Depression is a common psychiatric disorder that greatly diminishes the quality of life in adults (Noël et al., 2004). Globally, around 5 % of adults suffer from depression, with a higher prevalence of 5.7 % among adults older than 60 years (WHO, 2021). Depression has become the second leading cause of years lived with disability in China since 2010 (G. Yang et al., 2013), which has dramatically increased since the beginning of the COVID-19 pandemic (Huang and Zhao, 2020). Apart from greatly affecting an individual's health status (Kiecolt-Glaser et al., 2015), diminishing family function (Alexopoulos, 2005) and disrupting work (Hakulinen et al., 2019), the worst consequence of depression is suicide. >0.7 million people die from depression every year across the globe (WHO, 2021). Although effective treatments for depression have been established for years, over 75 % of individuals received no treatment in low- and middle-income countries in 2021 (Evans-Lacko et al., 2018). Identifying early modifiable risk factors of depression is indispensable for early detection and prevention to reduce the incidence rate and the social and economic burden for treatment.

Adverse childhood experiences (ACEs) have long-term effects on adult depression (Mandelli et al., 2015). A meta-analysis of Chinese college students reported that 64.7 % of them had suffered at least one type of childhood maltreatment from family members (Fu et al., 2018). Another study conducted in China indicated that 54.9 % of the population had been bullied by peers at some point during their childhood (Zhu and Chan, 2015). According to the socio-ecological model (McLeroy et al., 1988), family, peers and neighborhoods represent an individual's key socio-ecological systems, which impact his/her various health behaviors (e.g. unhealthy eating habits (Domínguez-Vásquez et al., 2008) and depression (Miao et al., 2019)). Several previous studies reported that poor childhood family relationships (CFR) (Parke and Ladd, 2016) childhood peer friendships (CPF) (Roosa et al., 2005) and childhood neighborhood quality (CNQ) (Hughes et al., 2017) were independently associated with adult depression. However, most of the studies only reported the risk of adult depression based on a retrospective and cross-sectional study design. Little is known about how poor childhood exposures affect adult depression longitudinally.

Poor childhood exposures often occur simultaneously and interact with each other, which can cause a vicious cycle and reduce one's coping resources chronically and comprehensively. For instance, it has been found that single-mother families were more likely to live in impoverished communities (Hoffmann, 2006). Parke et al. found that parent-child relationships and peer victimization mutually affected each other (Parke and Ladd, 2016). Roosa reported that children living in poorer-quality neighborhoods showed a closer relationship with deviant peers and more conflicts with parents, compared with children living in better-quality neighborhoods (Roosa et al., 2005). Thus, their combined exposures may increase the risk of depression. Previous studies documented that multiple ACEs significantly increased the risk of adult depression compared to those without ACE (Hughes et al., 2017). However, most of the ACEs involved in these studies came from the household, such as domestic violence, childhood abuse and neglect, and household criminality. Few studies explored the combined effects of multisystem childhood exposures, covering family, peers, and neighborhoods, on adult depression.

Several studies did report the effects of more than one childhood exposure from different social systems on adult depression. We identified two longitudinal studies that investigated childhood exposures from two systems and both were on adolescents. One study (n = 771) found that low levels of perceived family support and friendship quality at age 14 were positively associated with subsequent depressive symptoms at age 17 (van Harmelen et al., 2016). The other one reported that childhood neighborhood disadvantage and being bullied in childhood predicted future adolescent depressive symptoms (Choi et al., 2021). However, these studies only reported the independent impact of one childhood exposure with adjusting the other one as a covariate on depression, instead of assessing the impacts of combined exposures to them on subsequent depression. To our knowledge, no study investigates the longitudinal combined effects of multisystem childhood exposures from family, peers and neighborhoods, on adult depression. However, as mentioned earlier, these three are important parts of an individual's social-ecological system and are likely to influence each other. Furthermore, most of the previous studies on poor childhood exposures only tested the risk of depression as outcomes (Satinsky et al., 2021), without examining their impacts on the change of depression status (e.g. depression remission, consistent depression). These statuses reflect the dynamic process of depression, having significant clinical implications for health services and planning. It is known that low remission rates and recurrent depression are the leading cause of the increased burden of depression treatment to healthcare systems (Greden, 2001).

To fulfill these research gaps, this study aimed to test the longitudinal impacts of combined poor childhood exposures to CNQ, CPF, CFR on the level and status change of adult depression. In addition, we examined the relationships among the three types of exposures. We hypothesized that 1) CNQ, CPF, CFR would be positively correlated with each other. 2) combined poor childhood exposures in the three systems would increase the level and facilitate poor status change of adult depression.