When studying the migrant-related health disparities, scholars often use a native-born population in the host society as the reference group. Most of the studies demonstrated that migrants had poorer mental health, e.g., depressive symptoms, than native-born residents (6, 11, 29, 30). Depressive symptoms may lead to a series of complications that affect individual's quality of life and cause other critical health issues such as suicide, frailty, functional disability, and mortality (31–33). For migrants (including both international and internal migrants), depressive symptoms are also associated with their intention to settle down in the host place (4).

However, most studies on health outcomes in China only focused on the rural-to-urban migrants (6, 11, 17, 18). These studies suggest that there are significant mental health disparities within migrants. Furthermore, more information about migration status, e.g., the housing registration status of migrants, and reasons for migration, should be considered when examining the mental health disparities among migrants. In China, the most important characteristic that distinguishes the patterns of migration from those in other countries is the housing registration system (Hukou system) (13, 34). Since Hukou is linked with entitlement and benefits for individuals, whether migrants come from urban or rural is an indicator for their education background, financial status, and social support (13, 34–36), which may result in different mental health outcomes, such as depressive symptoms. To assess disparities in depressive symptoms among migrants, we proposed the following hypothesis:

There is no doubt that environmental context plays a crucial role in shaping psychological wellbeing (27, 37, 38). When investigating risk and protective factors for depressive symptoms of migrants, it is likely that community-level factors are at play.

It is well-recognized that individual and neighborhood-level factors, such as neighborhood socioeconomic status, social capital, and social support, have strong associations with residents' depressive symptoms (19–22, 37–43). A related concept of particular importance to the migrant community is social cohesion (27, 44). Social cohesion refers to communal bonds characterized by altruism, reciprocity, and shared norms and values (27). It generates mutual trust and support, collective efficacy, and a sense of belonging, which are all conducive to improved mental health (27, 45). Previous studies show that social cohesion was defined as a neighborhood-level factor, and lower neighborhood social cohesion was linked to a variably of health outcomes, such as depression and mortality (46–52).

Neighborhood social cohesion is the network of relationships, shared values, and norms of residents in a neighborhood and it shares some similarities to an individual's social network (53). Social cohesion, however, exists in a larger field. It accounts for value systems, degree of social interaction, and considers the cohesion of a broader community rather than cohesion within a neighborhood or a small group of individuals. However, little attention has been paid to the association of social cohesion, as a community-level social factor, with depressive symptoms among internal migrants in China. In a migrant community, social cohesion refers to the quality of social interactions which is closely related to individuals' mental health status (11, 20, 22, 41, 54). In a more cohesive community, migrants may feel a stronger sense of inclusion. Social cohesion has an impact on internal migrants' positive interactions including obtaining social support and help, or negative interactions such as suffering from social stress and/or exclusion from native-born residents. In a community where native-born residents and migrants trust each other, they will have more social connections and interactions. For migrants, the trust they have built will make it easier for them to establish new social ties and social networks that are important to adapt to a new life and maintain a good psychological status (20, 21, 42). In addition, a higher level of social cohesion would lead to less social stress or social exclusion, which may result in fewer depressive symptoms (6, 11, 12). Furthermore, in a well-integrated community, native-born residents and migrants are more likely to provide mutual social support, which is a protective factor against depressive symptoms (22, 41). Even if some migrants suffer from depressive symptoms, they are more likely to be able to receive timely assistance in a more cohesive community. To investigate the associations between social cohesion and depressive symptoms of both urban-to-urban and rural-to-urban migrants, here is the second hypothesis:

Although recent studies have shown the correlation between social cohesion and health outcomes (49, 50, 52, 55, 56), limited studies have been conducted to examine the mediating role of social adaptation on the association between social cohesion and migrants' mental health. Increasing attention has been paid to improve social adaptation as an effort to improve individual's mental health (57, 58). Social adaptation concerns the interactions between an individual and the environment. It refers to the performance in the activities of daily intercultural living (59) and involves the intercultural competence with emphasis on behavioral domains. As an important indicator of migrants' integration into the host community, social adaptation would mediate the pathways from social cohesion to depressive symptoms. Individual's allostatic load would be lower in a highly cohesive community where internal migrants have more positive interactions with local residents and a higher sense of being trusted and belonging, Individuals are more likely to “doing well” in the activities of daily intercultural living in the host city (59). Additionally, “doing well” in social adaptation may help buffer the effects of social stress and social exclusion they may face in the migration process (6, 11, 12); thus, avoiding the occurrence of psychological disorders. Therefore, no such studies have been conducted to examine whether social adaptation is a factor that could mediate between social cohesion and depressive symptoms. Thus, we propose the following hypothesis:

