As complex migration trajectories become increasingly more common, understanding potential protective factors of symptoms of depression is critical. Therefore, the aim of this paper was to understand how social support and religion may be associated with better outcomes for migrants’ mental health among Haitian migrants at different points in their intricate migration trajectories. Through a quantitative cross-sectional study using a non-probabilistic sampling, we surveyed 405 Haitian migrants in Santiago, Chile (n=207), and Tijuana, Mexico (n=198). Bivariate analyses revealed that participants in Chile had a longer time post-migration (4.6 years) than those in Mexico (3.7 months) (Chi-square p<0.001). Multivariable logistic regression findings confirm social support and religion are protective factors of depression symptoms but differ depending on the point in the migration trajectory. Having someone to trust was a significant protective factor for Haitian migrants in Chile (OR= 0.187, 95% CI 0.054 – 0.646, p=0.008) and Mexico (OR=0.080, 95% CI 0.007 – 0.858, p=0.037). Identifying with a religion reduced the odds of depressive symptoms by 60.3% for those in Chile (95% CI 0.170 – 0.923, p=0.032) but not in Mexico. Findings support re-classifying protective factors such as social support as dynamic, highlighting experiences in each migration step, rather than linear or non-static. Findings provide insight into how to best prevent and treat mental health throughout migrants’ diverse journeys. Ideally, future longitudinal studies should explore the nuances of social support during different migration phases in Latin America and their relation to mental health outcomes.

Economic, political, and social unrest, in addition to climate change, have driven increasingly more individuals to migrate in recent decades, with the estimated number of international migrants globally reaching 281 million people in 2020 (International Organization for Migration, 2021). Such trends, in conjunction with changes to immigration policies in some countries (e.g., United States (US), Chile) and the COVID-19 pandemic, have led to changes in migration journeys throughout Latin America and the Caribbean, including im/mobility, uncertainty, and multiple destination (Faret et al., 2021 ; Frank-Vitale, 2020 ; Menjívar, 2006 ; Zimmerman et al., 2011).

Throughout a migration trajectory (i.e., an open spatio-temporal process of migration (Schapendonk et al., 2020 )), an individual experiences a multitude of life changes and challenges that can have a deleterious effect on their mental health (Choy et al., 2021 ; Duncan , 2015 ; Jolof et al., 2022; Temores-Alcántara et al., 2015; Virupaksha et al., 2014). This is evidenced by migrants and refugee populations globally exhibiting increased rates of depression, anxiety, post-traumatic stress disorder (PTSD), and psychosis (Bourque et al., 2011 ; Breslau et al., 2011; Bustamante et al., 2018; Leiler et al., 2019). However, it is important to note that not all migration experiences are negative and some may not impact individual’s mental health (Virupaksha et al., 2014 ). For instance, migration and mobility may translate into more opportunities, reduced discrimination, or better healthcare access ( Lee, 2008 ; Steele, 2018 ). Such differences in the impact of migration largely come from the intersection of race, class, and legal status (Waters & Kasinitz, 2021).

Poor mental health outcomes can be explained in part as a response to the multitude of stressors pre-migration (e.g., poor socioeconomic opportunities, violence) as well as stressors during transit (e.g., violence, exploitation) and lack of access to continued healthcare and denial of basic needs and rights while in transit (Chen et al., 2022; Crocker, 2015 ; Fanfan et al., 2020; UNHCR, 2016; Virupaksha et al., 2014). Furthermore, the limbo period of being in-transit is characterized by uncertainty, waiting, and at times danger ( Frank-Vitale, 2020 ). Upon arrival in a receiving country, individuals may also experience alienation, discrimination, family separation, fear of detention or deportation, and structural vulnerability ( Crocker, 2015 ; Galvan et al., 2022; Gutierrez-Vazquez et al., 2018; Keys et al., 2015; Quesada et al., 2011). An ethnographic study of Mexican immigrants in the United States described the compounding effect of such stressors throughout a lifetime on individuals’ “emotional suffering” ( Crocker, 2015 , p. 8). Throughout Latin America and the Caribbean, migrants, especially irregular migrants, are often excluded from social protection systems, including healthcare and education, due to legal and administrative barriers (Blukacz et al., 2020 , 2023; Carreño Calderón et al., 2021; ECLAC, 2021). The lack of access to these resources further exacerbates the risk of depression (Blukacz et al., 2020 ).

