Traumatic experiences and depression among Haitian caregivers: Evidence from the Grandi Byen study Authors: Michael Galvin, Rachel Beth Zimmerman, Gifty Aboagye-Mensah, Najjuwah Walden, Sherlie Jean Louis Dulience, Michelle Dorce, Patricia Kohl, Lora Iannotti, Youri Encelotti Louis, Youri Encelotti Louis Published: 2026 ABSTRACT This study investigates the relationship between trauma and caregiver depression in Haiti, a country burdened by ongoing political unrest, natural disasters, and economic hardship. A preponderance of evidence shows the substantial impact of caregiver mental health on child development and intergenerational vulnerability. This cross-sectional analysis examined data from the Grandi Byen randomized controlled trial, including 480 caregiver-infant dyads in Cap-Haitien. Depression risk was assessed using the Zanmi Lasante Depression Symptom Inventory (ZLDSI), and trauma exposure was measured with a survey adapted from the Life Events Checklist for DSM-5 (LEC-5). Negative binomial and ordinal logistic regression models assessed the relationship between caregiver trauma and depression, adjusting for demographic, socioeconomic, and environmental conditions. The analysis revealed that trauma exposure was significantly associated with higher odds of depression risk (OR = 1.09; 95% CI: 1.001, 1.193). Household composition was identified as a protective factor for depression (OR = 0.81; 95% CI: 0.664, 0.910). Trauma exposure was significantly associated with caregiver depression in Haiti, likely exacerbating the mental health challenges faced by caregivers in the context of political, economic and environmental stressors. Given the limited mental health data available in Haiti, this study provides essential insights into the trauma and challenges Haitians experience amidst ongoing crises. Impact statement Caregiver mental health is a critical yet under-researched area in global public health, especially in fragile contexts like Haiti, where political instability, economic precarity, and environmental disasters intersect to create sustained trauma exposure. Within Haiti, there is a striking lack of recent mental health research, with Haitian studies currently contributing just 0.1% of regional mental health research output, leaving a significant gap in understanding the mental health challenges in this context. This study addresses that gap by investigating the relationship between trauma exposure and depression among Haitian caregivers, using data from a randomized controlled trial, Grandi Byen. Our findings reveal a significant link between trauma exposure and depression risk, while also highlighting household composition as a potential protective factor. These insights are especially critical in a setting where formal and informal mental health services are scarce, and caregivers often carry the compounded burdens of trauma, caregiving, and economic insecurity. By centering Haitian experiences within the global mental health landscape, this study provides an evidence base to support the development of culturally grounded, community-based interventions, as well as policy and resource allocation strategies that address the intergenerational impact of trauma and strengthen caregiver well-being in high-adversity environments. Introduction Depression is a significant contributor to the global burden of disease and the primary cause of health-related disability (Liu et al., 2020). Approximately 4.7% of the global population experiences a depressive episode annually (Herrman et al., 2022; Lu et al., 2024). Despite its prevalence and impact, access to effective care remains alarmingly low, with over 80% of people in high-income countries and over 90% of people in low- and middle-income countries (LMICs) not receiving adequate care (Vigo et al., 2022). These disparities are particularly acute in Haiti, where recent crises have exacerbated mental health challenges faced by the population. While scientific data remains limited in this context, one study found mental, neurological, substance use disorders and suicide collectively account for 9% of all disability-adjusted life years in Haiti, highlighting their substantial health burden (Pan American Health Organization, 2020). Trauma is a well-documented risk factor for depression. Its psychological consequences are often severe and enduring, with intergenerational impacts further compounding these effects (United States Center for Substance Abuse and Mental Health Services Administration, 2014). In Haiti, nuanced forms of trauma arise from complex and context-specific circumstances, including a history of political turmoil, natural disasters, limited resources, and persistent instability in the country. Recent overlapping crises in Haiti have placed caregivers, particularly mothers, under additional significant strain, exacerbating their already vulnerable position within Haitian society. Caregivers often bear a unique and compounded burden, shaped by persistent economic stress, entrenched gender inequalities and inadequate social support systems (Padgett and Warnecke, 2011; Cerdá et al., 2013; Pan American Health Organization, 2020; Singh et al., 2020; Usher et al., 2020; Lu et al., 2023; Rodrigues et al., 2023; El Khoury-Malhame et al., 2024). The intersection of these challenges not only heightens their susceptibility to depression but also has profound implications for their children’s well-being. Caregiver mental health is a critical determinant of child development, and the adverse effects of caregiver trauma and depression may ripple across generations, perpetuating cycles of vulnerability and hardship (Santos et al., 2024). These dynamics underscore the urgency of addressing the mental health needs of caregivers to safeguard both their well-being and that of their children. Given the numerous humanitarian crises and environmental disasters in Haiti in recent years, research on mental health remains limited and insufficient, with Haitian studies contributing 0.1% of the regional mental health research output (Humanitarian Coalition, 2024). These persistent challenges – natural disasters, political instability, economic hardship, and other longstanding structural issues – have significantly worsened mental health issues in Haiti. Expanding research efforts is essential to comprehensively assess and respond to the mental health needs of the Haitian population. The current study addresses a critical gap by investigating the trauma-depression relationship in Haiti, providing much-needed evidence to inform interventions tailored to this context. This study aimed to (1) explore the risk of depression among caregivers in