Among the participants of this study, reporting traumatic events and female gender were associated with higher probability of mental health problems such as depression and anxiety. However, none of the subscales of the MOS scale was significantly associated with mental health problems after Bonferroni correction.
The participants in this study might have experienced a large number of traumatic events. The mean number of traumatic events was 5.7 across all surveyed students. A study from Belgium, also using SLE, revealed an average number of 3.6 traumatic events among adolescent migrants who arrived in the country (aged between 11 and 18 years) . Like many other studies that have indicated an association between traumatic events and poor mental health status , there was a strong association between the number of adverse life events and mental health problems among the participants of this study.
Unexpectedly, a statistically significant association was not detected between perceived social support and mental health problems. Further studies covering more details such as the gender and source of support  might elucidate the roles of social support among adolescent students in boarding houses. Other studies involving Asian migrants or students have reported positive associations between social support and mental health status [41–44]. The participants of this study, in contrast, stayed in boarding houses and were thus separated from their families. Although this source of social support was not investigated in the present study, the participants’ main source of social support might thus be from peers, whereas the availability of family support would be limited. Notably, previous studies have suggested the important role of family support .
Furthermore, we found a relationship between affectionate support and higher RATS score, although the association was not significant after Bonferroni correction. However, reciprocity [46–49] and the specific source of the support [50, 51] may be considered as additional factors affecting this issue. Understanding support and negative outcomes must take into account many dimensional pathways , not all of which could be addressed in the present study.
Overall, female participants reported higher levels of anxiety and depression than did their male counterparts. However, female gender alone is not adequate to explain the association with reporting poorer mental health status. Rather, the different factors associated with gender among the participants of this study would be expected to influence the results in this direction [52, 53]. The notion of gender is socially constructed, encompassing culturally dictated conventions, roles, behaviors, and identities . Gender and health status should thus be understood in the context of the manners in which people work, live, eat, and recreate . In addition, the unique gender roles in Asian culture might influence the results in the translated scales . Further, females might be more sensitive than men to their health status and more willing to report symptoms of distress . Certain findings might thus be attributed to such gender differences in reporting depression and anxiety, and to an inherent gender bias in the measurement construct used in this study . In other words, the effect of gender bias of the scales might partially explain the results observed [58, 59].
Participants’ age was not a significant predictor variable in the stepwise multivariate regression analysis but weakly associated with the depression subscale in the forced entry model. However, few consistent results have hitherto been shown regarding the influence of age on mental health status among those affected by traumatic experiences . Hence, further investigation is needed to elucidate this point.
The number of years spent in Thailand was a significant protective factor for mental health status. The whereabouts of parents and the number of years spent in boarding houses, on the other hand, were not. However, these factors should be analyzed in context and should account for further details such as the participant’s level of acculturation, adaptation, adjustment to their new country, and expansion of their social network within the community [61–63].
Among the participants of this study, the mean scores on the HSCL-37A and RATS scales were 63.1 and 41.4, respectively. Although it is difficult to draw comparisons due to the limited number of participants in this study, results from a mixed sample of 1,294 adolescent immigrants and refugees aged 11–18 years surveyed in a Belgian study indicated a mean total score of 56.9 for the HSCL-37A and 39.3 for the RATS [30, 35]. This suggests a better overall mental health status than among the participants of this study. Notably, when the Belgian research sample was limited to only the 477 participants who were migrants, the mean score was 56.9 for the HSCL and 38.5 for the RATS.
This study has four primary limitations. First, the cross-sectional design could not assess the causal relationship between social support and the mental health status of the participants. A qualitative and longitudinal study would better serve to understand the process and effect of social support on the mental health status of students in boarding houses.
Second, data collected in this study did not cover detailed information such as the nature and role of the person(s) providing support to the students and building relationships with them [55, 64]. The protective effect of social support on an individual’s well being is explained in two ways [65, 66]. One is through a “buffering”-effect model of social support, which protects individuals from the influence of life stress. The other is through a main-effect model, by which social support directly benefits a person’s health status . As information on the daily life stresses experienced by participants was not collected in this study, the buffering effect of social support could not be assessed.
Third, choosing the Burmese version of questionnaire was associated with greater frequency of mental health problems among students. Those participants who elected to fill out the Burmese version of the questionnaire also reported a poorer mental health status than did those who chose the Karen version. It is important to note that using a particular language would not be expected to associate directly with mental health status. The language itself is not a factor, but rather, as mentioned in the case of gender bias, it should be regarded as a proxy for different backgrounds and personal experiences.
Furthermore, those who chose the Burmese questionnaire included the Burmese adolescents and other minority groups such as the Shan and Mon. Further studies should be conducted considering potential difference across the ethnic communities . Context validity of the questionnaire between Burmese and Karen versions might also be a source of bias.
Finally, although this study invited all 771 students in grades 7 through 12, the initial number of participants was only 428. The authors presumed that a major reason for the reduced participation was that the data collection was conducted only on Saturdays and Sundays, and that participation was voluntary, as a result of which many students would have elected to preserve their free time rather than participate. Furthermore, due to possible emotional distress that was associated with some questions, it was emphasized that participants could skip questions that they did not wish to answer. The questionnaire also contained a large number of items for the students to complete; thus some participants may not have been able to maintain consistent concentration throughout and could thus have overlooked items in the questionnaire. As a result, only 304 participants completed the questionnaires.
Despite these limitations, this study provides a valuable overview and insight into the mental health status of students in the boarding houses of Burmese migrant schools in Thailand. Mixed movement of refugees and migrants is a global phenomenon . The refugee and migrant populations use the same route to the destination country and share similar risks such as human trafficking . It is necessary to observe the nature and characteristics of the mobile population, rather than making generalized assumptions about illegal migration .
Psychiatric information about adolescents in low- and middle-income countries is generally sparse, but a troubling picture of depression and high suicide rates has been highlighted . More comprehensive improvement and expansion of social services offered is necessary such as upgrading of mental health assessment tools, treatment in primary care, availability of medication, national mental health programs, and training of mental health care professionals . At the same time, adolescents tend to underutilize mental health services due to stigma and other priorities in life . Further studies are necessary to make mental health promotion more successful in low- and middle-income countries, particularly within such vulnerable populations [68, 69].