Childhood adversity is becoming more prevalent for youth because of increasing exposure to natural and manmade disasters and the recent COVID-19 pandemic.Reference Rahmani, Muzwagi and Pumariega 1 Even before COVID-19, most youth reported at least 1 adverse childhood experience (ACE) by the age of 16.Reference Champine, Hoffman and Matlin 2 , 3 The developmental differences between adults and adolescents make adolescents more vulnerable to the mental and physical effects of disaster exposure.Reference Felix, Rubens and Hambrick 4 The most disenfranchised among the U.S. population are disproportionately harmed by disasters and national emergencies.Reference Fortuna, Tolou-Shams and Robles-Ramamurthy 5 Vulnerable youth and disadvantaged families carry compounding consequences that give rise to higher vulnerabilities, with fewer protective factors and resources.Reference Masten and Motti-Stefanidi 6
Disasters that cause catastrophic damage disrupt the operations of virtually all systems essential to human life and desirable for well-being, including family life, health care, work, education, economic and financial operations, transportation, manufacturing, emergency and other social services, recreation, and the functioning of the governments from local to national levels. Disasters can have both short- and long-term effects on youth’s health.Reference Tzaneva, Maeva, Erolova and Wei 7 Post-traumatic stress disorder (PTSD) and long-term morbidity are more prevalent among persons experiencing both trauma and social/economic inequities.Reference Fortuna, Tolou-Shams and Robles-Ramamurthy 5 Research on disasters experienced by youth has documented broad effects, including short- and long-term consequences for physical and mental health. For example, there is an increased risk for mental health problems (e.g., depression, anxiety, and post-traumatic stress), health risk behaviors (e.g., smoking, harmful alcohol use, illicit drug use), and other physical health problems.Reference Bellis, Hughes and Ford 8 –Reference Pumariega, Jo and Beck 10 Significant gender disparities have also been widely documented in disaster research.Reference Rahmani, Muzwagi and Pumariega 1 , Reference Felix, Rubens and Hambrick 4 For instance, girls are more likely than boys to report somatic, PTSD, and depression symptoms, and pre-existing gender disparities may become more prominent.Reference Felix, Kaniasty and You 11 –Reference Orengo-Aguayo, Stewart and de Arellano 13
As shown in Figure 1, youth in Puerto Rico have experienced a cascade of disasters starting in September 2017 with a catastrophic hurricane that devastated the island of Puerto Rico. Hurricane María collapsed the archipelago’s communication infrastructure, including 95% of cell networks; 14 collapsed the power grid, leaving all 3.5 million residents without power; 15 destroyed over 80% of agricultureReference Robles and Ferré-Sadurní 16 and over 300 thousand homes; 17 and caused an estimated 2975 deaths,Reference Lynch Baldwin and Begnaud 18 displacing over 200 thousand more.Reference Schachter and Bruce 19 The hurricanes exacerbated existing disparities.Reference Joshipura, Martínez-Lozano and Ríos-Jiménez 20 Two years later, a swarm of earthquakes starting December 2019 resulted in at least another 7.5 thousand individuals displaced. 21 Nearly 1 million citizens were left without power and more than 250,000 without water due to heavy structural damage to buildings,Reference Baratz 22 which led to the temporary closure of multiple schools.Reference Coto 23 A year after that, the COVID-19 pandemic began to affect Puerto Rico and the world. Youth again experienced multiple impacts with similar disruptions to school, social interactions, family life, and health care.
Figure 1. Timeline of Disaster Related Events in Puerto Rico. All icons were sourced from flaticon.com.