Haiti, (2) investigate the relationship between their traumatic experiences and risk of depression and (3) identify key factors influencing the relationship between trauma and depression within this context. Methods This study is a cross-sectional, observational analysis of the baseline data from the Grandi Byen trial. Setting The study activities were conducted in the Fort Saint Michel (FSM) Health Center in Cap-Haitien, Haiti. Cap-Haitien, the capital of the North Department (Nord), is the second largest city in Haiti after Port-au-Prince. FSM is located in Petite Anse, the city’s most impoverished commune, which is home to over 90,000 people. The area is densely populated and lies near a major road, a canal and an international airport. Due to its low-lying topography and inadequate water management infrastructure, the region is prone to flooding and mudslides during the rainy season. Healthcare access is limited, with one public hospital and two public clinics serving the population. The nearest mental health center, Sant Sante Mantal Mòn Pele, is ~7 km from FSM. Grandi Byen Grandi Byen is a three-arm longitudinal randomized controlled trial evaluating the impact of an integrated nutrition, responsive parenting and WASH (water, sanitation and hygiene) intervention on holistic child growth and development. Although the primary focus of the trial was child growth and development, a mental health component assessing caregiver depression and trauma was incorporated into the approved study protocol at a later stage of implementation. Six hundred caregiver-infant dyads living in Cap-Haitien, Haiti, and in the surrounding communities of Petite Anse, FSM, and Madeleine were recruited and randomized into one of the following groups: (1) standard well-baby care; (2) nutritional intervention (one egg per day for 6 months); and (3) multicomponent Grandi Byen intervention (12-week responsive parenting, nutrition, WASH training and one egg per day for 6 months). More information regarding the protocol of this study can be found in the referenced paper (Kohl et al., 2022). Caregiver-infant dyads were recruited through referrals from community health agents, maternity and well-baby units at Fort Saint Michel clinic, and community rally posts. Caregivers were eligible to participate in the Grandi Byen Trial if they were at least 18 years of age and served as the primary caregiver of an infant between 6 and 8 months old. Dyads were excluded if infants exhibited congenital medical conditions, severe disabilities or severe acute malnutrition (defined as a weight-for-length z-score below −3). Additional exclusion criteria included known allergies to animal-source foods (including eggs, milk or fish) and being part of a multiple birth (e.g., twin, triplet, etc.). The mental health assessments were introduced beginning with the second cohort; the first cohort did not participate in depression or trauma screening. As a result, the present analyses include a subsample of 480 caregivers across four cohorts who completed the mental health measures. For clarity, these cohorts are referred to as Cohorts 1–4 in the present study, although they correspond to Cohorts 2–5 of the original Grandi Byen Trial. Eligible participants for the mental health portion of the study included adult caregivers responsible for the enrolled child’s daily care, including mothers, fathers, grandmothers and other relatives. All participants provided written informed consent, which was amended before mental health data collection to include depression and trauma screening. Ethical approval for the Grandi Byen Trial, including the mental health component, was obtained from the Bioethics Committee of the Ministry of Public Health and Population (MSPP) in Haiti (#D) and the Institutional Review Board/Human Research Protection Office at Washington University in St. Louis (# 202101035; FWA00002284). This project is funded by the Eunice Kennedy Shriver National Institute of Child Health and Development (NICHD) (R01HD098255-02) and was registered on March 5, 2021, at https://clinicaltrials.gov/ (NCT04785352). Procedures were in place to ensure participant confidentiality, secure data storage and restricted access to data. Caregiver depression The outcome for this analysis was caregiver depression. Depressive symptoms were measured by the Zanmi Lasante Depression Symptom Inventory (ZLDSI) (Rasmussen et al., 2015a; Legha et al., 2020). This 13-item screening tool identifies and assesses depressive symptoms using expressions of depression that are relevant to the context of Haiti (Supplementary Table S1). Each of the 13 items corresponds to a clinical depressive symptom such as “low energy,” “crying or feeling like crying” and “difficulty sleeping without waking early.” Items are scored on a 4-point scale (0–3), where 0 represents the absence of the symptom over the past 2 weeks, 1 indicates experiencing the symptom for 1–5 days, 2 for 6–9 days and 3 for 10–15 days. Total scores range from 0 to 39, with a score of 13 or higher indicating a threshold for potential depression, warranting further clinical evaluation or referral. Those participants who scored above the threshold were referred to the Mental Health Center at Morne Pelé. For ordinal regression models, depression scores were separated into four categories: 0–9 points, 10–19 points, 20–29 points and 30–39 points. Caregiver trauma The primary predictor for this analysis was caregiver trauma. Caregiver trauma was measured through a survey adapted from existing trauma screeners, including the Life Events Checklist for DSM-5 (LEC-5) and informed by existing qualitative research of relevant traumatic events in Haiti (i.e., earthquakes and hurricanes) (Bolton and Gray, 2006; Grelotti et al., 2018). The 21-item survey includes 14 questions about specific traumatic experiences the caregiver may have encountered during their lifetime and 7 additional questions focused on traumatic experiences specifically related to or resulting from both the 2010 and 2021 earthquakes (Supplementary Table S2). Each item allowed the caregiver to respond whether they had experienced a particular event, and, if so, to rate how distressing it was. The response options were: “No,” “Yes, but no problem,” “Yes, a little problem,” “Yes, many problems” and “Yes, a big problem, and I couldn’t handle it.” For analysis, the total number of traumatic events experienced by each individual was summed. Small modifications were made based on the advice of local experts to be more relevant to northern Haiti, such as the removal of a question about combat or exposure to a war zone, as there is no war in this region. Other questions were