Authors: Lindsay Stark, Arturo Harker Roa, Carolina Rodriguez, Elvia Tamaity Ariza Pena, Julianne Deitch, Ilana Seff
Categories: Research, Adolescent girls, Forced migration, Mental health, Depressive symptoms, Self-esteem
Source: BMC Public Health
Authors: Lindsay Stark, Arturo Harker Roa, Carolina Rodriguez, Elvia Tamaity Ariza Pena, Julianne Deitch, Ilana Seff
Venezuelan migrants in South America are at heightened risk for mental distress due to factors such as discrimination, unstable employment, and gender-based violence (GBV). Adolescent girls are an especially vulnerable group, with family dynamics playing a crucial role in this population’s mental health status. This study explores how gendered household composition is associated with the mental health and well-being of forcibly displaced adolescent girls in Colombia.
This cross-sectional analysis employs baseline data from a pilot randomized controlled trial evaluating “Sibling Support for Adolescent Girls in Emergencies” **(**SSAGE), a whole-family, gender-transformative program aimed at improving adolescent girls’ mental health and well-being. Data were collected from 186 adolescent girls. Key outcomes of interest included the Revised Children’s Anxiety and Depression Scale (RCADS-25) and the Rosenberg Self-Esteem Scale (RSES); predictors of interest comprised the presence of various individuals residing in the household and a measure of the number of household tasks performed by girls in the last two weeks compared to an adolescent male in the household. Linear regression models were used to estimate the relationships between predictors and outcomes for younger and older adolescent girls, separately.
Living with a male relative other than a father, brother, or grandfather was associated with higher anxiety and depressive symptoms and lower self-esteem for 13-16-year-old girls, as compared to not living with such a male relative. Adolescent girls ages 13–16 years old who reported performing a greater number of household tasks relative to their adolescent male counterpart were found to have lower self-esteem scores. For 17-19-year-old girls, living with a mother and sister was associated with lower anxiety and depressive symptoms and living with a mother was linked with greater self-esteem as compared to those not living with those family members. Additionally, 13-16-year-old girls who reported performing a greater number of household tasks relative to their adolescent male counterpart were found to have lower self-esteem scores.
Findings underscore the significance of household gender composition in relation to the mental health and self-esteem of forcibly displaced adolescent girls. Findings highlight the potential protection of a gender-transformative approach within family-centered interventions to improve forcibly displaced girls’ well-being.
Colombia hosts approximately 2.8 million refugees, making it one of the top destinations for displaced individuals [1]. The majority of migrants in Colombia are Venezuelan, driven by Venezuela’s prolonged economic and social collapse, which has led to instability, hyperinflation, rising crime, and food and medical shortages [2, 3]. These migrants face extreme vulnerability, including limited access to social protection systems, labor exploitation, targeted violence, and further displacement within Colombia [4, 5, 6]. In addition to those settling in Colombia, increasing numbers of Venezuelans engage in regular cross-border movement, known as “pendular migration” [2, 7, 8].
Forced displacement, compounded by conflict, loss of social networks, and economic instability is a well-documented risk factor for mental distress [9]. Refugees and conflict-affected populations experience high rates of depression, PTSD, and anxiety, with one meta-analysis estimating the prevalence at 32%, 31%, and 11%, respectively [10]. Among Venezuelan migrants in South America, factors such as discrimination, unstable employment, and gender-based violence (GBV) have shown to further increase mental health risks [11]. Conflict and displacement often intensify existing gender inequities by disrupting traditional protection mechanisms, and increasing women and girls’ exposure to economic dependency, sexual exploitation, and interpersonal violence [12]. When families experience such stressors as poverty, loss of housing, or lack of livelihoods, these conditions can elevate tensions at home and heighten the risk of GBV, especially in contexts where patriarchal norms are unchallenged [13, 14]. Adolescent girls are particularly vulnerable, exhibiting higher rates of anxiety and depression compared to boys [15, 16, 17, 18]. A study on adolescents impacted by the Venezuelan-Colombian crisis found that parental abandonment, violence, and lack of support structures significantly harmed the mental health of adolescent refugee girls, while also limiting their access to psychosocial health services [4, 19, 20, 21, 22].