Although there is increasingly more focus on disaster-related research and trauma, consistent associations have not been found for youth and a lack of understanding on cultural differences in terms of resiliency underscores the need for further research.Reference Pumariega, Jo and Beck 10 This study analyzes data from a longitudinal study originally designed to estimate reliability of change for the AQOL-MHS using methods of Cranford et al.Reference Cranford, Shrout and Iida 24 It examines quality of life (QOL) with the backdrop of the emergence and persistence of psychological reactions in response to multiple compounding disasters. Associations between QOL and PTSD were evaluated, as well as prevalent mental health diagnoses. Specifically, differences in QOL reported by adolescents were quantified both overall and by domain (emotional regulation, self-concept, social context). Multivariable regression analyses were run with time as a linear trend to observe variables associated with change in QOL and explore the implications for disaster burdened youth.
Participants and Procedures
Two hundred and twenty-seven (227) participants with at least 1 of 5 prevalent mental disorders: attention deficit hyperactivity disorder (ADHD), conduct disorder (CD), oppositional defiant disorder (ODD), generalized anxiety disorder (GAD), and major depressive disorder (MDD), were recruited from 5 clinics in the San Juan Metropolitan area. These clinics included APS Healthcare in San Juan, APS Healthcare in Bayamón, APS Healthcare in Caguas, Dr. Antonio Ortiz Pediatric Hospital, and the Child and Adolescent Clinic at the Medical Sciences Campus, as well as some private practices. The inclusion criteria for the study were as follows: a) the youth was at least 12 years old but not older than 18; b) the youth had received 3 or fewer sessions with a mental health professional in the current clinical setting; and c) the youth had not received mental health treatment at other settings in the 6 months prior to recruitment. In addition, the study excluded youth with cognitive impairment, history of severe brain injury, pervasive developmental disorder, or evident sensory impairment.
After an initial visit to inform clinic staff of the study, study staff made routine visits to collect referrals with information on age, diagnosis, type of treatment/modality, severity, and prognosis. Parents/Caregivers were called to complete the screening process and validate the referral information. Once the participants were deemed eligible, the staff explained the research project, and those who expressed interest were scheduled for an initial interview. Parents/Caregiver and adolescent dyads signed a consent/assent form in which they acknowledged that their participation was voluntary and could withdraw from the study at any time. They also agreed to participate in the study over the course of a year. Baseline appointments were scheduled at the participants’ homes or alternate convenient locations. The follow-up waves were self-report interviews using a tablet and scheduled during routine clinical appointments. A private room was made available at participating clinics to provide a quiet space for participants to respond. More than 1 adolescent would be scheduled to join the private room to take part in the study. Due to pandemic restrictions, the final set of follow-up interviews was conducted remotely with the use of the Zoom platform. All participants were receiving services at the time of the baseline assessment and were tracked for follow-up appointments regardless of treatment status.
Data collection spanned over a 2-year period that included the prolonged recovery post hurricanes (2018-ongoing), a swarm of earthquakes (2019-2020), and part of the COVID-19 pandemic (March-June 2020), with 4 assessments 3 months apart for every participant. Participants received a research incentive to participate. The incentives were incremental for each study visit. For the initial interview, they were offered $30.00; for follow-up interview 1, $35.00; for follow-up interview 2, $40.00; and for follow-up interview 3, $45.00. The initial interview lasted 1.5 hours on average, and the follow-up interviews lasted 45 minutes on average.
Measures
The Adolescent Quality of Life-Mental Health Scale (AQOL-MHS)Reference Chavez, Mir and Canino 25 , Reference Chavez, Ramirez and Garcia 26, is a quality of life instrument for adolescents with mental health problems. This instrument differs from other QOL instruments developed for children or adolescents (e.g., the KIDSCREEN, 27 PedsQL,Reference Varni, Seid and Kurtin 28 YQOL-RReference Patrick, Edwards and Topolski 29) in that it is a specific instrument developed for clinical mental health populations with established psychometric properties. The AQOL-MHS has also been studied using generalizability theory methods to track changes in adolescent reports as their underlying health or mental health condition changed. The results support the utility and applicability of the measure to estimate change over time. The AQOL-MHSReference Chavez, Mir and Canino 25 , Reference Chavez, Ramirez and Garcia 26 , Reference Chavez, Shrout and Garcia 30 is a patient reported outcome measure with 21 items that can be scored as a whole or by domain for each of 3 QOL scales: emotional regulation (8 items), self-concept (6 items), and social context (7 items). Example of items: emotional regulation “When I have been angry, I haven’t been able to think straight,” self-concept “I have felt I would be able to achieve my goals,” and social context “I have enjoyed sharing and doing things with my family.” Participants evaluate whether the item description applies to them using an 11-point scale that ranges from 0= “completely disagree” to 10= “completely agree.” The responses to the emotional regulation scale were reverse coded so that larger numbers correspond to more positive QOL.