added, such as being the victim of a mudslide – which is common in the hilly terrain of northern Haiti – or being the victim of a kidnapping, a common form of extortion by gangs in the region. Demographic, socioeconomic and environmental predictors Demographic, socioeconomic and environmental predictors regarded pertinent to control for after stepwise selection – a statistical method that iteratively adds or removes variables based on their contribution to the model’s explanatory power – include the caregiver’s age, caregiver’s education level, caregiver’s marital status, number of adults living in the caregiver’s household, caregiver’s home ownership status and study cohort. Participants of this study may have experienced traumatic events related to significant political or social upheaval, depending on the period they were involved in the study, as the participants’ period of involvement was defined by their cohort (Figure 1). Cohort 4 was specified as the reference group, as this group experienced the lowest average depression symptoms of the four cohorts.Figure 1.Timeline of mass traumatic events during the project period. Figure 1. long description.A vertical timeline from March 2022 to early 2025 showing traumatic events across four study cohorts. Events include disease outbreaks, natural disasters, and political instability. See long description. Timeline of mass traumatic events during the project period. Figure 1. long description. Statistical analyses Caregiver characteristics were summarized and evaluated to provide context for the baseline population and identify variables relevant to the analyses. All statistical analyses were conducted in R 4.4.2. Bivariate linear models were employed to assess the association between caregiver depression and each individual predictor, while general linear regression models were used to examine the relationship between caregiver depression, caregiver trauma and the demographic, socioeconomic and environmental predictors mentioned above. Analyses followed a two-stage approach. First, the number of traumatic events experienced by caregivers was modeled as an outcome to examine demographic, socioeconomic and cohort-related factors associated with trauma exposure. Due to overdispersion in Poisson regression models, negative binomial regression was used for this analysis (Cairney and Boyle, 2004). Second, caregiver depression was modeled as an ordinal outcome, with traumatic events included as a key predictor alongside demographic, socioeconomic and environmental variables. Ordinal logistic regression was used to assess associations across increasing levels of depressive symptom severity. Stepwise selection was applied to identify these key variables for inclusion in all models. Diagnostics assessed collinearity and model fit using likelihood ratio tests and information criteria (Akaike and Bayesian Information Criteria [AIC and BIC, respectively]); overdispersion and residual diagnostics were evaluated for the negative binomial model, and the proportional odds assumption was assessed for the ordinal logistic model. Results Population characteristics The sample includes 480 caregivers, with the majority being mothers (92.1%, 442 caregivers) and female (96.7%, 464 caregivers) (Supplementary Table S3). Two-thirds of caregivers were unmarried, with an average age of 29 years. Over three-quarters of caregivers completed secondary school. Caregivers averaged 2 days per week working outside the home. Households were typically composed of a few adults and children, and 64% of caregivers reported an income of 10,001 or more HTG, with a 62% receiving money transfers from abroad. Homeownership was relatively uncommon; less than one-third of households owned their home, while the majority rented, shared or borrowed housing. Only 16% had access to electricity, and cooking practices primarily involved methods requiring more time or labor, such as cooking with charcoal or wood. The vast majority of homes had concrete floors at 83%, and a slight majority had aluminum or tin roofs. Water treatment was infrequent, and nearly three-quarters of households had outdoor toilets. Caregiver depression and trauma The depression scores among the caregivers revealed that 14.8% of participants (71 caregivers) were identified as being at high risk for depression, scoring 13 or higher with ZLDSI (Table 1). The mean depression score (x̄ = 5.87) across the population was below the threshold (13 or higher) in all cohorts (Cohort 1 x̄ = 5.66; Cohort 2 x̄ =6.40; Cohort 3 x̄ = 6.10; Cohort 4 x̄ = 5.33). The most common symptoms experienced among caregivers were “thinking too much” (49.8%, 239 caregivers), “feeling you have a constricted heart” (44.2%, 212 caregivers), “feeling tired or having little energy” (40.2%, 193 caregivers) and “low energy” (39.6%, 190 caregivers) (Supplementary Figure S1).Table 1.Depression and trauma among the participant mothers Table 1. long description.(n = 480)Depression score (SD)5.87 (7.83)Risk of depression (%)High risk of depression (depression score exceeding the cutoff)14.8Low risk of depression (depression score below the cutoff)85.2Number of traumatic events experienced (SD)3.00 (2.50)Type of event experienced (%)Motor or car accident16.2House fire7.1Hurricane or flood35.6Mudslide3.54Severe illness11.5Loss of livelihood25.8Robbed with a weapon10.4Rape1.7Witness a motor or car accident24.4Sudden death of a family member or close friend38.5Witness a person harmed9.4Witness a sudden death14.8Family member disappeared or was kidnapped7.1Earthquake34Neighborhood damaged2.3House destroyed1Serious injury2.1Death of a family member or close friend14.2Belongings lost2.7Forced to move1.7Housed displaced persons1.5Depression symptom experienced (%)Low energy39.6Feeling you have a constricted heart44.2Thinking too much49.8Crying or feeling like crying33.5Feeling you have lost the taste for doing anything38.5Feeling down, discouraged, or totally hopeless37.5Difficulty falling asleep34.6Feeling tired or having little energy40.2Having no appetite30.4Feeling you are a failure or feeling bad about yourself31.7Moving or speaking so slowly that people have noticed24.6Thoughts that you would be better off dead, or of hurting yourself in some way18.3Difficulty sleeping without waking early25.6A data table detailing depression scores, trauma history, and symptom prevalence among 480 participant mothers. See long description. Table 1.Depression and trauma among the participant mothers Table 1. long description.