Data were drawn from the “2017 Urbanization and New Migrant Survey,” a cross-sectional survey focusing on policy issues such as population migration, social mobility, and social integration in Chinese adults. A multi-stage stratified sampling strategy was used, and data were collected from 10 cities in China, including economically developed cities (GDP per capita of is more than 100,000 RMB), such as Zhengzhou, Tianjin, Xiamen, Guangzhou, and Changsha, and less developed ones (GDP per capita of is <100,000 RMB), including Harbin, Changchun, Yanji, Shenyang, and Anshan. Data were collected during in-home interviews by well-trained interviewers. Inclusion criteria for participation were: (1) full-time residence in this city >6 months in the past year; (2) aged 18–65 years; and (3) capable of communicating answers to interview questions and giving consent.

A sample of 2,752 adult internal migrants (i.e., the movement of people between usual residences within national states) were drawn from the survey data. After excluding 168 adults with missing values, the final analytical sample consisted of 2,584 internal migrants. Among them, 1,152 reported living in another urban area before moving to their current locations, defined as urban-to-urban migrants, and 1,432 reported living in a rural area before moving to their current locations, defined as rural-to-urban migrants.

Descriptive characteristics of the final weighted analytical sample are presented by migration status in Table 1. Rural-to-urban migrants were found to have more clinically significant depressive symptoms than urban-to-urban migrants (32.86 vs. 25.12%, p < 0.001). Approximately 43.87% of rural-to-urban migrants reported low social cohesion compared to 31.96% of urban-to-urban migrants. Approximately 16.46% of rural-to-urban migrants reported fair or low social adaptation compared to 7.99% of urban-to-urban migrants. Moreover, rural-to-urban migrants were less educated than urban-to-urban migrants. They tended to stay in the hosting city for a shorter duration (12.26 vs. 15.99 years, p < 0.001).

Weighted prevalence of depressive symptoms by migration status, social cohesion, social adaptation and covariates are presented in Table 2. Results showed that being rural-to-urban migrants as compared with urban-to-urban migrants [Odds Ratio (OR) = 1.46, 95% Confidence Interval (CI) = 1.456, 1.461, p < 0.001], had lower perceived social cohesion (OR = 1.46, 95% CI = 1.458, 1.463, p < 0.001), and poorer social adaptation (OR = 1.94, 95% CI = 1.932, 1.941, p < 0.001), are associated with higher odds of having clinically significant depressive symptoms. Moreover, being married or with a partner (OR = 0.68, 95% CI = 0.679,0.682, p < 0.001) are associated with lower odds of having clinically significant depressive symptoms. Self-reported physician-diagnosed type 2 diabetes (OR = 1.71, 95% CI = 1.707,1.723, p < 0.001) and heart diseases (OR = 1.18, 95% CI = 1.172,1.183, p < 0.001) are associated with higher odds of having clinically significant depressive symptoms.

As shown in Table 3, fair or poor social adaptation was associated with a higher odd of migrants with a low perceived social cohesion for both urban-to-urban migrants (OR = 2.91, 95% CI = 2.89, 2.92, p < 0.001) and rural-to-urban migrants (OR = 4.12, 95% CI = 4.11, 4.13, p < 0.001). After adjustment for all covariates, the odds ratio reduced to 2.69 (95% CI = 2.67, 2.70, p < 0.001) for urban-to-urban migrants and 3.66 (95% CI = 3.65, 3.67, p < 0.001) for rural-to-urban migrants.

Table 4 shows the associations between social cohesion and depressive symptoms. As shown in Model 1, migrants with a low perceived social cohesion had a higher CES-D score for urban-to-urban migrants [β = 1.17, Standard Error (SE) = 0.005, p < 0.001] and rural-to-urban migrants (β = 1.81, SE = 0.004, p < 0.001) compared to those with moderate-high social cohesion. As shown in Model 2, the addition of covariates reduced the β-coefficient for low social cohesion to 0.78 (SE = 0.005, p < 0.001) for urban-to-urban migrants and 1.74 (SE = 0.004, p < 0.001) for rural-to-urban migrants. Moreover, the addition of social adaptation to Model 3 further reduced the β-coefficient for low social cohesion to 0.66 (SE = 0.005, p < 0.001) for urban-to-urban migrants and 1.41 (SE = 0.004, p < 0.001) for rural-to-urban migrants, and the associations remained significant.