Post-traumatic stress disorder was assessed for both parents/caregivers and youth by the Post-Traumatic Stress Disorder Checklist (PCL-5), which has been considered a psychometrically sound measure by several studies and has been reliably used in different contexts and samples.Reference Blevins, Weathers and Davis 31 , Reference Roberts, Lotzin and Schäfer 32 The PTSD Checklist for Diagnostic and Statistical Manual of Mental Disorders (DSM)Reference Weathers, Litz and Keane 33 was developed to assess DSM-5 PTSD. It includes 20 self-report items. Participants reported how much they were bothered by a symptom over the past month using a 5-point Likert scale (0= “Not at all,” 1= “A little bit,” 2= “Moderately,” 3= “Quite a bit,” 4= “Extremely”). The total score can range from 0 to 80, and we used the developers’ suggested cutoff score of greater or equal to 33 as indicative of probable PTSD. Participants were asked to complete the PCL-5 using last month as the time frame and in relation to the traumatic experience that troubled them the most. Preliminary analyses concluded that the PCL-5 has satisfactory psychometric properties, as measured with internal consistency, test-retest reliability, and aspects of convergent validity and sensitivity to clinical change. Discriminant validity was not strong; however, this might reflect the overlap in symptoms between depression, general stress, and PTSD.Reference Forkus, Raudales and Rafiuddin 34 , Reference Sveen, Bondjers and Willebrand 35
Psychiatric diagnoses were measured by the Diagnostic Interview Schedule for Children (DISC-IV). The presence of a past year DSM-IV psychiatric disorder was assessed using the youth version of the DISC-IV.Reference Bravo, Ribera and Rubio-Stipec 36 , Reference Shaffer, Fisher and Lucas 37 The DISC-IV is a structured diagnostic instrument for the assessment of psychiatric disorders in children and adolescents (collectively youth) designed to be administered by lay interviewers. Previous versions of both the English and Spanish DISC have generally shown adequate test-retest reliability in both clinic and community samples.Reference Jensen, Roper and Fisher 38 –Reference Shaffer, Schwab-Stone and Fisher 40 Criterion, concurrent, discriminant, and predictive validity for earlier versions of the DISC have been evaluated and reported.Reference Jensen, Roper and Fisher 38 , Reference Goodman, Schwab-Stone and Lahey 41 –Reference Weathers, Litz and Keane 33 The test-retest reliability of the current version of the DISC has been reported for Spanish (using Puerto Rican youth) and English-speaking youth samples yielding comparable results.Reference Bravo, Ribera and Rubio-Stipec 36 , Reference Shaffer, Fisher and Lucas 37 To ease response burden, only the modules for the 5 target disorders (ADHD, CD, ODD, GAD, and MDD) were administered.
Disaster exposure questions were selected from the Hurricane Exposure Questionnaire developed by Felix et al.Reference Felix, Hernandez and Bravo 44 to assess the impact of the 1998 Hurricane Georges in Puerto Rico. The questionnaire has been used in other studies to explore the direct impact of disasters on Latino children and families.Reference Felix, You and Canino 45 –Reference Rubens, Felix and Hambrick 47 Questions cover a range of areas that include experiences, personal loss, infrastructure loss, financial impact, preparing for a new hurricane season, and positive consequences. Questions were developed for both caregivers and youth. Reported impact in Table 1 includes only parents’/caregivers’ reports. Higher scores indicate greater exposure. The sample was divided into 2 categories (no/limited impact vs major impact).