From a socio-ecological perspective, family dynamics play a crucial role in adolescent mental health. Parental stress, harsh discipline, and domestic violence are all known intermediaries between displacement and mental distress in young people [23, 24]. Additionally, gendered household dynamics, such as gender inequitable roles and expectations, have been shown to negatively impact adolescent girls’ self-esteem and psychological well-being [25, 26, 27, 28, 29]. Forcibly displaced households may adopt rigid gender norms as a coping strategy, further limiting girls’ mobility, educational opportunities, and participation in decision-making– factors that compound both mental health and vulnerability to abuse. Conversely, caregiver mental health, family cohesion, and peer support have been shown to foster resilience [30, 31, 32, 33]. Studies in Gaza, Liberia, and Sri Lanka, highlight the importance of supportive family relationships in adolescent wellbeing [30, 34]. Families that foster secure attachments, gender equity, and positive communication can help protect adolescent girls from violence and mental distress [30, 31, 32, 35].
Despite the well-documented influence of family on adolescent mental health, most interventions have focused on children and adolescents directly, with broader family-based approaches only recently gaining attention [24]. Understanding how household members may influence adolescent girls’ mental health and well-being is critical to improving whole-family programming. For instance, addressing sibling dynamics, in addition to exploring gendered relationships between adolescent girls and caregivers, has been shown to lead to both short-term and long-term benefits for adolescent girls’ mental health and protection from violence [36].
This study explores how gendered household composition is associated with the mental health and well-being of forcibly displaced adolescent girls. Employing cross-sectional survey data with displaced families in Colombia, we examine how living with different male and female household members and the division of labor between adolescent girls and boys relate to adolescent girls’ anxiety, depression, and self-esteem. We hypothesize that co-residence with men will be associated with higher levels of depression and anxiety and lower self-esteem among adolescent girls, while co-residence with women will correspond to lower levels of depression and anxiety and higher self-esteem. Our hypotheses regarding which specific male and female household members drive these associations are exploratory.
The study was conducted in Cartagena, a city in Colombia’s Bolívar department. Cartagena is a major destination for forcibly displaced Venezuelans due to its proximity to the Venezuelan-Colombian border, which facilitates transit between the two countries [37].
The present analysis employs cross-sectional baseline data from a pilot randomized controlled trial (RCT) evaluating “Sibling Support for Adolescent Girls in Emergencies” **(**SSAGE), a whole-family, gender-transformative program aimed at improving adolescent girls’ mental health and well-being [38]. Participants included forcibly displaced families living in Cartagena, with baseline data collected from adolescent girls within these households. The study also included Colombian returnee families who had previously lived in Venezuela for at least five years before returning to Colombia. Mercy Corps Colombia, an NGO supporting refugees and displaced populations since 2006, led the study recruitment process. Potential participants were identified using a database of households who had previously received or been eligible for programming from Mercy Corps. Households were contacted by a Mercy Corps staff member to confirm eligibility for the present study. To be eligible, adolescent girls had to be between 13 and 19 years old and be willing to participate in the three-month SSAGE intervention. They also needed to be available for a follow-up survey questionnaire within one month following the completion of the intervention. A total of 186 adolescent girls from 186 families were recruited into the study. Additional details on the sample size calculations, which were informed by the objective to detect a medium effect size in outcomes of interest in the pilot RCT, have been published elsewhere [38].
Prior to randomization and the start of the intervention, all 186 participants completed a baseline survey in August 2024. Informed consent was obtained prior to survey administration. Research staff explained the study’s purpose, procedures, potential benefits and risks, and the referral process. The referral process was initiated if a participant indicated risk of self-harm, as measured through two questions in the DSM-V cross-cutting youth, which was part of the survey questionnaire. In the event of such a disclosure, the participant was referred to Mercy Corps, who employed their standard referral pathway for the necessary support. Additionally, for any respondent exhibiting distress as a result of a disclosure, real-time, remote support was provided by a social worker trained in psychological first aid. Both caregivers’ written consent and participants’ written assent were obtained for girls under 18.