There has also been limited research demonstrating how mental health outcomes vary during a migration trajectory in the Latin American and Caribbean migration context (Chen et al., 2022; Gil Everaert, 2021 ). Migrants whose transit has been interrupted by uncertain mobility onward to a final destination have been found to have a higher prevalence of common mental disorders (Bojorquez et al., 2022 ). Similarly, migrants awaiting the outcome of their application for asylum or with uncertain status reported higher psychological distress compared to migrants who have received immigration decisions (Leiler et al., 2019 ; Walther et al., 2020). Additionally, Chen et al. ( 2022) found that individuals intending to migrate from an interception point to a final destination had significantly fewer depression symptoms. However, Carreño et al. ( 2023 ) found nuances in the mental health experiences in international migrant children and adolescents who migrated to Chile with narratives of fear and uncertainty but also hope while in transit. Further research is needed to fully understand mental health outcomes throughout complex migration trajectories.

There are several factors that have been found to be protective of poor mental health outcomes (e.g., education, age, language proficiency) ( Bhugra, 2004 ). For example, as demonstrated in Siriwardhana et al.’s systematic review (Siriwardhana et al., 2014 ), resilience has been found to be a protective factor for migrants and people on the move. However, not only is resilience interrelated with social networks, religion and social support have been found to mediate the relationship between resilience and mental health outcomes (Revens et al., 2021; Siriwardhana et al., 2014). Thus, the present study will focus on the role of social support and religion on migrants’ mental health status.

Though social support has been defined in many ways, it can be broadly characterized as the “support accessible to an individual through social ties to other individuals, groups, and the larger community” (Lin et al., 1979 , p. 109). Social support is often broken down into four domains: material (i.e., the exchange of resources or goods), instrumental (i.e., supporting someone perform a task), emotional (i.e., shared bonds that result in trust, understanding, and affection), and appraisal (i.e., relaying of information between people) ( Tardy, 1985 ).

Social support has long been found to be positively associated with better mental health outcomes and well-being in the general population (Kessler & McLeod, 1985). This relationship persists among migrants. Several studies have found that social support is inversely associated with different measures of psychological distress in various Latin American migrant collectives (Brunnet et al., 2018 ; Gutierrez-Vazquez et al., 2018; Mulvaney-Day et al., 2007; Revens et al., 2021). The importance of social support during migration is backed by narrative interviews of Central American women in transit through Mexico who noted social support as a key source of resilience (Lemus-Way & Johansson, 2020). Beyond emotional support, social support networks can offer appraisal support. Ethnographic fieldwork and qualitative interviews by Plasencia (2021 ) give insight into the financial support, such as job recommendations, and informational support, such as sharing how to access health resources, that peer social networks provide among an older Latinx immigrant population in the United States.

Until recently, social networks and social support have been viewed statically and linear, rather than evolving throughout a migration trajectory that may involve several steps such as immobility for unknown periods of time, deportation, and diverse destinations ( Menjívar, 2006 ; Schapendonk , 2015 ; Zimmerman et al., 2011). Limited studies on social support throughout a migration trajectory indicate that while some new relationships are added and other social ties are maintained throughout a migration trajectory, existing relationships not only evolve but can weaken (Behrendt et al., 2022 ; Ryan, 2021 ). A longitudinal qualitative study on social support throughout the migration trajectories of unaccompanied refugee minors found that the quality and contexts of existent relationships changed throughout a migration trajectory (Behrendt et al., 2022 ). Furthermore, migrants’ changing social networks can further influence their migration trajectories, at times enabling or restricting onward migration (Wissink et al., 2020 ). A wide gap in the literature remains in our understanding of how varying social support throughout a complex migration trajectory affects mental health outcomes and whether there are differences in domains of social support throughout a migration trajectory.