Table 1. Distribution by Hurricane María’s impact
* p < 0.05, ** p < 0.01, ***p < 0.001.
¹Psychiatric disorders were child reported except for ADHD.
Parents/caregivers completed questions about themselves (residence, country of origin, educational level, race/ethnicity, family structure, economic status based on perception of poverty, marital status, and household composition) and their youth (age, grade, and sex).
Statistical Analysis
Data on the demographic and clinical characteristics of the study sample were analyzed using descriptive statistics (means and percentages). To compare samples, independent t-tests (for continuous variables) and chi-squared tests of independence (for categorical variables) were performed. For mean differences on selected variables, effect sizes were calculated for continuous variables using Cohen’s d. Skewness and kurtosis were examined for the dependent variable and normal distribution assumptions were met. Independent sample t-tests were assessed for homogeneity of variances as used in the SAS t-test procedure. Variance Inflation Factors (VIF) were calculated to assess multicollinearity among the independent variables in the regression models. A VIF value greater than 1.5 was considered indicative of high multicollinearity. All regression variables were under this value. For linear regression, the homoscedasticity of residuals was evaluated using the Breusch-Pagan test, and the test did not indicate heteroscedasticity. In terms of multivariable outliers, data points were reviewed, and none were identified that could disproportionally affect the regression results. By ensuring these assumptions were met, we aimed to maintain the robustness and validity of our parametric tests and linear regression models.
To determine which variables were significantly associated with QOL during the study period, a multivariable regression analysis was performed with time (quarterly) as a linear trend. A linear fixed effects model was chosen for its ability to control for unobserved individual heterogeneity by leveraging within-individual variation over time. Fixed-effects models use the variation within individuals over time to estimate parameters. Tests were conducted to assess if QOL differed across subscales by introducing interaction terms between variables of interest. The following variables were included as independent variables: age, sex, having a diagnosis, having PTSD, and time. Four models were examined: Model A includes only main effects with the time-variant variable, Model B adds the time by diagnosis interaction term to the estimation model, Model C adds a time by PTSD (TP) interaction, and Model D adds a time by sex interaction.
The University of Puerto Rico, Medical Sciences Campus Institutional Review Board (IRB) for the protection of human subjects reviewed and approved this study (Protocol #2290033886).
Results
Sociodemographic Characteristics
The mean age of caregivers was 43 years old and 96% were female. Parents/caregivers represent 88% of mothers that responded, 4% of fathers, and 8% of “other” caregivers, such as grandparents or legal tutors. About half of the caregivers attained some college education or higher (50%), and 22% reported PTSD symptomatology. For adolescents, the mean age was 14 years old, 41% were female, 14% were classified as having PTSD, and 50% were scored as having a psychiatric disorder based on the DISC. The percentage of adolescents with an externalizing disorder was 39% and 22% with internalizing.
In Table 1, Hurricane María’s impact illustrates additional information about the study sample. Significant differences were observed for perception of poverty (parent/caregiver reported). Those caregivers who reported living poorly were the most affected by the hurricane. In separate analyses (data not shown), those parents/caregivers who reported income loss due to the hurricane also reported significantly greater impact due to the hurricane (P ≤ 0.001). Parents/caregivers with PTSD also reported greater impact due to the hurricane (P ≤ 0.01; Table 1).
QOL Mean Differences by Sex, PTSD, and Any Diagnosis
Mean differences and effect sizes for sex, any diagnosis, and PTSD for the 4 survey waves are depicted as line graphs in Figures 2-4.
Figure 2. AQOL-MHS Mean Comparisons for Youth by Sex.
*Means range from 0 to 10.