Baseline surveys were conducted through enumerator-facilitated interviews using computer-assisted personal interview (CAPI) software; surveys lasted approximately 25–45 minutes, on average. Participants received a small transportation stipend to cover the costs associated with their travel to the data collection site. All data collectors were trained in the survey measures, research ethics, and the referral protocol. Only female data collectors were used to maximize participant comfort, and all interviews took place in a private space to ensure confidentiality. Survey tools were translated into Spanish and back translated for quality assurance, and all surveys were administered in Spanish. Ethics approval was obtained from the Institutional Review Boards of Universidad de Los Andes and Washington University in St. Louis.
The present analysis examines the following key outcomes of a combined measure of depressive and anxiety symptoms as measured through the Revised Children’s Anxiety and Depression Scale (RCADS-25) and self-esteem as measured through the Rosenberg Self-Esteem Scale (RSES).
The RCADS-25 is a 25-item scale measuring levels of depression and anxiety in children [39]. For each item, the respondent is asked to indicate the frequency of symptoms in the previous two weeks using a 4-point Likert scale, where 0 = Never, 1 = Sometimes, 2 = Frequently, and 3 = Always. Examples of items measuring depressive symptoms include, “I am tired a lot,” “I feel sad or empty,” and “nothing is much fun anymore.” Anxiety items include, “I worry when I have done poorly at something,” “I worry what other people think of me,” and “I feel restless.” The final score is a summed total of all items (0–75), where higher scores signal greater depression and anxiety symptoms. A Spanish version of this measure has been previously validated with Latin American youth [40].
The RSES is a 10-item scale that measures self-esteem and has been widely used in several countries and languages for nearly 60 years [41, 42]. Respondents are asked to indicate their level of agreement with each statement using a 4-point Likert scale, where 1 = Strongly agree, 2 = Agree, 3 = Disagree, and 4 = Strongly disagree. Items include, “on the whole, I am satisfied with myself,” “I feel that I have a number of good qualities,” and “I certainly feel useless at times.” After reverse-coding five of the ten items, the final score [10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40] is taken to be the sum of all items, where higher scores represent greater self-esteem.
In order to explore how cohabitating with different household members may relate to adolescent girls’ outcomes, presumably through consequential intrahousehold dynamics, the analysis examines how the outcomes of interest vary across girls’ cohabitation with different household members. To construct the main explanatory variables (i.e. presence of each household member type), respondents were asked to select the individuals currently living in the household from a list of the mother, father, sister, brother, grandparent, another male relative, other female relative, and other non-relative. One dichotomous variable was constructed for each household member. For example, the variable for living with one’s mother was coded as a ‘1’ if the respondent reported living with her mother and a ‘0’ otherwise.
Previous research has shown that a more gender equitable division of household responsibilities is associated with greater mental well-being for women [43]. As such, the final predictor of interest captures information on the number of household responsibilities the respondent had performed in the previous two weeks, in comparison to the number of household responsibilities performed by an adolescent male in her household. Respondents were asked to indicate the frequency with which they had performed each of the following nine tasks in the last two cleaning the house (mop, sweep), cleaning the bathrooms, making the beds, organizing the rooms, washing the dishes, cooking and/or serving meals, washing clothes, taking care of younger siblings, cousins, nephews or nieces, and taking younger siblings, cousins, nephews or nieces to school. Adolescent girls were also asked to answer the same set of questions for the adolescent boy in their household that would be participating in the SSAGE intervention. A summed score was created for the adolescent girl and her male adolescent counterpart indicating the number of activities each individual had performed at least once in the previous two weeks. The final indicator included in the analysis was operationalized as the total number of activities performed by the adolescent girl minus the total number of activities performed by the adolescent boy. As such, a value of ‘0’ signals that household responsibilities are gender equitable between the adolescent boy and girl and a value of ‘10’ indicates a girl performed ten more activities than their male counterpart in the last two weeks.