There has been a recent push for the increased inclusion of religion when studying social support ( Merino, 2014 ). Religion has been found to have an inverse relationship with mental health outcomes (i.e., anxiety, substance use, depression, and suicide) among adults, including in South America, though there are differences based on sex and age with older women having the strongest associations (AbdAleati et al., 2016 ; Formoso-Suárez et al., 2022; Santero et al., 2019). There are many facets of religion that can be measured, particularly intrinsic and extrinsic religiosity. Intrinsically, religion provides individuals with psychological support through coping methods, strength, and motivation (Boyas et al., 2018 ; Eppsteiner & Hagan, 2016; Silva et al., 2017).

Religion can also act as a mediator in the relationship between social networks and mental health outcomes (Revens et al., 2021; Siriwardhana et al., 2014). Extrinsic religion can offer individuals expanded nonkin social networks, closer social ties, as well as increased social support building a stronger community ( Bradley , 1995 ; Merino, 2014 ). Religious social ties have the added element of common beliefs as well as understandings and expectations about support ( Merino, 2014 ). The social networks and social capital religious affiliation offers has been found to improve both health and life satisfaction (Lim & Putnam, 2010; Yeary et al., 2012).

Thus, the present study focuses on social support as well as religion, throughout complex, interrupted migration trajectories. Specifically, this study will measure the impact of different emotional and instrumental social support as well as religious affiliation on depressive symptoms.

Our study population are Haitian migrants, who have experienced the aforementioned complex migration trajectories. Haitians had a history of migrating primarily to the Dominican Republic, however, factors like racism, discrimination, and xenophobia spurred Haitians to move to Chile and Brazil (Keys et al., 2015 ). The first wave of Haitian migration to Chile began in 2015 as, at the time, Chile offered tolerant immigration policies and economic opportunity ( Bartlett, 2021; Yates, 2021 ). By 2020, Haitians became the third highest migrant collective in Chile (Servicio Jesuita a Migrantes, 2020). However, a 2018 migration policy reform made obtaining a visa uniquely difficult for Haitians and introduced further bureaucratic challenges and barriers to accessing resources for Haitians living in Chile ( Bartlett, 2021; Carreño Calderón et al., 2021; Servicio Jesuita a Migrantes, 2020; Yates, 2021 ). Such barriers, as well as economic hardship (in part due to the COVID-19 pandemic) and rumors that the Biden administration in the United States (US) would adopt more favorable policies towards migrants, spurred some Haitians to decide to migrate to border cities, such as Tijuana, in Mexico in hopes of migrating to the US in 2021 ( Bartlett, 2021; Organización Internacional para las Migraciones (OIM), 2022 ; Yates, 2021 ). By the beginning of 2022, 36,500 Haitians were expected to arrive in cities in northern Mexico ( Rivera, 2021). This migration from Chile to Mexico is often traversed over land through South and Central America and is characterized as one of insecurity, vulnerability, and danger ( Barba-Sánchez, 2019 ; Taylor, 2022 ; Yates, 2021 ). Vulnerability continues when migrants reach the US/ Mexico border and often live in migrant camps, where they can be exposed to not only violence but the politics of humanitarian aid and ideas of deservingness ( Wurtz, 2022 ).

Uncertainty does not end once migrants reach their intended final destination though, as there have been verified reports of the deportations of Haitian migrants from the United States (Alden & Tippett, 2021; Watkins et al., 2023). Since September 2021, 2,000 Haitians who crossed the US-Mexico border have been deported to Haiti and 8,000 Haitians have been returned to Mexico (Alden & Tippett, 2021). Some of the individuals who were deported to Haiti and had a valid Chilean visa or a visa request in progress, migrated back to Chile (Watkins et al., 2023). This trajectory is one of many that Haitian migrants take. Dias et al. ( 2020 ) explore 34 other migration journeys that Haitian migrants often take and, overall, characterize Haitian migration as an “open movement” through “flexible spaces” that have been “shaped through the intricate, repeated, and habitual movements of migrants” (p. 15). Thus, it is evident that Haitian migrants’ migration trajectories are complex; their migration is often forced and is characterized by multiple interception points, re-routing, and many migration phases or steps (Dias et al., 2020 ; Schapendonk et al., 2020). Given this complexity, it is unsurprising that Haitian migrants in several different receiving countries have been found to have elevated rates of depression, anxiety, and PTSD (Brunnet et al., 2018 ; Fanfan et al., 2020; Kaiser et al., 2015).

