The results presented describe systematic human rights violations, traumatic events, daily stressors, mental health symptoms, and functional impairment, and examine the associations between these factors.
Demographics
Of the total households selected for inclusion, 168 (34%) were either not home or did not have an eligible respondent to complete the survey (often due to being a minor headed household, or because eligible respondents were not at home). In addition, 13 eligible respondents declined to participate in the survey. The final sample of 495 participants included 264 women (53%) and 231 men (47%), which closely matches the camp population gender breakdown of 56% women and 44% men. For more demographic data, see Additional File 1.
Systematic Human Rights Violations
Response options for systematic human rights violations were, 1 = “Not at all”, 2 = “A little”, 3 = “Quite a bit”, and 4 = “Extremely.” Respondents endorsed restrictions against the Rohingya at the following rates: “blocked from . . .” obtaining citizenship (M = 3.99; 100% endorsed quite a bit or extremely), working in government positions (M = 3.99; 100% endorsed quite a bit or extremely), obtaining official documentation (M = 3.99; 99.8% endorsed quite a bit or extremely), using the name Rohingya (M = 3.98; 99.6% endorsed quite a bit or extremely), expressing thoughts and feelings publicly (M = 3.98; 99.8% endorsed quite a bit or extremely), meeting in groups in public (3.97; 99.6% endorsed quite a bit or extremely), travelling (3.96; 99.6% endorsed quite a bit or extremely), religious practices (3.96; 99.6% endorsed quite a bit or extremely), voting (3.96; 99.2% endorsed quite a bit or extremely), legal services (3.95; 100% endorsed quite a bit or extremely), pressure to accept unwanted documentation (3.95; 99.8% endorsed quite a bit or extremely), restrictions in building or repairing homes (3.90; 99.6% endorsed quite a bit or extremely), pursuing education (3.90; 99.6% endorsed quite a bit or extremely), restrictions related to marriage (3.81; 99.6% endorsed quite a bit or extremely), medical services (3.80; 99.6% endorsed quite a bit or extremely), working (3.78; 98.8% endorsed quite a bit or extremely), and having children (3.65; 98.8% endorsed quite a bit or extremely). For qualitative interpretative data related to these results, see Additional File 2.
TABLE 1 HERE
Trauma Events
Rohingya refugees endorsed a variety of potentially traumatic events occurring in Myanmar. The average number of trauma events endorsed for Myanmar was 19.4. The most frequently endorsed events related to exposure to violence, and included “exposure to frequent gunfire” (98.6%), “witnessed destruction burning of villages” (97.8%), “witnessed dead bodies” (91.8%), and “witnessed physical violence against others” (90.4%). Other events were related to directly experienced physical violence, including torture (55.5%), being beaten (46.1%), stabbed (29.4%), shot (5.1%), etc. Experiences of sexual assault were endorsed by both men and women. Women (33.1%) and men (34.3%) endorsed sexual assault at very similar rates. However, rape (by both security forces and others) was endorsed at a higher rate by women (3.1%) than men (0.8%). For qualitative interpretative data related to these trauma events results, see Additional File 2, and for a list of trauma events disaggregated by gender, see Additional File 3.
TABLE 2 HERE
Daily Stressors
Rohingya refugee respondents reported varying levels of daily stressors during the last month in Bangladesh and previously in Myanmar. In Bangladesh the most frequently endorsed stressors were regarding difficulties with sufficient income (95%), food (79%), limited access to education (72%), and travel (66%). Problems with living space (62%), sanitation facilities (62%), physical health (62%), and water (60%), were other problems endorsed by a majority of participants linked to current life in the camps in Bangladesh. Regarding stressors during their previous time living in Myanmar, the following was most commonly reported: serious problems - because of harassment by police (98%), harassment by the local population (97%), with travel (96%), and with education (84%). The average number of daily stressors endorsed by participants previously in Myanmar was 6.17 while currently in Bangladesh the average number of daily stressors endorsed was 6.34, although notably, the type of stressors differed across the two contexts.
TABLE 3 HERE
Posttraumatic Stress Symptoms
On a scale from 1-4 (1 = not at all, 2 = a little, 3 = quite a bit, 4 = extremely), all PTSD symptoms were endorsed at an average severity score of 2.5 or higher. The items with the highest average severity scores included “recurrent thoughts or memories of the most hurtful or terrifying events” (3.56), “feeling as though the event is happening again” (3.42), “feeling as if [they] don’t have a future” (2.91), and “recurrent nightmares” (2.83). Although the PTSD subscale of the HTQ has not been validated for use with the Rohingya population, a composite cut-off score has typically been used to indicate scores that are diagnostic of PTSD [24], 61.2% of participants endorsed posttraumatic stress symptoms typically diagnostic of PTSD, with the average score for all participants being 2.80. For endorsement rates of all PTSD symptoms, see Additional File 4, and for more information related to the scoring and interpretation of the HTQ, see Additional File 5.