Descriptive statistics for all variables of interest were summarized for the full sample, and for girls ages 13–16 years old and girls ages 17–19 years old. Differences between the two age cohorts were tested using paired T-tests. Bivariate linear regressions were used to estimate the relationship between each predictor and the two outcomes of interest. Fully adjusted models controlling for all predictors of interest and age simultaneously, were then estimated. All models were estimated separately for each age cohort for two reasons. First, the literature shows that mental disorders and self-esteem may be patterned differently for younger vs. older adolescent girls [44, 45]. Second, adolescent girls participating in SSAGE are stratified by age group in order to accommodate certain program content that is appropriate for some ages and not for others. As a result, for both theoretical and practical reasons, we opted to stratify the analysis by age group in order to maximize both comparability with existing literature and utility in program implementation. All assumptions of linear regression models held and no data were missing. Analyses were conducted using Stata16.
Descriptive statistics for the sample are presented in Table 1. Approximately 54% of the sample were between the ages of 13 and 16 years old, while 46% of the sample were between 17 and 19. Nearly 85% of the sample reported living with their mother, 42% with their father, 73% with a brother, and 41% with a sister. Additionally, 19% reported living with a grandparent, 51% with another male relative (who is not a father, brother, or grandfather), 26% with another female relative (who is not a mother, sister, or grandmother), and 6% with a non-relative household member. No differences between these household compositions were observed between age group.
Table 1Descriptive statisticsFull sample (n = 186)13-16-year-olds (n = 101)17-19-year-olds (n = 85)P-value (T-test of difference)Age15.83 [1.87]14.37 [1.15]17.56 [0.73]Currently in school153 (82.26%)89 (8.12%)64 (75.29%)0.023Living Mother157 (84.4%)88 (87.1%)69 (81.2%)0.267 Father78 (41.9%)46 (45.5%)32 (37.6%)0.279 Brother135 (72.6%)71 (70.3%)64 (75.3%)0.449 Sister77 (41.4%)43 (42.6%)34 (40.0%)0.724 Grandparent36 (19.4%)22 (21.8%)14 (16.5%)0.364 Other male relative95 (51.1%)49 (48.5%)46 (54.1%)0.449 Other female relative48 (25.8%)24 (23.8%)24 (28.2%)0.490 Other non-relative11 (5.9%)5 (5.0%)6 (7.1%)0.546Gender difference in # of HH tasks (girls - boys)3.11 [2.86]2.56 [2.95]3.76 [2.61]0.004Revised Child and Anxiety Depression Scale (RCADS-25)23.38 [12.04]21.51 [11.89]25.59 [11.90]0.021*Rosenberg Self-Esteem Score (RSES)19.21 [4.19]19.21 [3.81]19.21 [4.61]0.995Note: Values are n(%) or mean[SD]. Differences between age groups are estimated using T-tests and are statistically significant at *p < 0.05, **p < 0.01, ***p < 00.01
On average, respondents performed an additional 3.11 household tasks in the previous two weeks as compared to their adolescent male counterpart and this gender gap was larger for 17-19-year-olds (3.76 greater tasks) as compared to 13-16-year-olds (2.56 greater tasks; P = 0.004). The average RCADS25 and RSES scores were found to be 23.38 and 19.21, respectively. The RCADS-25 score was greater in the 17-19-year-old group as compared to the 13-16-year-old group (25.59 v. 21.51; P = 0.021).
Results from the regression analyses predicting the RCADS-25 scores are found in Table 2. Overall, living with a male relative other than a father, brother, or grandfather was associated with higher anxiety and depressive symptoms among the younger cohort, as compared to not living with such a male relative. For the older cohort, living with a mother and sister was associated with lower anxiety and depressive symptoms as compared to those not living with a mother or sister. In the fully adjusted model, 13-16-year-old girls who reported living with a male relative other than a father, brother, or grandfather exhibited 5.59 higher (95% CI [0.50,10.67]; P = 0.032) RCADS-25 scores as compared to those girls who did not report living with another male relative. In other words, living with a male relative other than a father, brother, or grandfather was associated with a 0.47 increased standard deviation (the standard deviation for the RCADS-25 score among 13-16-year-olds was 11.89) in the RCADS-25 score for the younger cohort. In contrast, 17-19-year-old girls who reported living with their mother or sister had RCADS-25 scores 8.88 points lower (95% CI [-16.31,-1.45]; P = 0.020) or 7.16 points lower (95% CI [-12.31, -2.02]; P = 0.007), respectively, as compared to those who did not report living with their mother or sister; the effect sizes correspond to RCADS-25 scores that were 0.75 and 0.60 standard deviations lower, respectively.