(n = 480)Depression score (SD)5.87 (7.83)Risk of depression (%)High risk of depression (depression score exceeding the cutoff)14.8Low risk of depression (depression score below the cutoff)85.2Number of traumatic events experienced (SD)3.00 (2.50)Type of event experienced (%)Motor or car accident16.2House fire7.1Hurricane or flood35.6Mudslide3.54Severe illness11.5Loss of livelihood25.8Robbed with a weapon10.4Rape1.7Witness a motor or car accident24.4Sudden death of a family member or close friend38.5Witness a person harmed9.4Witness a sudden death14.8Family member disappeared or was kidnapped7.1Earthquake34Neighborhood damaged2.3House destroyed1Serious injury2.1Death of a family member or close friend14.2Belongings lost2.7Forced to move1.7Housed displaced persons1.5Depression symptom experienced (%)Low energy39.6Feeling you have a constricted heart44.2Thinking too much49.8Crying or feeling like crying33.5Feeling you have lost the taste for doing anything38.5Feeling down, discouraged, or totally hopeless37.5Difficulty falling asleep34.6Feeling tired or having little energy40.2Having no appetite30.4Feeling you are a failure or feeling bad about yourself31.7Moving or speaking so slowly that people have noticed24.6Thoughts that you would be better off dead, or of hurting yourself in some way18.3Difficulty sleeping without waking early25.6A data table detailing depression scores, trauma history, and symptom prevalence among 480 participant mothers. See long description. Depression and trauma among the participant mothers Table 1. long description. A total of 395 caregivers reported experiencing at least one traumatic event, indicating that a significant proportion of the sample (~82%) had exposure to trauma. On average, participants reported approximately three traumatic events, reflecting a notable level of cumulative adversity within the group. The most common traumatic events experienced were the sudden death of a family member or close friend, hurricane or flood, earthquake and loss of livelihood (Table 1). Regression models: Caregiver trauma and depression We used stepwise regression and the evidence base to test covariates with the potential for an association with trauma and depression. Those covariates showing significant associations and/or improving model fit were retained. The dyad characteristics assessed in these models, including the child-level characteristics, are reported as descriptives now in Supplementary Table S3. The first model (Table 2) explored the relationship between traumatic events (dependent variable) and demographic and socioeconomic exposures (independent variables). Cohort 1 was the only statistically significant predictor of trauma among Grandi Byen mothers (OR = 0.888, 95% CI: [0.794, 0.992], p = .035), indicating that mothers in Cohort 1 had lower odds of experiencing traumatic events compared to the reference cohort. Caregiver characteristics, including age, education level, marital status and household factors such as home ownership and the number of adults in the home, were not significantly associated with trauma levels.Table 2.Regression model for factors associated with trauma among Grandi Byen mothers Table 2. long description.CoefficientsEstimateStd. error p-value95% CI (Lower)95% CI (Upper)(Intercept)2.880.2570.00011.764.72Cohort 10.8880.05660.0351 0.7940.992Cohort 21.010.03260.6570.9521.08Cohort 30.9930.02310.7680.9491.04Caregiver’s age1.010.006100.07890.9991.02Caregiver’s education level0.9390.04120.1290.8681.02Home ownership0.9470.06060.3730.8391.07Number of adults in the home1.010.02940.8100.9541.06Caregiver’s marital status0.9970.09390.9740.8301.20A regression model table showing factors associated with trauma among Grandi Byen mothers across nine variables including cohorts and demographics. See long description. Note: Bold values are significant p-values (p > .05) Table 2.Regression model for factors associated with trauma among Grandi Byen mothers Table 2. long description.CoefficientsEstimateStd. error p-value95% CI (Lower)95% CI (Upper)(Intercept)2.880.2570.00011.764.72Cohort 10.8880.05660.0351 0.7940.992Cohort 21.010.03260.6570.9521.08Cohort 30.9930.02310.7680.9491.04Caregiver’s age1.010.006100.07890.9991.02Caregiver’s education level0.9390.04120.1290.8681.02Home ownership0.9470.06060.3730.8391.07Number of adults in the home1.010.02940.8100.9541.06Caregiver’s marital status0.9970.09390.9740.8301.20A regression model table showing factors associated with trauma among Grandi Byen mothers across nine variables including cohorts and demographics. See long description. Note: Bold values are significant p-values (p > .05) Regression model for factors associated with trauma among Grandi Byen mothers Table 2. long description. Note: Bold values are significant p-values (p > .05) The second model (Table 3) investigating the relationship between depression risk (outcome), trauma and demographic and socioeconomic predictors reveals that a higher number of traumatic events experienced is associated with an increased risk of depression. Cohort membership showed limited significant effects, with only one cohort exhibiting a marginally higher risk (Figure 2). Interactions between trauma and cohort membership were not associated with depression risk. A larger number of adults in the household was associated with a decreased depression risk. Caregiver characteristics, including age, education, marital status and home ownership, were not significantly associated with depression risk.Table 3.Regression model for factors associated with depression risk among Grandi Byen mothers Table 3. long description.CoefficientsOdds ratioStd. error p-value95% CI (Lower)95% CI (Upper)Traumatic events1.09320.04460.0455 1.00071.1925Cohort 11.37650.23780.17910.86782.2160Cohort 20.99800.14480.98890.74611.3194Cohort 31.02820.10980.80020.82411.2700Caregiver’s age1.00260.01710.87770.96931.0365Caregiver’s education level0.90820.11360.39630.72411.1315Home ownership0.76620.16740.11160.55081.0629Number of adults in the home0.81350.09690.0331 0.66390.9695Caregiver’s marital status0.65570.26600.11260.38531.0961Traumatic events*Cohort 10.93740.05430.23420.84121.0421Traumatic events*Cohort 21.01690.03370.61880.95201.0869Traumatic events*Cohort 30.96600.03030.25450.90811.0240A regression model table showing factors associated with depression risk among Grandi Byen mothers. Significant factors include traumatic events and the number of adults in the home. See long description. Note: Bold values are significant p-values (p > .05) Figure 2.Relationship between depression risk and trauma across cohorts. Figure 2. long description.A scatter plot with linear regression lines showing a positive correlation between traumatic events and depression scores across four cohorts. See long description. Table 3.Regression model for factors associated with