Social support in Haiti, tèt ansanm in Creole, is a key facet of Haitian culture that migrants have reported is missing after they begin their migration trajectory (Kaiser et al., 2015 ). Tèt ansanm is characterized by reciprocal social support and solidarity (Kaiser et al., 2015 ). Despite the salience of social support in Haitian culture, little is known about how social support affects mental health in the Haitian migrant collective (Chen et al., 2022; Kaiser et al., 2015). Chen et al. ( 2022) found that social support variables had a direct protective effect on depression symptoms in Haitians living in Chile. A mixed-method study examining the role of social support among Haitian migrants living in the Dominican Republic found that a lack of social support was associated with increased depression, anxiety, and mental distress (Kaiser et al., 2015 ). However, in-depth interviews with participants from the same study also described how social networks were a source of stress (e.g., expectations of remittances, maintaining relationships from another country) (Kaiser et al., 2015 ).

Further examination of social support at different stages of a complex migration trajectory for Haitian migrants is needed in addition to investigating how different types of social support affect mental health outcomes. As detailed in Chen et al. ( 2022), many Haitian migrants in Santiago intended to leave Chile to continue migrating after almost seven years in this country. Santiago was the first point of this study and Tijuana, Mexico was the second, the United States was often the intended destination of this onward migration ( Bartlett, 2021; OIM, 2022; Yates, 2021 ). Many of the participants in this study have migrated before to the Dominican Republic or Brazil; thus, Chile was not their first step in their migration journey. Therefore, the primary objective of this paper is to investigate the relationship between social support, religion, and symptoms of depression among Haitian migrant populations in Santiago, Chile and Tijuana, Mexico.

Our analysis was rooted in Bhugra’s Migration Phases and Psychiatric Disorders framework with an added focus on the differences in social support and religion’s effects of diverse points of a migration trajectory ( Bhugra, 2004 ).

The present explicative sequential mixed-method study sought to survey and subsequently performed in-depth interviews with Haitian migrants living in Santiago, Chile and Tijuana, Mexico. Study sampling took place from February 2021 to March 2022 in Santiago and from January to March 2022 in Tijuana. Before conducting the surveys, the research team spent considerable time with the community. We built a field team composed of Haitian migrants and strengthened ties with migrants (e.g., Fundación Urgencia País). We conducted extensive ethnographic fieldwork that entailed spending time with the community, sharing food and experiences, participant observation, informal conversations with other migrant collectives who also live in the migrant camp, and in-depth interviews with 31 participants in both Chile and Mexico, which helped us inform the discussion of the present manuscript. Ethnography enabled us to capture the nuances and details of the migration trajectories in an exceptionally complex scenario (i.e., the COVID-19 pandemic and the emigration of Haitians from Chile and Brazil). The details of the ethnographic fieldwork and the community involvement can be found in McLaren et al. ( 2024). The present paper details quantitative findings from the surveys included in the larger mixed-method study.

Eligibility criteria for the surveys included (a) being 18 years or older; (b) speaking Spanish or Creole; (c) willing and able to provide informed consent; (d) born in Haiti, and (e) living in Chile or Mexico at the time of participation. Study instruments (e.g., informed consent, survey, information flyers) were prepared in Spanish and translated to Creole by the study field staff.

When the present study was designed, a power analysis was conducted to determine the necessary sample size in both Santiago and Tijuana to detect depression symptoms (Pourhoseingholi et al., 2013 ). This power analysis was based on a conservative lower prevalence of symptoms of depression (assuming a prevalence of 13.9%) as well as the recommended 95% confidence interval and effect size of 1.5 (Martinez Tyson et al., 2011; Medina-Mora et al., 2003). The sampling approach in Santiago was non-probabilistic and originally consisted of respondent-driven sampling, however, after data collection began the research team observed that participants who were intending to continue migrating onward were less likely to refer contacts to the study. To adjust for this potential bias, the research team adopted two-stage conglomerate sampling. First, the metropolitan region of Santiago was stratified into subsamples with a higher density of Haitian migrants and where research staff had community presence using public information on the presence of the Haitian community in Santiago, Chile (Contreras, 2020). Based on this, the neighborhoods Cerillos, Santiago Centro, and Cerro Navia were chosen. Additionally, a key informant was located in a large migrant camp in Santiago, and therefore, data was also collected in this community. The second stage of sampling consisted of individuals in those areas meeting eligibility criteria being invited to participate in the study.