Table 2Estimating the Revised Child and Anxiety Depression Scale (RCADS-25)13–16-year-olds17–19-year-oldsUnadjustedAdjustedUnadjustedAdjustedB [95% CI]B [95% CI]B [95% CI]B [95% CI]Living Mother0.152.3-7.13*-8.88*[-6.90,7.20][-5.78,10.39][-13.55,0.71][-16.31,-1.45] Father-1.54-0.192.974.5[-6.27,3.18][-5.13,4.76][-2.32,8.26][-1.44,10.43] Brother-2.350.27-3.39-1.56[-7.49,2.79][-5.66,6.20][-9.33,2.55][-8.38,5.26] Sister-1.42-1.63-6.67*-7.16**[-6.19,3.34][-6.53,3.26][-11.73,1.60][-12.31,-2.02] Grandparent-1.82-2.080.920.72[-7.53,3.89][-8.03,3.87][-6.04,7.88][-6.53,7.98] Other male relative4.635.592.032.05[0.00,9.26][0.50,10.67][-3.13,7.20][-4.49,8.59] Other female relative0.20.11-1.11-5.06[-5.35,5.74][-6.05,6.26][-6.84,4.62][-11.13,1.02] Other non-relative member5.565.1-6.01-8.18[-5.26,16.38][-5.76,15.96][-16.01,3.98][-19.18,2.83]Gender difference in # of HH tasks (girls - boys)0.800.80.410.04[0.01,1.59][-0.05,1.66][-0.58,1.40][-0.96,1.04]Note: Fully adjusted models also control for age. Beta coefficients are statistically significant at *p < 0.05, **p < 0.01, and ***p < 0.001
While performing more tasks as compared to the adolescent male in the household was associated with greater anxiety and depression for the younger cohort (B = 0.80; 95% CI [0.01,1.59]; P = 0.046) in the bivariate regression, this association was only significant in the unadjusted model.
Similar patterns were observed when estimating RSES scores (see Table 3). Overall, living with a male relative other than a father, brother, or grandfather, and living with a non-relative household member were both associated with lower self-esteem for the younger cohort and living with a mother was associated with greater self-esteem for the older cohort. In the fully adjusted model, 13-16-year-old girls living with a male relative other than a father, brother, or grandfather exhibited 1.67 lower (95% CI [-3.22,-0.12]; P = 0.035) RSES scores (i.e., RSES scores that were 0.44 standard deviations lower) as compared to those girls who did not report living with another male relative. Girls living with a non-family household member scored 3.51 points lower (95% CI [-6.82,-0.20]; P = 0.038) on the RSES (i.e., RSES scores that were 0.92 standard deviations lower) compared to girls who did not. Among girls ages 17–19 years old, those who reported living with their mother had RSES scores that were 4.60 points greater (95% CI [1.52,7.68]; P = 0.004) (i.e., RSES scores that were one standard deviation lower) as compared to those who did not live with their mother.
Table 3Estimating the Rosenberg Self-Esteem Scale13–16 year olds17–19 year oldsUnadjustedAdjustedUnadjustedAdjustedB [95% CI]B [95% CI]B [95% CI]B [95% CI]Living Mother1.30.094.194.60[0.94,3.54][-2.37,2.55][1.79,6.58][1.52,7.68]Father0.3-0.13-0.34-1.16[1.22,1.81][-1.64,1.38][-2.41,1.73][-3.62,1.30]Brother0.49-1.190.98-0.99[1.16,2.14][-2.99,0.62][-1.34,3.29][-3.81,1.84]Sister-0.65-0.891.170.5[2.17,0.88][-2.38,0.60][-0.86,3.20][-1.63,2.63]Grandparent-0.91-0.31-1.28-0.54[2.73,0.92][-2.12,1.51][-3.97,1.41][-3.54,2.47]Other male relative-1.43-1.67-1.12-1.01[2.92,0.05][-3.22,-0.12][-3.12,0.87][-3.72,1.70]Other female relative-2.08*-1.58-1.110.52[-3.80,0.35][-3.45,0.30][-3.32,1.11][-1.99,3.04]Other non-relative member-4.43*-3.51*-1.480.19[-7.79,1.06][-6.82,-0.20][-5.38,2.42][-4.36,4.75]Gender difference in # of HH tasks (girls - boys)-0.35**-0.38**-0.08-0.07[-0.60,0.10][-0.64,-0.12][-0.46,0.31][-0.48,0.34]Note: Fully adjusted models also control for age. Beta coefficients are statistically significant at *p < 0.05, **p < 0.01, and ***p < 0.001
Similar to the pattern observed for anxiety and depression, 13-16-year-old girls who reported performing a greater number of household tasks relative to their adolescent male counterpart were found to have lower self-esteem scores (B=-0.38; 95% CI [-0.64,-0.12]; P = 0.004). Each additional task a 13-16-year-old girl performed relative to her adolescent male counterpart was associated with a 0.10 decreased standard deviation in the RSES, reflecting a small effect size. This relationship was not observed for the older cohort.