depression risk among Grandi Byen mothers Table 3. long description.CoefficientsOdds ratioStd. error p-value95% CI (Lower)95% CI (Upper)Traumatic events1.09320.04460.0455 1.00071.1925Cohort 11.37650.23780.17910.86782.2160Cohort 20.99800.14480.98890.74611.3194Cohort 31.02820.10980.80020.82411.2700Caregiver’s age1.00260.01710.87770.96931.0365Caregiver’s education level0.90820.11360.39630.72411.1315Home ownership0.76620.16740.11160.55081.0629Number of adults in the home0.81350.09690.0331 0.66390.9695Caregiver’s marital status0.65570.26600.11260.38531.0961Traumatic events*Cohort 10.93740.05430.23420.84121.0421Traumatic events*Cohort 21.01690.03370.61880.95201.0869Traumatic events*Cohort 30.96600.03030.25450.90811.0240A regression model table showing factors associated with depression risk among Grandi Byen mothers. Significant factors include traumatic events and the number of adults in the home. See long description. Note: Bold values are significant p-values (p > .05) Regression model for factors associated with depression risk among Grandi Byen mothers Table 3. long description. Note: Bold values are significant p-values (p > .05) Relationship between depression risk and trauma across cohorts. Figure 2. long description. Discussion This study examined a large sample of primarily female Haitian caregivers living in extreme poverty, with little access to basic necessities such as electricity, gas for cooking, potable water, and indoor plumbing. Amidst a failing state and widespread insecurity, the vast majority of these caregivers had at least one traumatic event that significantly impacted their mental health, with an average of three per caregiver. Related to these traumatic events, this research also found relatively elevated levels of risk for depression among a sample of non-care-seeking caregivers, with many expressing both somatic and affective symptoms of depression. While some cohorts of this study exhibited higher rates of depression than others – potentially related to proximity to political violence in the country – overall models highlighted a strong relationship between traumatic experiences and depression, as well as poverty-related indicators – such as having a larger number of adults in the home – and depression. Due to ongoing political and social crises in Haiti, few studies have been able to examine mental health among large population samples in recent years. Understanding the severity of the political and social upheaval that affected the lives of every Haitian is essential for the interpretation of this study’s findings. Recruitment for this study began 8 months after the assassination of Haitian President Jovenel Moïse on July 7, 2021, an event that set in motion the collapse of the state and the eventual takeover of large parts of the country by armed gangs. Moïse’s death was followed by the unelected appointment and resignation of Prime Minister Ariel Henry, and a 7.2 magnitude earthquake in the South Department (Diaz, 2021; Boulet-Groulx, 2022; United Nations Office for the Coordination of Humanitarian Affairs [OCHA], 2022). In addition to the political events, since March 2023, the country has also experienced torrential rainfall resulting in increased flooding, landslides, cholera, dengue fever and skin infections, in addition to being affected by hurricane Melissa last year (Crisis24, 2023; Chéry, 2024a, 2024b, 2024c, European Commission Humanitarian Aid and Civil Protection (ECHO), 2024; Pan American Health Organization (PAHO), 2023; Partners In Health (PIH), 2023). Later in 2023, the Dominican Republic–Haiti border closed, and the Dominican Republic government suspended issuing visas to Haitians, preventing work and essential food and medicine supplies from crossing the border (Assessment Capacities Project (ACAPS), 2023; Famine Early Warning Systems Network (FEWS NET), 2023; Kanno-Youngs and Enecia Pérez, 2023). Since then, nearly 71,000 deportations from the Dominican Republic to Haiti have taken place, and nearly 1 in 2 Haitians reported experiencing hunger as of October 2024 (Enecia Pérez and Robles, 2024; World Food Programme (WFP), 2024). In response to political unrest, gang activity has also increased as a concerted effort to control Port-au-Prince, as well as other areas of the country, such as Artibonite (U.S. Committee for Refugees and Immigrants (USCRI), 2024). Gang attacks and barricades have caused many schools, hospitals and banks to close, and an estimated 700,000 people were internally displaced across the country (IOM, UN Migration, 2024; Sanon and Luxama, 2024). The consequences of these challenges have both directly and indirectly affected mental healthcare access, resulting in widespread suffering in the population (Cénat et al., 2025). As it has now been several years of unusually high levels of instability and lawlessness, impacts are compounded and affect the most vulnerable – notably women and children, as highlighted in this study – even more significantly. This study’s assessment of mental health among 480 primarily female caregivers highlights a population that is largely very poor, with little formal education. Additionally, this is a sample of caregivers who were largely unmarried, therefore often struggling to provide for their children while living on a single income. Previous research in Haiti has found a strong link between poverty and mental health, underscoring the important role that poverty can play in factors that can lead to mental distress (Smith Fawzi et al., 2012; Wagenaar et al., 2012; Eugène, 2020; Cénat et al., 2022). In such conditions, social support can play a significant role in supporting single mothers, and having more adults in the household was found to be protective for depression in this study. This underlines the unique role that social support can play in bolstering mental health, particularly among unmarried female caregivers who often need help raising and supporting their young children. Particularly in the current context of Haiti, where the state has effectively ceased functioning, vulnerable individuals are most reliant on social support from friends and family members for survival. Additionally, several studies have stressed the importance of social support with regard to preventing depression among similar populations (Amédée et al., 2024; Hogan et al., 2024). With regard to traumatic experiences, many of the caregivers reported events related to violence, displacement, and economic loss that influenced depression risk. Several studies have documented high levels of traumatic events among Haitians, and particularly among caregivers (Jaimes et al., 2008; Auguste and Rasmussen, 2019). A strong