The sampling approach in Mexico was similarly non-probabilistic. First, the research team identified areas in Tijuana where large groups of migrants congregated in their attempts to navigate the US asylum process. This was usually next to the San Ysidro point of entry where migrant shelters and local organizations were located. Second, individuals who were eligible for the study and wanted to participate were invited.

This research was reviewed and approved by the ethics review committees of Universidad Mayor in Santiago, Chile (#0169) and Universidad de Xochicalco, Tijuana in Mexico. Furthermore, this research project was designed and executed in collaboration with a Community Advisory Board (CAB) to ensure cultural competence and relevance to the Haitian migrant collective, as detailed in McLaren et al. ( 2024).

Surveys administered to the samples in Santiago and Tijuana included questions related to symptoms of depression, social support, religious affiliation, and sociodemographic variables. Variables were operationalized as follows:

Depression Symptoms: Depression symptomology was measured using the 20-items Center for Epidemiologic Studies scale (CESD-R-20), administered in Spanish or Creole ( Radloff, 1977 ). The CESD-R-20 has a range of 0-60 and it specifically measures symptoms of sadness (dysphoria), loss of interest (anhedonia), appetite, sleep, thinking, guilt, tired (fatigue), movement (agitation), and suicidal ideation ( Radloff, 1977 ). Individuals with a score of 16 or greater were coded as displaying symptoms of depression while individuals with a score less than 16 and therefore did not meet the threshold were considered to no symptoms of depression.

Social Support Predictors: Respondents were asked four dichotomous (1= yes, 0= no) questions from the Medical Outcome Study Social Support Survey (Moser et al., 2012), which were chosen based on the recommendations of the Community Advisory Board. These items asked whether respondents had someone they could trust, someone to help them when they got sick, someone to lend them money, and someone to show them love and affection. Additionally, whether an individual identified with any religion was included as a social support question.

Other Relevant Variables: Sociodemographic questions and other relevant variables for symptoms of depression were included as well. Sex (1= female, 0= male), age (in years), level of education (1= more than primary school, 0= less than primary school), if they had children (1= yes, 0= no), if they were working/employed (1=yes, 0=no), whether they identified as a migrant, Haitian, Chilean, and/or Afro-descendant (participants could identify as multiple categories; 1= identified as migrant, 0= did not identify as a migrant), and how much time they had spent in the country of interview (in months).

Analyses: Descriptive statistics were conducted to better understand the samples. Bivariate analyses were performed using Chi-square or Fisher’s Exact test for categorical variables and Pearson correlations for continuous variables to assess whether there were significant differences between those who had and did not have depression symptoms (<tabl2>Table 2</tabl2>). Given that that depression prevalence is so different between countries, and the circumstances in which the data was collected too, we chose to run two multivariable logistic regression models per country and not a model with the entire data set. Sociodemographic variables and migration related variables were then included as control variables in the multivariable logistic regression models for both Chile and Mexico. Regression models and adjusted odds ratios (AOR), 95% confidence intervals (95% CI), with p-values less than 0.05 considered to be significant. The software SPSS and R were used for this paper’s analyses (IBM Corp, 2021; R Core Team , 2022).

Handling missing cases: Multiple imputation with chained equations (MICE) were used to impute missing data (m=20). Evidence suggest that the data was missing at random. However, MICE has proved to yield to unbiased results even when the data is not missing at random (Schafer & Graham, 2002). Most variables had no missing cases or those were lower than 2%. Relative Increase variance was 0.003, showing an increase of variance below 1%.