Sex
In general, girls fared worse for all 3 subscales of QOL in baseline through follow-up 2 interview reports, and improvement was finally seen on the third and final follow-up interview; emotional regulation and social context scores were no longer significantly different to boys (P ≤ 0.52, P ≤ 0.13, respectively). Even though self-concept showed the most improvement, the sex difference is still significant at the follow-up 3 interview (P ≤ 0.01). Boys reported similar scores across time for self-concept and social context throughout the study. Girls reported lower self-concept scores through follow-up 2 interviews and showed a marked improvement at survey wave 3 (Figure 2).
PTSD
Those with PTSD also scored lower than those without for all scales and waves of data, with the sole exception being baseline social context (P ≤ 0.22), which is possibly due to the post hurricane devastation experienced by all. For those with PTSD, social context had the greatest decline on the first follow-up of data collection but with progressive recovery. The emotional regulation scale had the lowest scores for those with PTSD, and these remained significantly lower than the non-PTSD group through the end of the study. Self-concept is the only scale that had follow-up wave 3 scores surpass baseline scores for non-PTSD participants. For those with PTSD, there was also noticeable improvement (Figure 3).
Figure 3. AQOL-MHS Mean Comparisons for Youth PTSD.
*Means range from 0 to 10.
Mental health diagnosis
Adolescents with a mental health diagnosis reported lower QOL for all scales. The largest difference between groups was reported for baseline self-concept scores (P ≤ 0.001). Although there was a parallel trend of improvement in both groups, the differences remained significant to the end. Scores reported for the emotional regulation scale are the lowest across the board. Those with a diagnosis (Dx) showed improvement in emotional regulation by follow-up wave 3, and, although scores remained lower, they were no longer statistically different to the non-diagnosis group (non-Dx; P ≤ 0.14). Social context groups did not differ for the first 3 assessments (ranging from P ≤ 0.06-0.12). Results are similar to baseline scores for this scale in our PTSD analyses; this might be due to the prolonged recovery efforts post hurricane. By the final survey follow-up, the Dx/non-Dx groups differed (P ≤ 0.05), but this was due to those with no diagnoses showing slightly greater improvement (Figure 4).
Figure 4. AQOL-MHS Mean Comparisons for Youth Diagnosis.
*Means range from 0 to 10.
Longitudinal Mixed Model Analyses
Tables 2-4 summarize the main findings for each scale from the multivariable mixed-effects regressions for the AQOL-MHS. Model 1 includes main effects only. The successive models (2-4) include interaction terms for the variables of interest. The longitudinal time trend estimate confirms that QOL worsened during the study period for emotional regulation but remained stable for the self and social scales in all 4 models.
Emotional regulation
Having a mental health diagnosis and PTSD were clear indicators of worse emotional QOL in all 4 models (see Table 2). A significantly positive time by PTSD interaction indicates that for those with PTSD there is greater improvement across the study periods. A similar trend was also observed which signaled improvement for those with a diagnosis. The time interaction for sex was not significant, but, nevertheless, the sex main effect for model 4 indicated that girls had worse emotional QOL (Table 2).
Table 2. AQOL-MHS mixed longitudinal analysis for emotional regulation
*p ≤ .05; **p ≤ .01; Ɨp ≤ .0001; ᵃp = .06–.10. All models use PCL at baseline.
AQOL-MHS, Adolescent Quality of Life-Mental Health Scale; Dx, diagnosis; PTSD, probability of post-traumatic stress disorder; SE, standard error.
Self-concept
Girls showed significantly worse self-concept compared to boys in all models (see Table 3). No change was observed in this pattern over time. Having a mental health diagnosis and PTSD were also indicators of worse self-concept in all 4 models. The only significant time interaction was with PTSD (P ≤ 0.002), which indicated improvement across time for those with the condition. No age group differences were observed. This contrasts with emotional and social context QOL, where a statistical signal is observed for all but 1 model (Model 4 in emotional regulation in Table 2).