relationship between trauma and poor mental health has long been established in the scientific literature, further reinforced by the findings of this study (Martsolf, 2004; Belik et al., 2007; Byansi et al., 2023). In particular, a significant percentage of study participants reported being robbed with a weapon, having a family member kidnapped or disappeared or being the victim of rape, highlighting increases in crime and violence following the collapse of the state. In addition, over one-third said they had been the victims of a natural disaster such as a hurricane, flood, or earthquake. Lastly, over a quarter reported a loss of livelihood due to events outside their control. Previous mental health research in Haiti found similar rates of natural disaster and violence-related traumatic experiences (Bolton et al., 2012; Brewis et al., 2022). Overall, this study seeks to contribute to the limited mental health literature in Haiti that focuses on interventions for caregivers, with the goal of developing future longitudinal and mixed-methods research in this domain. Limitations and strengths This study has several limitations. The data collection period overlapped with the onset of several significant periods of political and social upheaval, including a presidential assassination, environmental crises, and lockdowns. However, the current study does not examine how the timing of these events correlates with the onset of depressive symptoms. Instead, it focuses on the overall number of traumatic events and their contribution to differences in depression onset and severity. Given that data on depressive states among this population before these events is unavailable, this limitation highlights the need for future research that evaluates the impact of trauma on depression over time. Although we cannot assess changes before and after these events, understanding the duration of residual trauma is critical to understanding the progression of depression and the timeliness of depression treatment (Wang et al., 2023). As this study was quantitative in nature, it primarily focused on measurable outcomes, leaving a gap in understanding the unique emotional and physical experiences of each caregiver. A qualitative evaluation of psychological and psychosomatic symptoms could have provided deeper insights. Additionally, the study was limited in its ability to quantify distress or trauma severity, as no validated protocol exists for scoring the available response categories; consequently, we relied on event count alone, which may not capture differences in distress intensity. In addition to limitations, this study also has several strengths. First, this research interviewed a relatively large sample of participants compared to previous studies in this region. Next, this study is important as it examined the mental health of primarily female caregivers in northern Haiti. Studies in Haiti and elsewhere around the world have shown significantly higher rates of depression among women compared to men – with mothers at even higher risk (Patel et al., 2002; Wagenaar et al., 2012; Rasmussen et al., 2015b). Therefore, this research focuses on an at-risk group with regard to mental health. Finally, while several studies have examined mental health in the south of Haiti (Hagaman et al., 2013; Blanc et al., 2016; Cadichon et al., 2017), only recent research by one study team has begun to examine it in the north of the country (Galvin et al., 2023). This research, therefore, fills an important gap in assessing mental health impacts in this region of Haiti. Conclusion This study explored the impact of trauma on caregiver mental health in Haiti and revealed that trauma exposure is a major factor contributing to caregiver depression in Haiti, with significant implications for child development. Social support factors were identified as protective elements, suggesting that community and familial support can help mitigate some of the adverse effects of trauma. Given the limited mental health data available in Haiti, this study provides essential insights into the trauma and challenges that Haitians experience amidst ongoing crises. Additionally, this work has important clinical significance, as formal and informal mental health services are extremely limited. Ultimately, given the profound effects of caregiver depression on child development, interventions tailored to the socio-cultural context of Haiti are essential to addressing the long-term impact of trauma and improving outcomes for both caregivers and children. Supporting information Galvin et al. supplementary material Galvin et al. supplementary material Thank you for the opportunity to review this manuscript. It is timely, given the many structural challenges facing caregivers in Haiti. Some limitations need to be addressed to strengthen the manuscript. 1. Please include the consent procedures and ethical approval. Were any procedures in place for caregivers with clinical levels of depression? Did you provide any incentives for participation in the study? How were participants recruited in the main study? 2. For all the measures, it is essential to indicate their validity and reliability in this context. For the depression scale, what do the four categories correspond to or mean in terms of depression risk? 0-9 points, 10-19 points, and what is the clinical relevance of this categorization? 3. In the analytical approach, it is not clear why the negative binomial analysis was used on a predictor variable. Depression (ordinal variable) was the outcome, and traumatic events (count) were the predictor variable. The authors need to clarify why NBR was used in the analysis. 4. There is a typo on page 6, line 46. Additionally, please specify the currency for the income. In the population’s characteristics, the authors write “with a significant portion receiving money transfers.” This seems incomplete, from whom or where? 5. In the first model, page 7, line 18. “experiencing a greater number of traumatic events was significantly associated with higher trauma levels”. This sentence is confusing and not supported by Table 2. Please clarify? 6. In the discussion, the authors argue that they found a high risk for depression. Given the low mean (5), I would argue that this level is not high risk. The average in the sample is below the threshold of 13. I would suggest you tone down the language to reflect this overall statistic. It would be important to comment on the clinical significance of this work in the context of Haiti, where formal and informal mental health services are severely limited. 