Overall, 405 Haitian migrants participated in the survey with 207 individuals participating in Chile and 198 participating in Mexico. Many sociodemographic variables were similar in both samples (as seen in <tabl1>Table 1</tabl1>). Both samples were relatively evenly split by gender with 53.1% of the Chilean sample being male and 51.5% of the Mexican sample being male. Average age was 34.08 years old (standard deviation: 7.3) in Chile and 33.23 years old (standard deviation: 8.3) in Mexico. Whether the respondent had children was also similar (82.1% in Chile compared to 71.2% in Mexico). There were, however, also variables that differed between the samples. Notably, while 36.2% of the sample in Chile was employed, only 1% of the sample in Mexico reported being employed. Whether someone was married also differed, with nearly all (97.5%) of the sample being unmarried (compared to 46.4% in Chile). Level of education also varied between the sample with 78.8% of the sample in Mexico having completed primary school compared to only 57.0% of the Chilean sample.

There were statistically significant differences in the prevalence of depression symptoms between the country of interview; the prevalence of depression symptoms in Mexico was 92.4% compared to 37.2% in Chile (Chi-square p<0.001). The amount of time spent in the country of interview also varied significantly between Chile and Mexico, with individuals in Chile having spent an average of 56.2 months (or 4.6 years) in the country and participants in Mexico having only spent an average of 3.7 months in the country (Chi square p<0.001).

As seen in Table 2 , bivariate tests were run to assess differences between those with and without depression symptoms in each sample. In the sample in Chile, the only demographic or migration related variable that differed significantly between those who had and did not have depression symptoms was whether the respondent identified as a migrant (Chi-square p=0.024). Among the sample surveyed in Mexico, whether someone had children (Fisher’s Exact test p=0.039) and whether someone was employed (Fisher’s Exact test p=0.005) were the only demographic or migration related variables that differed significantly between those who reported and did not report symptoms of depression.

Contrastingly, bivariate tests already indicated significant differences for some independent variables in both samples ( Table 2 ). Among the sample in Chile, a lower proportion of individuals with depression symptoms (70.1%) had someone to trust (compared to 94.6% among those without depression symptoms; Chi-square p<0.001). Similarly, in Chile, a lower proportion of individuals with depression symptoms had someone to lend them money (51.9% compared to 68.5% without depression symptoms; Chi-square p=0.018), had someone to take care of them while sick (79.2% compared to 93.8%; Chi-square p=0.001), had someone to show them love and affection (80.5% compared to 96.2%; Chi-square p<0.001), and identified with a religion (72.7% compared to 86.9%; Chi-square p=0.011). However, among the sample in Mexico, only whether someone had someone to trust and whether someone had someone to take care of them while sick differed significantly between those with and without depression symptoms, with a lower proportion of those with depression symptoms reporting they had someone to trust (47.5% compared to 93.3%; Chi-square p<0.001) or had someone to take care of them while sick (32.2% compared to 80.0%; Chi-square p<0.001).

Multivariable logistic regressions were then run separately for each country (Chile: N=207; Mexico: N=198), as seen in Table 3 . All migration related (e.g., did individuals identify as a migrant, time in country) and sociodemographic variables (e.g., sex, age) were included as control variables in these models. For Haitian migrants who were interviewed in Chile, having someone to trust remained a protective factor, reducing the odds of depression symptoms by 81.3% (AOR= 0.187, 95% CI 0.054 – 0.646, p=0.008). The same was true for Haitian migrants interviewed in Mexico, where having someone to trust reduced the odds of depression symptoms by 92% (AOR=0.080, 95% CI 0.007 – 0.858, p=0.037). Religion was a protective factor for those in Chile alone. Identifying with a religion reduced the odds of depression symptoms for those in Chile by 60.3% (AOR= 0.397, 95% CI 0.170 – 0.923, p=0.034). Additionally, being married/ partnered reduced the odds of depression symptoms in Chile by 55.4% (AOR= 0.446, 95% CI 0.217 – 0.915, p=0.028) but not in the population surveyed in Mexico, and having children increased the odds of depression symptoms 4.235 times in Mexico (AOR= 4.235, 95% CI 1.005 – 17.855, p=0.049).