Depression and Anxiety Symptoms
On a scale from 1-4 (1 = not at all, 2 = a little, 3 = quite a bit, 4 = extremely), all anxiety and depression symptoms were endorsed at an average severity score of 2.0 or higher. The anxiety and depression symptoms with the highest average severity scores included “worrying too much about things” (3.49), “feeling sad” (3.40), and “loss of interest in things you previously enjoyed doing” (3.04). Anxiety symptoms with the highest average severity scores included “feeling tense or keyed up” (3.13), “faintness, dizziness, or weakness” (2.73), and “headaches” (2.57). Investigator-developed items were endorsed at the following rates, “feeling humiliated/subhuman” (2.69), “bodily pain from distress/tension” (2.66),[1] “feeling disrespected” (2.54), and “feeling helpless” (2.47). For qualitative interpretative data related to “feeling humiliated/subhuman”, see Additional File 2. Although the HSCL-25 has not been validated for the Rohingya population, a composite cut-off score for the combined anxiety and depression sub-scales has typically been used to indicate scores that are “checklist positive for some type of unspecified emotional distress” related to anxiety and depression [24],[2] 84.0% of respondents endorsed anxiety and depression symptoms typically indicative of emotional distress, with average score for all participants being 2.64. For endorsement rates of all anxiety and depression symptoms, see Additional File 4. For more information related to the scoring and interpretation of the HSCL-25, see Additional File 5.
Table 4: Percentage of respondents reaching diagnostic cutoff scores
Scale
|
Mental health composite score threshold
|
%
|
PTSD
|
Respondents who scored higher than the typically diagnostic cutoff score of 2.5
|
61.2%
|
Emotional Distress (Anxiety and Depression)
|
Respondents who scored higher than the typically diagnostic cutoff score of 1.75
|
84.0%
|
Functioning
Participants endorsed difficulties with daily functioning on a scale from 1-4 (1 = not at all, 2 = a little, 3 = quite a bit, 4 = extremely). Participants on average indicated difficulty carrying out daily tasks (2.87), caring for their hygiene (2.67), and engaging in social activities (2.39). There was a much lower level of difficulty in engaging in religious activities (1.60). For participants that indicated any level of functional difficulty, a follow-up item inquired “What do you attribute these difficulties to?” Respondents were instructed that they could choose more than one response and were given four response options including “current living situation” (71.6%), “mental health” (62.3%), “physical health” (48.2%), and “Other (Specify).” Of the specified qualitative responses, the most common reasons given were related to lack of income/opportunity (5.9%), displacement/statelessness (1.8%), and monsoon season (1.5%). For full text of functioning difficult items and endorsement rates, see Additional File 6: Functioning Difficulty Items.
Prediction Models
A series of initial multiple linear regression models were conducted in order to identify the most robust predictors for the final regression models. Generally, predictors were chosen that exceeded a β cutoff of .1; however, some variables with less than a β of .1 were included based on their broadly documented relationship with outcome variables, as well as their clinical and cultural significance in relation to outcome variables.
The three final models, presented here, predict –
- PTSD symptoms,
- emotional distress (anxiety and depression), and
- functioning[3]
As a reminder, the variable ‘Myanmar systematic human rights violations’ is a sum score that combines most of the items on the systematic human rights violations scale. ‘Trauma history’ is a sum score that combines the lifetime trauma events endorsed by a respondent in both Myanmar and Bangladesh, although nearly all events endorsed occurred in Myanmar. ‘Bangladesh daily stressors’ is a sum score that combines all the daily stressors endorsed in Bangladesh in the last month, while ‘Myanmar daily stressors’ is a sum score of the same stressors, except faced when the participants previously lived in Myanmar. ‘Depression symptoms’ is the composite score of HSCL depression items.
PTSD Symptoms. The final model predicting PTSD symptoms included age, sex/gender, Bangladesh daily stressors, Myanmar daily stressors, trauma history, Myanmar systematic human rights violations, feeling humiliated/subhuman, and feeling helpless.[4] The full regression model was significant in predicting PTSD scores F(8, 469) = 82.05; p < .001, and R2 = .58. Older age (β = .097, p < .01), being a woman (β = -.094, p < .05), a higher number of lifetime trauma events (β = .185, p < .001), higher levels of systematic human rights violations in Myanmar (β = .095, p < .01), a higher number of daily stressors in Myanmar (β = .000, p < .05), higher levels of feeling humiliated/subhuman (β = .313, p < .001), and higher levels of feeling helpless (β = .366, p < .001), significantly predicted higher PTSD scores.