Table 3. AQOL-MHS mixed longitudinal analysis for self-concept
**p ≤ .01; Ɨp ≤ .0001. All models use PCL at baseline.
AQOL-MHS, Adolescent Quality of Life-Mental Health Scale; Dx, diagnosis; PTSD, probability of post-traumatic stress disorder; SE, standard error.
Social context
Girls again showed significantly worse social context QOL in all 4 models (see Table 4). A statistical trend is observed for age (P ≤ 0.09) and having a mental health diagnosis (P ≤ 0.07). Younger adolescents fared better in their social environments as measured by the scale, while those with a diagnosis did worse. Having PTSD seems to be unrelated to social context. There were also no significant time interactions, which indicates no change throughout the study period (Table 4).
Table 4. AQOL-MHS mixed longitudinal analysis for social context
*p ≤ .05; **p ≤ .01; ᵃp = .06–.10. All models use PCL at baseline.
AQOL-MHS, Adolescent Quality of Life-Mental Health Scale; Dx, diagnosis; PTSD, probability of post-traumatic stress disorder; SE, standard error.
Discussion
Results are based on the responses of adolescents in Puerto Rico collected after surviving the 2017 Hurricane María as the starting point. The series of events that unraveled in their lives may be unique, but in many ways might also mimic the trauma experienced by other youth in disaster related circumstances. Youth of every race, ethnicity, and cultural background experience adversity, but youth from disadvantaged backgrounds and historically minoritized and traumatized groups are at heightened risk for both trauma exposure and developing PTSD.Reference Cohen, Deblinger and Mannarino 48 , Reference Jaycox, Stein and Kataoka 49 The sample came from economically disadvantaged families who had existing mental health problems and experienced adverse events. All these stress-ridden conditions can exacerbate existing disorders, result in a more severe course of mental health disorders, and lead to new stress-related conditions.Reference Fegert, Vitiello and Plener 50 A holistic understanding, which includes the environmental context, is central to addressing the needs of families and youth affected by trauma and adversity.
This study aimed to understand the QOL in adolescents by tapping into the AQOL-MHS’ 3 domains and drawing on the longitudinal data collected starting in 2018 to the beginning of summer 2020. There are 3 major findings. First, recovery takes time. This finding is reflected mostly by the emotional regulation QOL dimension. All emotional regulation models produced a negative time estimate, which indicated a decline. Second, there is a clear sex disparity for girls. This finding is consistent with results observed in numerous studies.Reference Luijten, van Muilekom and Teela 51 –Reference Zhou, Zhang and Wang 54 The self-concept scale highlights these differences most dramatically with effect size differences as high as 0.81 (second wave follow-up). This dimension of QOL explores self-esteem and a prospective outlook, which are both heavily influenced by circumstances. Third, the 3 quality of life domains addressed by the AQOL-MHS underline how the challenges experienced by disasters impact different areas of QOL in different ways and for different groups. An unexpected finding was that there were no significant differences for those with PTSD in the multivariable longitudinal regression analyses for social context. As much as the social context scale reflects the distressing post disaster realities experienced by all, the differences that were observed at the mean level of analyses were attenuated by a significant sex difference and by trends which indicate that the younger adolescents are less burdened. Also, those with a diagnosis conveyed worse social context QOL.
There is compelling evidence that, in youth, positive changes can produce a cascade of improved resilience capacity and improve health.Reference Bartlett 55 , Reference Masten and Cicchetti 56 Changes at home, school, and community are examples of how multiple aspects of resilience interact and affect the individual. Resilience research provides hope about the long-term outcomes of most youth and families in the wake of disasters.Reference Masten and Motti-Stefanidi 6 , Reference Masten, Cicchetti and Cicchetti 57 Study results were analyzed across time to explore for evidence of resilience. Observations revealed that even under repeated threats, improvement can be found for specific dimensions of QOL. Of special interest post disasters, children with PTSD reported better emotional regulation and self-concept dimensions of QOL across time. The challenges that they will have to face during their lifespans are numerous, and all efforts made to help them along the way will be beneficial. Gaining mental health ground in disaster related prevention, maintenance, recovery, and remission can make a great impact for current and future generations.