The present study examined the associations between different measures of social support and depression symptoms in Haitian migrants at different points during a migration trajectory. Consistent with previous literature, our findings confirm that social support and religion are protective of depression symptoms in migrant populations ( Bendixsen, 2020 ; Brunnet et al., 2018; Chen et al., 2022; Gutierrez-Vazquez et al., 2018; Moreno et al., 2020; Mulvaney-Day et al., 2007; Revens et al., 2021; Schlechter et al., 2021). However, this research also highlights the differences in depression symptoms and how protective social support and religion are at different points in a migration trajectory.

Our findings highlight the differential impact that different types of social support provide at different points in a migration trajectory. In the separate models for Chile and Mexico, having someone to trust was protective for both those in Mexico and in Chile. Conversely, being affiliated with a religion was only protective for those in Chile. The differences in social support at different points in migration trajectories emphasize the importance of considering protective factors as a temporary or varying factors throughout a migration trajectory, rather than as static factors ( Schapendonk , 2015 ). Our findings substantiate the need for future longitudinal studies and to further investigate what different forms of social support offer individuals during each migration phase. Furthermore, the nuances of each migration phase and how social interactions and support may vary over time stress the importance of studying mental health within the context of migration and mobility and emphasize the need for appropriate preventive and mental health services throughout migration trajectories.

Lastly, our findings show that while having a partner was protective in Chile, having children increased the odds of depression symptoms for those in Mexico 4.235 times (p=0.049). Though having children would seemingly contribute to an individual’s social network, having children may have increased individual’s risk for depression symptoms as parents may have been separated from their children. This would be consistent with Guo et al.’s (2016) findings that parent/ child separation among migrants increased parent’s depression. Alternatively, if parents migrated with their children, this may have increased the parent’s responsibility or worries while migrating, as children are particularly vulnerable to exploitation and health concerns while in transit (Carreño et al., 2023 ). These possibilities indicate the need for future investigation into migrant parents’ mental health. And while these characteristics cannot be altered, with further examination these findings can be used to inform mental health services for migrants.

There were significant differences in prevalence of depression symptoms based on the country of interview (37.2% in Chile vs 92.4% in Mexico, p<0.001). The prevalence of depression symptoms in both countries are higher than the prevalence of depression found in the general population during COVID-19, as Duarte and Jiménez-Molina (2021, 2022) found the prevalence of depression in Chile during COVID-19 was 22.6%. The sample in Mexico’s prevalence of depression symptoms was also higher compared to other migrant collectives during the COVID-19 pandemic, who had prevalence’s of 23% and 38.1% respectively (Attal et al., 2020; Barbato & Thomas, 2021).

We hypothesize that the high prevalence of depression symptoms in our samples, and the higher prevalence in Mexico compared to Chile, can be explained by a multitude of factors related to the migratory context in each country. Firstly, as our analyses indicate, individuals in Mexico had a more recent migration (average time in country = 3.70 months) compared to participants in Chile (average time in country 4.6 years). This is supported by Borjorquez et al. ( 2022) who found that time since migration was protective against common mental health disorders. Secondly, data collection in Chile occurred earlier in the pandemic (data collection was from February, 2021 to March, 2022 in Chile compared to January to March, 2022 in Mexico). The conditions of the COVID-19 pandemic had a serious deleterious effect on migrant’s mental health outcomes (Spiritus-Beerden et al., 2021). Furthermore, the period of time when data was collected in Mexico was particularly politically and socially tense. Speculations that the US’s immigration policy “Title 42” would be lifted caused tensions between migrant groups who were all waiting at the US-Mexico border in Tijuana. Migrants of different nationalities (e.g., Salvadoreans, Hondurans, and Mexicans) had been living in provisory precarious camps on the Mexican side of the border (El Chaparral) for over a year awaiting decisions on their requests for asylum in the US ( Ayala, 2022; Rocha Jiménez & Lazos, 2022). Violence and discrimination towards Haitians at the border was widespread and basic needs like housing and health were not being fulfilled in Tijuana ( Ayala, 2022). Thus, the resulting feelings of insecurity, precariousness, and fear may also explain the high levels of symptoms of depression reported by the Haitians interviewed in Mexico. While conducting ethnography in Tijuana, we heard from migrants that a lot of the anxiety and stress came from not having a place to sleep at night, a job, or any certainty if they could cross to the United States at some point. Additionally, individuals in Mexico had survived the treacherous journey from South America to North America, crossing the Darien Gap, which can result in complex psychological trauma ( Taylor, 2022 ).