Formal mediation analysis was conducted to examine the mediating role of social adaptation on the associations between social cohesion and depressive symptoms in both groups. The Sobel mediation tests were significant for both groups at the p = 0.001 level. Social adaptation partially mediated the association between social cohesion and depressive symptoms by explaining 15.39% of its effect for urban-to-urban migrants and 18.97% of its effect for rural-to-urban migrants.

Our findings that rural-to-urban migrants had a higher prevalence of depressive symptoms than urban-to-urban migrants confirmed the mental health inequalities within migrants. Whether they migrated from urban or rural may have implications for their housing registration status (Hukou), educational background, financial status, social capital, and sense of wellbeing. All these factors are related not only to the challenges they may face in the migration process but also to their abilities to overcome the challenges that could affect their probability of suffering from depressive symptoms.

The limited resources and opportunities due to their Hukou could partially explain rural-to-urban migrants' higher prevalence of depressive symptoms. Rural Hukou limits their access to both economic and social resources in the host city. In this situation, the social stress and social exclusion they experienced may have significant mental health implications (6, 10–12). Our results showed that a higher proportion of rural-to-urban migrants had less perceived social cohesion and fair or poor social adaptation than urban-to-urban migrants. Although China has gradually relaxed its policies on its Hukou system, it remains challenging for rural-to-urban migrants, especially those with low education, to settle permanently in the receiving city or are entitled to the same social benefits as urban residents (14, 35, 36). These inequalities could limit their opportunities for upward mobility and ultimately affect their mental health outcomes.

Our results also suggest that, among rural-to-urban migrants, involuntary migrated participants had higher depressive symptoms than those who voluntarily migrated. Several reasons may shed a slight on the findings. For example, these involuntary migrants may have limited social connections because family members are likely to be the only social network they can interact with. In addition, involuntary migrants have limited access to formal financial support, social capital, and health insurance, all are considered protective factors for maintaining good mental health (19).

Findings on the association between social cohesion and depressive symptoms are in line with previous studies which demonstrated that a higher level of neighborhood social cohesion is associated with better mental health outcomes (49, 50, 52, 55, 56).

Findings from linear regression models indicate that the level of depressive symptoms of rural-to-urban migrants is more dependent on social cohesion. It seems that social environmental context is even more important for rural-to-urban migrants' psychological wellbeing than for urban-to-urban migrants. In a host city, rural-to-urban migrants face much greater mental health risks than urban-to-urban migrants (6, 11, 17, 18); including (i) lack of urban Hukou that comes with a variety of economic and social resources (i.e., social security, unemployment insurance, and health insurance). (ii) Lack of sufficient level of education that could help them achieve a stable occupation and a higher socioeconomic position. (iii) Lack of sufficient social capital that is an important protective factor of psychological wellbeing. Although social capital appears to exist among rural-to-urban migrants, they have less-developed organizational social networks. Most of them rely on family members and friends from hometowns as primary personal networks (71). Thus, we speculate that they are more in need of support and help from a well-integrated community than urban-to-urban migrants. Tailored social strategies are needed to target rural-to-urban migrants for addressing mental health disparities among migrant population.

Additionally, our study identified the mediating role of social adaptation on the association between social cohesion and depressive symptoms for both groups. Reflecting the interactions between local residents and migrants, community-level social cohesion impacts whether migrants could “doing well” in the activities of daily living (59) and achieve good psychological wellbeing in the host city.

Our study findings demonstrate that social cohesion matters in mental health outcomes among migrants in China. Improving social cohesion could be an effective social strategy to promote the mental health of the migrant population. Social cohesion is undoubtedly influenced by governance, socio-economic and public policies (26). Thus, public programs and intervention strategies are needed to construct a more integrated community that serves both local residents and internal migrants.

Some limitations of this study need to be acknowledged. First, this study used cross-sectional survey data; we are only able to examine the associations of the variables of interest. First, this study used cross-sectional survey data; we are only able to examine the associations of the variables of interest. From a life-course perspective, the impact of migration and migration-related changes accumulate throughout the life-course; therefore, longitudinal studies are needed for future research. Second, this study was conducted in 10 cities, which is not representative of all migrant-receiving cities in China. Third, to better identify the pathways of social cohesion on depressive symptoms, we need to explore more mediating factors, such as social stress and social exclusion, in future studies. We thus call for future studies to develop and evaluate socio-behavioral interventions in addressing the mental health problems of the increasing number of internal migrants in China.