Limitations
This study has multiple strengths, which include a rigorous methodological and longitudinal design that measured the compounding impact of disaster events on youth and the use of a valid and reliable measure of QOL that was developed for Latino adolescents in Puerto Rico. Like most studies, however, there are also some limitations. First, the development of the instrument for a specific population could hamper its generalizability. Previous work has identified that the psychometric properties of the AQOL-MHS extend to community samples. An English language version was also used in the U.S., and a Castilian Spanish version has been used in Spain, and both have promising results.Reference Magallón-Neri, Chávez and Ortiz 58 Nonetheless, additional testing of its psychometric properties with Latino groups in the U.S. mainland and in other countries is recommended. Second, data collection ended during the first few months of the COVID-19 shelter-in-place orders. Several recent studies have found increases in mental health symptoms associated with the onset of the COVID-19 pandemic.Reference Hussong, Midgette and Thomas 59 Therefore, there is a need to consider the unique historic circumstances and possible consequences. Additional waves of data or additional new studies might demonstrate how these youth will fare in the future. Third, clinical youth may behave or react differently to adversity compared to non-clinical youth; therefore, results should also be framed within these considerations. Even so, mental health risks post pandemic should be considered for all youth, especially those who are disadvantaged and/or minoritized and have been disproportionally burdened. Finally, because the disasters occurred sequentially and given the study’s longitudinal design, the analyses cannot be disaggregated by type of disaster.
Conclusions
Adverse experiences during youth impact health outcomes and serve as mechanisms through which the detrimental effects of social determinants of health multiply across the lifespan.Reference Allen, Balfour and Bell 60 There is a need to continuously monitor and address the impacts of disasters on youth beyond the immediate impact phases and well into the long-term recovery periods.Reference Felix, Rubens and Hambrick 4 Supporting initiatives to prevent the onset and reduce the chronicity of existing mental health problems should be at the forefront of every health care system and policymaker’s list of priorities. As adversities continue to unfold and affect families, researchers and practitioners must consider how adversity-related demands placed on youth impact their mental health, behavioral risks, school performance, and how the consequences differ from those experienced by adults.Reference Roche, Huebner and Lambert 53 , 61 The mental health of young people must be supported as an upstream investment in long-term prevention because youth will experience the residual mental health effects of the disasters and the pandemic across their lifetimes.Reference McCray and Rosenberg 62 Screening individual children for adversity is a first step in preventing and mitigating its negative effects.Reference Bartlett 55 One essential way to support the mental health of youth is to provide universal screening in schools. Although this study recognizes that caregivers and pediatricians play an important role in recognizing mental health problems, we pin-point schools as the fitting place to conduct routine screenings. By building a health safety net into education, a lifeline can be created for young people who lack support in the home or are housing unstable.
Supplementary material
To view supplementary material for this article, please visit http://doi.org/10.1017/dmp.2024.175.
Acknowledgements
This research was supported by NIH Research Grants #HD060888 and #GM109326. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. We thank additional members of our study team, Karen Mir, PhD., Gisella Battisti, Marlys Massini, and the clinics that cooperated with our recruitment process: mental health outpatient clinics of the APS Healthcare System of Puerto Rico, the Dr. Antonio Ortiz Pediatric Hospital of the University of Puerto Rico, the Adolescent and Children Mental Health Clinic of Administration of Mental Health and Anti-Addiction Services (ASSMCA), and the invaluable contributions of the participants.
Author contribution
LMC designed study concept, analysis, collected and curated the data, and wrote the manuscript. PAG carried out data management, analysis, and interpretation. KV collaborated in the conceptualization of study design and writing. JS and AO participated in drafting the manuscript. JTSR edited, reviewed the paper, and prepared tables and figures.
Competing interest
All authors declare no financial or ethical conflicts of interest regarding the contents of the submission.