Emotional Distress (Anxiety and Depression). The final model predicting emotional distress (anxiety and depression) symptoms included age, sex/gender, Bangladesh daily stressors, Myanmar daily stressors, trauma history, and Myanmar systematic human rights violations. The full model was significant in predicting distress scores F(6, 471) = 48.47; p < .001, and R2 = .38. Older age (β = .109, p < .01), a higher levels of daily stressors in Bangladesh (β = .105, p < .01), higher levels of daily stressors previously in Myanmar (β = .337, p < .001), a higher number of lifetime trauma events (β = .341, p < .001), and higher levels of systematic human rights violations in Myanmar (β = .160, p < .001) significantly predicted higher emotional distress scores.
Functioning Difficulties. The final model predicting functioning difficulties included the following predictor variables, age, sex/gender, Bangladesh daily stressors, trauma history, PTSD symptoms, and depression symptoms. The full regression model was significant in predicting functioning difficulties F(6, 483) = 66.26; p < .001, and R2 = .45. Higher numbers of Bangladesh daily stressors (β = .336, p < .001), higher levels of depression symptoms (β = .362, p < .001), and higher levels of PTSD symptoms (β = .140, p < .05) significantly predicted higher levels of functioning difficulty.
Table 5: Models predicting PTSD, emotional distress, and functioning difficulty
Variables
|
Model 1: Predicting PTSD symptoms
R2 = .583, F (8, 469) = 82.05 p < .001
|
Model 2: Predicting emotional distress (anxiety and depression)
R2 = .382, F (6, 471) = 48.47 p < .001
|
Model 3: Predicting functioning difficulty
R2 = .451, F (6, 483) = 66.26 p < .001
|
1. Systematic human rights violations
|
Stand. β = .095
|
Stand. β = .160
|
Variable not included in this model
|
t = 2.756
|
t = 3.890
|
p = .006**
|
p = .000**
|
2. Trauma history
|
Stand. β = .185
|
Stand. β = .341
|
Stand. β = .033
|
t = 5.417
|
t = 8.461
|
t = .834
|
p = .000**
|
p = .000**
|
p = .405
|
3. Bangladesh daily stressors
|
Stand. β = .000
|
Stand. β = .105
|
Stand. β = .336
|
t = .007
|
t = 2.725
|
t = 9.258
|
p = .994
|
p = .007**
|
p = .000**
|
4. Myanmar daily stressors
|
Stand. β = .000
|
Stand. β = .337
|
Variable not included in this model
|
t = 2.461
|
t = 9.029
|
p = .014*
|
p = .000**
|
5. Sex
|
Stand. β = -.094
|
Stand. β = .037
|
Stand. β = -.065
|
t = -2.488
|
t = .868
|
t = -1.850
|
p = .013*
|
p = .386
|
p = .065
|
6. Age
|
Stand. β = .097
|
Stand. β = .109
|
Stand. β = -.017
|
t = 3.147
|
t = 2.911
|
t = -.471
|
p = .002**
|
p = .004**
|
p = .638
|
7. Feeling humiliated/subhuman
|
Stand. β = .313
|
Variable not included in this model
|
Variable not included in this model
|
t = 6.885
|
p = .000**
|
8. Feeling helpless
|
Stand. β = .366
|
Variable not included in this model
|
Variable not included in this model
|
t = 8.807
|
p = .000**
|
9. PTSD symptoms
|
Variable not included in this model
|
Variable not included in this model
|
Stand. β = .140
|
t = 1.996
|
p = .047*
|
10. Depression symptoms
|
Variable not included in this model
|
Variable not included in this model
|
Stand. β = .362
|
t = 4.939
|
p = .000**
|
*p < .05, **p < .01
For additional data from final regression models, see Additional File 7.
[1] The bodily pain item was used as part of the anxiety sub-scale and total emotional distress score; however, the remainder of the investigator-created items were examined individually.
[2] Unspecified emotional distress will be referred to in this report as “emotional distress,” and is a combination of anxiety and depression subscales from the HSCL-25.
[3] The multicollinearity statistics were examined for all models, and all predictors had acceptable VIF scores < 5.
[4] Feeling humiliated/subhuman and feeling helpless were items that were developed from discussions with Rohingya key informants and focus groups. These expressions of distress seemed to be distinctly different than items included in the mental health scales utilized in this study. Feeling humiliated/subhuman captures the feelings associated with being referred to as an animal or forced to survive in living situations not suited for human beings. Feeling helpless captures the feeling associated with the lack of empowerment or agency to be able to change or improve one’s situation. Although, during the course of the questionnaire refugees were asked if they felt this way in the last two weeks, it seemed that these feelings built-up over the course of a long period of time. Focus group participants reported feeling both humiliated/subhuman and helpless both historically in Myanmar and currently in their lives in Bangladesh. Results from the correlation matrix show that these items were both significantly correlated with PTSD scores, and were included as independent variables in the final regression model predicting PTSD symptoms.