The analysis examines relationships between gendered household composition, adolescent boys’ and girls’ participation in household responsibilities, and symptoms of anxiety, depression, and self-esteem across two cohorts of adolescent a younger cohort, consisting of girls aged 13–16 and an older cohort, comprising girls aged 17–19. For the older cohort, living with a female family member, such as a mother or sister, was found to be associated with lower anxiety and depressive symptoms and higher levels of self-esteem. For the younger cohort, living with a male relative other than a father, brother or grandfather, was associated with greater anxiety and depressive symptoms and lower levels of self-esteem. Furthermore, while both younger and older girls performed significantly more household responsibilities than their adolescent male counterparts, this inequity was further linked to lower self-esteem in the younger cohort.
Our study highlights the potentially protective roles female family members may play with respect to both mental health and self-esteem among adolescent girls in humanitarian contexts. Although there is currently limited evidence exploring the content and quality of relationships between adolescent girls and other girls and women in their household in humanitarian contexts, previous literature within marginalized or high-risk communities finds that having female role models, maternal figures, and social support can provide adolescent girls with examples of healthy coping and increases overall psychological well-being [46, 47, 48]. In under-resourced and humanitarian settings, informal female networks outside the household have been shown to positively influence health and access to resources [49, 50, 51]. More formalized women’s collectives and adolescent girl groups have also been shown to improve agency, empower women and girls, and buffer against current or future violence [38, 47, 49]. Our findings suggest that female relationships within the household may similarly bolster resources for adolescents in settings of forced displacement, pointing to the importance of strengthening these bonds in whole-family programing. Further, programming should consider connecting adolescent girls who do not live with a sister or mother—which is not uncommon in forced displacement—with other female mentors outside the home.
Relatedly, our findings point to the potential risk that living with male family members outside the nuclear family can present for forcibly displaced adolescent girls’ mental health and self-esteem. Buehler and Gerard [52] describe the family as a key socialization context for children as they move through early adolescence. Gender inequitable norms and socialization within the household that impact adolescent girls—such as limited mobility, reduced involvement in family decision-making, and greater expectations to complete household tasks—have well-documented effects on their mental health and well-being [53, 54, 55, 56]. Indeed, the present analysis revealed an association between the relative number of tasks adolescent girls complete in comparison to adolescent boys and adolescent girls’ self-esteem. Further, while the relative pervasiveness of inequitable gender norms among forcibly displaced populations in Colombia has been documented, the impact of this inequity may be more pronounced when more men are living in the household [57]. The increased risk of compromised well-being found for younger adolescent girls living with male relatives in our study may also be related to safety concerns [58, 59].
Gender inequality in the family can perpetuate violence against women and girls at the family level, with humanitarian emergencies exacerbating instances of violence [60]. Research examining the interrelated drivers of household violence against women and children in displaced Colombian communities has highlighted how relocation and ongoing insecurity exacerbate family violence, with family composition in continual flux and members facing economic precarity and increased substance use [13]. Indeed, conflict, migration, and economic hardship have been shown to add strain to the family, intensifying interfamilial conflict and increasing risk of GBV in contexts around the globe [14]. The recent displacement of families within our study may have similarly intensified gender-based conflict for families with “other” males in the household, resulting in increased levels of anxiety and depression for adolescent girls.