7. In the background, the authors document the impact of structural challenges on the mental health of caregivers. Although these are not measured, I would strongly suggest that the authors highlight how these impact overall health in Haiti specifically (poverty, natural disasters, political violence, etc.). The authors have primarily focused on political violence. The strength of this paper is that it addresses a critical and underexplored public health issue by examining the relationship between trauma exposure and caregiver depression in Haiti, a setting marked by compounded political, environmental, and economic stressors. This research specifically contributes novel evidence to a severely limited body of mental health research in Haiti and offers valuable insights with implications for caregiver support, child development, and intergenerational health equity. METHODS : Although the authors make reference to the original protocol for the study, basic characteristics of the study sample should be presented in the methods section including participants Eligibility and inclusion criteria, age range, and caregiving role definitions. Ethical Approval, local ethics approval in Haiti and Informed Consent procedures should be described in the methods section. Lines 9- 19: The covariates / confounding variables section would benefit from having a description of the children’s characteristics in the dyad sample. Also, an explanation of Why specific covariates were selected could strengthen the paper. Lines 32-56: Your measures are well chosen and contextually appropriate, but the section would benefit from greater methodological transparency. The authors alternate between “caregiver” and “maternal” depression/trauma. It is unclear whether all caregivers are mothers, or mothers are a subgroup of caregivers. Specify the timeframe of measurement (Lifetime, adult-only, or since caregiving began) to establish causal interpretation in trauma–depression analyses. Line 50: You state the trauma survey was “adapted” from the LEC-5 but do not describe how items were selected or modified; and whether local experts or pilot testing informed adaptation. Scoring and Operationalization of Trauma, it is unclear why distress severity was collected but not used; and whether earthquake-related items were included in the total score or analyzed separately. A brief justification for using event count versus severity, and clarification of scoring decisions would help establish construction and analytic justification. RESULTS: The Results section requires substantial revision to ensure that all analyses are clearly and consistently reported. Specifically, relevant statistical information should be explicitly incorporated for each analysis conducted. Strengthening the presentation of statistical results will improve clarity, transparency, and interpretability of the findings and ensure alignment with the journal reporting standards. Line 46: Check spelling for ”reporedt”. STRENGTHS AND LIMITATIONS: The strengths of the study is described as part of the discussion section. However, the manuscript would benefit from the addition of clearly delineated Strengths and Limitations sections. Including these sections would strengthen the overall interpretation of the findings, align the manuscript with standard reporting practices, and enhance its credibility for both research and practice audiences. DISCUSSION: The Discussion section provides a rich and well-documented contextual narrative that effectively situates the findings within Haiti’s complex political, environmental, and humanitarian landscape. A major strength of this section is its clear articulation of how cumulative trauma, poverty, and instability intersect to shape caregiver mental health, supported by extensive and relevant citations. That said, the Discussion would benefit from greater focus and synthesis to more clearly center the study’s primary findings. At times, the extensive background detail on political and environmental crises overshadows the interpretation of statistical results. Consider streamlining contextual descriptions and more explicitly linking them back to the key quantitative findings (e.g., trauma–depression associations, protective role of household composition). Additionally, the Discussion could be strengthened by more clearly distinguishing between results supported directly by the data and broader contextual interpretations drawn from the literature. The section addressing traumatic experiences is informative but would benefit from greater analytic integration. Rather than listing frequencies of traumatic events, the authors could more explicitly discuss how specific categories of trauma (e.g., violence, displacement, economic loss) may differentially influence depression risk, even if explored descriptively. Similarly, while limitations related to timing, cross-sectional design, and lack of pre-crisis mental health data are acknowledged, these could be more succinctly framed and consolidated to avoid redundancy. Moreover, the Discussion would be strengthened by a clearer closing synthesis that emphasizes the study’s contribution to the limited mental health literature in Haiti, its implications for caregiver-focused interventions, and directions for future longitudinal and mixed-methods research. Finally, some of the contextual information described in the Discussion would be better placed in a Background section, and a smaller portion could be streamlined or referenced in the Methods. Thank you for this valuable research. This study provides important data on the impact of traumatic events and associated depression risks in Global South countries facing complex situations. I particularly appreciate the use of culturally validated tools, such as the Zanmi Lasante Depression Symptom Inventory and the adapted Life Events Checklist for DSM-5. After reviewing this first draft, I have the following questions and suggestions: - Since the Grandi Byen trial focused on child growth and development rather than adult trauma, could you confirm that the approved protocol included a mental health component? I don’t seem to find it in the Grandi Byen Protocol. Please also confirm that participant consent was obtained for the collection and use of mental health data as part of the Grandi Byen Trial as well as how confidentiality and ethical access to data was part of the process. - While the Grandi Byen Trial mentions 600 caregiver-infant dyads, this study includes 480 caregivers. Could you clarify how the inclusion criteria differed ? - For ethical clarity, what interventions or follow-up resources were provided to the 14.8% of participants identified as having a high risk of depression? - Given the focus on maternal mental