culture, as well as instrumental and emotional support [38]. However, in our study, the age and participants were inconsistent with previous studies which made the result contrary. During our study, the elderly might feel burdened by childcare and housework rather than receiving support from the adult child, and more conflicts might arise between and among family members in a multi-generational household [11]. It can be hypothesized that a parent who lives with a child and a spouse has to pay more attention to caring for both children and spouse, and performs housework.

During the logistic regression analysis in our study, we found that the association between living arrangements and depressive symptoms was statistically significant after controlling several covariates. We also discovered that the association still existed when we changed the cut-off point in Additional file 1: Table S1 We could draw a conclusion that such models were sensitive and robust because the association in our study was valid although we added different variables in the model and changed the coding in dependent variable.

However, the association become meaningless when we controlled the health behavior in Model 3. Taking such result into consideration, we found that the primary association in our study didn’t change in the four models, meaning that living arrangements had positive effect on depressive symptoms even though the OR became insignificant in Model 3. What’s more, the 95% CI of living with a spouse and child was (1.00, 1.32) which was so close to 1.00, meaning that such influence may be significant once we expand our sample size.

The results of our study were concordant with those of a previous study showing that women are more likely to have depressive symptoms [11, 36]. But there was no statistical significant between living arrangements and depressive symptoms in women. In terms of regular gender roles, based on traditional Chinese values, women are more likely to be responsible for the family, so they might be more likely to have depression [10, 11, 36]. However, there was no significant association between living arrangements and depressive symptoms for women. Women in China have always felt a greater responsibility to care for their families, no matter who lived with them. Therefore, the composition of the domestic unit had little impact on them. Nevertheless, men living with a spouse and a child had the strongest depressive symptoms. This may be because older men in China have always occupied the main position in the family and society. Therefore, when they live with a child and spouse, they would overthink their child, their spouse and their family’s circumstances, which might make them anxious and depressed.

Similar to previous studies, we confirmed that low SES (Socioeconomic status) especially lower education levels were more likely to have depressive symptoms [41]. And the health conditions including ADL disability, self-rate health, and the number of chronic diseases were strongly associated with depressive symptoms, similar to previous reports [10, 18, 22, 41]. To make our model more preferably, we adjusted the BMI as a covariate health condition factor in our study which was fewer included in the model. The odds of people living with a spouse and child increased from 1.15 to 1.23 after adjusting the health conditions. Although it was not our focus, we can draw a conclusion that health conditions played an important role in depressive symptoms especially in the elderly living with a spouse and child.

We identified several political implications in our study. As older people seems to be more vulnerable to loneliness and social isolation, we should pay more attention to their mental health [42] especially for people living alone or living with a spouse and child. Regarding the community, which had been proved to be the most accessible way to expand the social coverage in the older nowadays [43], the infrastructure and healthcare facilities in community should be kept more attention to be improved by policymakers because such aspects had been prove to be more likely to lead to depression in Chinese older adults [44]. In addition, since men living with a spouse and a child were more likely to be depressed, different social services should be provided according to different kinds of household composition. We should not only focus on the mental health in older people, but also concern with the child and their spouse who would influence their psychological state directly.

The strength of our study is the use of a national sample among older adults in China, so our conclusions are more representative. Previous studies had focused on either rural or urban China [14, 18, 21], but our study focused on the whole areas, making our conclusions more universal.

Several limitations of our study should be noted, however. First, the data we used in this survey were cross-sectional, not necessarily causal. We could not draw conclusions about whether depressive symptoms were due to living arrangements. In addition, we can’t avoid the occurrence of endogeneity which may come from the genetic, reverse causality and so on. However, we have made the endeavor to decrease the effect of the potential variates by adding the covariates, which had been proved to have effect on depression previously, as much as possible in Model 4. We had tried to avoid the endogeneity by sensitivity analysis we done above. Second, as our study was secondary data, the influence factor in our model is circumscribed. We focused on the association between living arrangements and depressive symptoms. So many factor likes the mainly subjective loneliness which had been proved as a predicted factors increased the depression scores were not taken into account [45]. Thus further study containing more potential risk factors should be explored in the future.

Older adults living alone or living with a spouse and child are at higher risk to have depressive symptoms. Men living with a spouse and a child are more likely to have depressive symptoms. Finding from our discovery in China may have implications for other countries which have undergone rapid urbanization and aging that older adults living with child is the main family support system. Government should better complete the age-old system of providing social services, particularly for the infrastructure and healthcare facilities in community. Different housed composition should be provided different social services not only for the older but also for their cohabitant.