While further research is needed to understand why the presence of non-nuclear male family members may negatively impact the well-being of younger adolescent girls, in particular, previous research points to a few possible pathways of influence. First, younger adolescent girls may be less likely than their older counterparts to have developed healthy and effective coping mechanisms to deal with the stress of complex household dynamics. Previous studies around use of coping strategies from a life course perspective found that developmental age has an impact on use of different coping techniques, with more adaptive, problem solving-oriented coping strategies observed in adolescents 16 and older [61, 62]. Second, younger adolescent girls may face greater supervision from older family members. In settings of conflict or displacement, especially, increased levels of stress and perceptions of insecurity can lead male family members to limit younger girls’ mobility [32, 36, 63]. Future research should seek to further unpack interpersonal dynamics within multigenerational or extended households, to better understand the impact of living with non-nuclear family male relatives on forcibly displaced adolescent girls’ well-being. These dynamics are especially important for displaced families in non-Western contexts, given the increased prevalence of larger, multi-generational households in countries across Latin America, Asia, and Africa [64].
Overall, findings stress the importance of gendered household composition for adolescent girls’ mental health and self-esteem, highlighting the criticality of a gender-transformative approach within family-centered interventions to improve forcibly displaced girls’ well-being. Gender-transformative programs aim to not only shift behavior and raise awareness around gender equity but to also transform unequal power dynamics around gender [65]. Critically, engaging men and boys in these types of discussions and programs is paramount to inducing change [66]. The past decade has seen increases in programs that aim to improve the health and wellbeing of adolescents through family and caregiver support. However, evidence suggests that family interventions that fail to engage male family members are limited in their potential to impact adolescent girls’ experience of violence and mental health [29]. Nonetheless, a recent review of whole-family programs targeting violence against adolescent girls in humanitarian settings identified only two interventions that were also gender-transformative [67]. SSAGE, one of these interventions, has been piloted with forcibly displaced families in Nigeria and Jordan and has shown potential to improve girls’ safety networks, supportive parenting behaviors, and attitudes around gender norms [68]. More robust evaluations of gender-transformative whole-family interventions are needed to assess whether these impacts extend to positive changes for adolescent girls’ mental health and well-being. Future research should also seek to better understand the mechanisms through which the engagement of different family members, including men and boys, can impact girls’ well-being.
Findings from this study should be considered alongside a few limitations. First, the cross-sectional nature of our data limits our ability to make causal claims, especially given the potential for unmeasured confounding variables [69]. Although we find significant associations between gendered household composition and adolescent girls’ mental health and self-esteem outcomes, we cannot establish the causality of this relationship. Additionally, the single time point measurement cannot capture how these relationships may evolve over time or how developmental transitions during adolescence might influence these associations. Second, it is important to acknowledge that the household task measurement relies on adolescent girls reporting on both their own and their male counterparts’ behaviors, which could be subject to recall bias. Third, although our analysis offers important insights into adolescent girls’ mental health outcomes, the use of purely quantitative assessment cannot fully capture the complexity and nuance of family dynamics among forcibly displaced households. Future research would benefit from incorporating qualitative or mixed methods approaches to better capture the lived experiences of adolescent girls, allowing for a deeper understanding of how family dynamics and displacement intersect to influence mental health and well-being. Finally, our study focuses exclusively on forcibly displaced adolescent girls living in Cartagena, Colombia. Although gender-based violence and gender inequity are global issues, the manifestation and impact of gender norms are contextual. As such, our findings may have limited generalizability to other contexts.
Migration, along with exposure to conflict, loss of family members, and other life-threatening experiences, are recognized predisposing factors for mental distress in displaced youth, with adolescent girls facing additional gendered exposures. Our study underscores the significance of household gender composition in shaping the mental health and self-esteem of forcibly displaced adolescent girls. The family unit has the potential to prevent mental illness and promote healthy development, reinforcing the need for gender-transformative, family-centered interventions. Addressing gender inequities within the household and engaging female and male members of the household in supportive caregiving and decision-making structures may be a critical step in fostering adolescent girls’ mental health, resilience, and wellbeing.