Lastly, as previously discussed, there are differences in social support and its associated protectiveness in Chile and Mexico. A higher proportion of individuals in Chile consistently reported having different types of social support compared to those Mexico, as seen in <tabl1>Table 1</tabl1>(e.g., 92.4% of those in Chile reported having someone to care for them if they were sick compared to only 46.2% in Mexico). Since social support has been found to be key for migrants, especially while in transit and during travel, this is likely a large driving force behind the observed differences in symptoms of depression (Brunnet et al., 2018 ; Gutierrez-Vazquez et al., 2018; Mulvaney-Day et al., 2007; Plasencia, 2021 ; Revens et al., 2021). Furthermore, social support has been found to act as a moderating factor between traumatic stress and depression (Ward et al., 2018 ). More research is needed to understand this relationship between trauma, social support, and mental health in this Haitian migrant population and migrant populations that experience complex migration trajectories.

Our findings also indicate that identifying with a religion played a different protective role depending on the point in an individual’s migration trajectory, as religious affiliation was only protective for those in Chile. Religious affiliation can offer individuals a social support system, especially when individuals are away from other social networks like family and friends, as was found in Bendixsen’s (2020) ethnographic study of irregular migrants living in Norway who detailed how partaking in religious practices reconstituted their sense of belonging. Since participants in Chile had significantly more time post-migration and had built a strong community, we hypothesize that they had longer to establish social networks associated with their religious affiliation. In contrast, individuals in Mexico who were in-transit, likely relied more on intrinsic religiosity, (e.g., relying on religion for strength and motivation) rather than benefiting from the extrinsic social support of religious affiliation. This hypothesis is supported by Hovey et al.’s (2014) findings that the association between religion and mental health was mediated by social support factors.

There a few important limitations of this study to note. Firstly, the scales used to measure social support were for self-reported perceived social support. There has been some research that suggests that there are differences in the protectiveness of perceived versus received social support (Haber et al., 2007 ; Prati & Pietrantoni, 2010). Additionally, the survey did not inquire about who participants perceived the source of social support to be (e.g., kinship relationship, friend, etc.) and whether they were transnational (i.e., in another country from where they interview occurred). Some studies have found that some migrant populations do not benefit from transnational social ties ( Horst, 2006 ; Koser Akcapar, 2010). However, a study on German migrants in Great Britain found that transnational social support relationships are more effective in providing emotional support (e.g., someone to confide in) rather than instrumental support (e.g., lending money, taking care of someone when sick) ( Herz, 2015 ). Ethnographic fieldwork conducted by Warner (2007 ) on Q’eqchi’ refugees in Maya Tecún, Mexico also draws attention to the importance of kinship relationships for psychological distress in Mayan culture. Similar ethnographic studies of social support relationships among the Haitian diaspora would be beneficial.

Also, while Haitian migrants were surveyed about their religious affiliation, their beliefs in or practicing of Vodou (sometimes referred to as voodoo), an important and longstanding part of Haitian culture, were not inquired about ( Laguerre, 1973 ). It is therefore unclear if the protection of religious affiliation came from affiliating with Vodou, other organized religions, or both. The relationship between Vodou and mental health, especially as it pertains to treatment, has begun to be explored (Auguste & Rasmussen, 2019), however, future research should continue to explore the relationship between of Vodou, social support, and mental health in Haitian migrants.

In conclusion, the protectiveness of different types of social support and religion for symptoms of depression differ throughout a migration trajectory. While having someone to trust, a form of emotional social support, was protective among both those interviewed in Chile and those interviewed in Mexico, religion was protective against symptoms of depression among Haitian migrants in Chile alone. These findings indicate that identifying social support networks, or the lack thereof, could be a useful way to identify individuals vulnerable to depression symptoms and to develop programs and resources tailored to this vulnerable cohort.