health, the study would be more impactful if you highlighted data for female caregivers separately and provided the exact number of female participants. - Finally, I believe a qualitative evaluation of symptoms would have further deepened the findings. But it’s already included as limitation. Thank you again for your contribution to this field. Best regards, The reviewers have identified several shortcomings that need to be addressed. Please provide a response to each of the points raised. No accompanying comment. The authors were very receptive to the feedback and suggestions provided to improve the article’s quality. Great work detailing the research methodology. The models used and the statistical analysis appear impeccable. I am confident that this article provides essential mental health data to the Haitian scientific community that will effectively inform future strategies and programs. The authors have addressed my previous comments. I have one comment on the discussion section (first paragraph, line 51). The authors indicate that poverty-related indicators were associated with depression. In tables 2 and 3, there are no findings to support this statement. Please clarify. Thank you for your attention to the previous reviewer comments and for addressing several of the suggested revisions. This manuscript addresses a critical and underexplored public health issue by examining the relationship between trauma exposure and caregiver depression in Haiti, contributing valuable evidence to a limited body of mental health research in this context. The following comments are intended to further strengthen the methodological transparency, reporting of results, and overall organization of the manuscript. Methods Ethical Approval: Please include the Federal Wide Assurance (FWA) number associated with the Institutional Review Board (IRB) that approved this study. Providing this information is necessary to ensure compliance with ethical reporting standards. Caregiver Trauma Section: While a prior comment was partially addressed by stating that “small modifications were made based on the advice of local experts to be more relevant to northern Haiti,” the manuscript still does not sufficiently describe what those modifications entailed. For example, the trauma survey is described as being adapted from the LEC-5; however, the process by which items were selected, modified, or omitted is not explained. Please provide additional detail regarding how the instrument was adapted for this population and specify what “small modifications” were made to improve cultural relevance. Results: The Results section would benefit from substantial revision to ensure that analyses are clearly and consistently reported. Relevant statistical information should be explicitly incorporated for each analysis conducted. Strengthening the presentation of statistical results will improve clarity, transparency, and interpretability of the findings, as well as ensure alignment with journal reporting standards. For example, all reported percentages should be accompanied by the corresponding sample sizes. Additionally, the statement that “the mean depression score across the population was below the threshold in all cohorts” is presented without supporting statistical values. Please provide the relevant descriptive statistics (e.g., means and standard deviations) to support this assertion. Similarly, the statement that “the most common symptoms experienced among caregivers were ‘thinking too much,’ ‘feeling you have a constricted heart,’ ‘feeling tired or having little energy,’ and ‘low energy’” is not accompanied by any statistical data. Please report the frequency or proportion of participants endorsing these symptoms. The statement that “Cohort 1 showed experiencing a notably significantly lower estimated amount of traumatic events compared to other cohorts” also requires supporting statistical information. Please provide the appropriate comparative statistics to demonstrate differences between Cohort 1 and the other cohorts. Discussion: The primary focus of the Discussion section should be on interpretation of the study findings. Currently, this section includes six paragraphs, but only one directly addresses the results of the study. It would be beneficial to strengthen the Discussion by providing a more balanced interpretation of the findings in relation to the existing literature, rather than focusing predominantly on the broader context of sociopolitical unrest in Haiti. Additionally, some of the contextual information included in the Discussion may be more appropriately placed in the Background section. A portion of this material could also be streamlined or referenced in the Methods section where relevant. Please respond and address all of the reviewers' comments. No accompanying comment. Thank you for your attention to the previous comments and for addressing many of the comments. This manuscript addresses a critical and underexplored public health issue by examining the relationship between trauma exposure and caregiver depression in Haiti, contributing valuable evidence to a limited body of mental health research in this context. The following comments are intended to further strengthen the methodological transparency, and reporting of results of the manuscript. Methods: Caregiver Trauma Section: The following statement is not necessary and can be removed; this information is already well articulated in the methods section “The most significant changes were the 7 additional questions specific to the 2010 and 2021 earthquakes which caused significant disruption and hardship across the country” While a prior comment was partially addressed by stating that “small modifications were made based on the advice of local experts to be more relevant to northern Haiti,” the manuscript still does not sufficiently describe what those modifications entailed? Results: The Results section would benefit from minor revisions to ensure that analyses are clearly and consistently reported. It is not clear what is considered “Most” or “Many” – The following statements should be supported by the related statistics. Line 34 -Most caregivers were unmarried…………. Line 37 - Many caregivers worked outside the home, but most worked only a few days per week…. Line 38 - Most caregivers reported an income…. Line 44- Most do not have access to electricity…... Line 46 - Most homes had concrete floors…… Lines 35-43- There are no statistics reported to support the claims. Please address all points raised by the reviewer. No accompanying comment. Thank you for this valuable contribution to the field of global mental health. No accompanying comment.