This article has been cited by other articles in PMC.
Historically, refugees and asylum seekers have had a high probability of experiencing politically motivated torture.1 The United States has resettled many groups of refugees, including these study populations from Somalia and Ethiopia. Having experienced civil war and a lack of formal government for more than a decade,2 Somalis have often suffered traumatic events. Oromos claim ongoing political oppression since their territory was incorporated into the country of Ethiopia at the end of the 19th century.3–5
Estimating the prevalence of torture in community samples of refugees is extremely difficult, often impossible, and rarely attempted. The political and emotional sensitivity of torture makes it difficult to study, and refugees are challenging groups for research.6 Between 5% and 35% of refugees have been tortured, according to the most frequently cited review.7 Existing studies of torture and associated factors typically conducted in refugee clinics and in other treatment settings,1,6,8–16 report that posttraumatic stress, anxiety, depression, and somatization are common.17,18 However, these studies have rarely included control groups, generally have had small samples, and cannot address the prevalence of torture survival in communities. Any consequences specifically associated with torture, compared with other traumatic events that refugees commonly experience, still need to be identified and the effects quantified.19,20
Only a few studies with large samples (n > 500) have examined torture prevalence and posttraumatic stress disorder (PTSD) rates, the focus of our article. From national samples, De Jong et al.21 studied postconflict populations in Algeria, Cambodia, Gaza, and Ethiopia, finding rates of PTSD ranging from 16% to 37%. The prevalence of PTSD in the 1200 Ethiopians surveyed was 16%, higher for torture survivors (P < .001) than for those not tortured. Modvig et al.22 randomly surveyed 1033 household representatives in East Timor and found a torture prevalence rate of 30%.
Symptom levels tend to be higher in refugee camps than in resettlement populations.23–26 Mollica et al.,27,28 studying 993 Cambodians in a Thai refugee camp, found that a third had PTSD. Comparing 526 Bhutanese torture survivors in a Nepalese refugee camp with matched controls, Shrestha et al.29 found higher posttraumatic stress and anxiety. Van Ommeren et al.30 subsequently randomly sampled 810 (418 tortured and 392 nontortured) Bhutanese refugees from the same frame (the general source from which the sample population is selected), finding that torture survivors had more occurences of PTSD (43% vs 4%).
Among refugees living in nearby countries, Iacopino et al.31 found 4% torture prevalence among 1180 randomly sampled households of Kosovars living in Macedonia and Albania. Smaller samples included Senegalese refugees in Gambian camps (16% torture prevalence),32 Tibetan nuns and lay students tortured in Tibet but living in India (54% anxiety vs 29% in controls),33 and Burmese political dissidents in Thailand (23% PTSD).34
In smaller prisoner populations in Turkey, Paker et al.35 estimated that tortured Turkish prisoners had significantly more PTSD compared with other prisoners, and Basoglu et al.36 showed higher rates of lifetime (33%) and current (18%) PTSD than controls.
In Western resettlement populations, Thonneau et al.37 found 8% torture prevalence among 1194 refugee applicants to Canada. Smaller samples have shown PTSD rates among Cambodians of 12%,38 50%,39 and even 86%.40
Our 5-year, multiphased, community-based epidemiological study aimed to identify demographic characteristics, pre- and postmigration factors, torture prevalence, and the association of torture survival with health and social problems in 2 resettled refugee communities. The findings of a quantitative survey of 1134 adult participants are presented here. Two subsequent surveys, using subsets from this sample, will compare torture survivors with nontortured refugees using (1) structured instruments to assess symptoms, disability, coping, social support, and family function and (2) a brief neurological screen to identify soft signs of impairment and the Schedules for Clinical Assessment in Neuropsychiatry41 to make Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition,42 and International Classification of Diseases, 10th Revision43 psychiatric diagnoses.
METHODS
Questionnaire Development and Interview Administration
Questionnaires were administered by 8 Somali and Oromo staff with experience in health care or interviewing and backgrounds in medicine, law, engineering, biology, or social services.
Investigators provided training in survey methods and research protocols. To prevent participants from associating the study with torture, we used the terms health problems and physical or psychological abuseinstead of the word torture in the informational documents and consent forms used. Three authors (M. S., C. R., D. J.) held biweekly sessions with staff, providing debriefing and ongoing supervision during data collection.
Because using structured assessment instruments for the entire study was too expensive and impractical for a large sample, we chose a multiphased design. We developed a comprehensive closed-ended questionnaire for this first interview phase. Although questionnaires to assess trauma and torture are described in the literature,44–47 their validity for East African refugees has not been documented. Our questionnaire consisted of 188 questions with 537 response items adapted from published studies and the authors’ clinical and research experiences. The questionnaire elicited biographical information; current and earlier life circumstances; health status; and experiences of violence, deprivation, and physical or psychological trauma and torture. Questions were asked in a variety of formats, including yes/no, Likert scale, and checklist. The most sensitive trauma and torture questions were asked at the end of the questionnaire, after staff had the opportunity to gain participants’ trust. Checklists, an accepted assessment method for torture techniques,46 were used, and scales were developed to estimate physical, psychological, and social problems.
Because PTSD is one of the most common and controversial diagnoses associated with torture, the PTSD Checklist (PCL-C) was administered in this interview phase and, for comparison, in the next 2 phases. The PCL-C, a self-report Likert scale with 17 items, has shown high internal consistency and reliability and strong correlation with PTSD diagnosis using the Clinician Administered PTSD Scale.48,49 Translated versions of the PCL-C into Oromo and Somali produced high reliability with Cronbach’s α (.93).
Staff members were matched with participants by gender and ethnicity. The questionnaire was self-administered for literate participants (50.4%) and interviewer administered for those who were illiterate (49.6%). These percentages were not significantly different across the groups exposed and unexposed to torture (χ2 = 0.36, df = 1, P = .55). Staff remained with all participants during questionnaire administration to ensure that the participants understood the items and to observe for signs of distress. The questionnaire was translated from English into the Somali and Oromo languages and back-translated using standard techniques.50–60
Sampling and Classification
We estimated that a sample size of 1200 would allow us to detect a doubling of the rate of PTSD in the tortured group from a base prevalence of 10% (power = 80%, α = .05, 2-tailed).
Oromos or Somalis living in the Minneapolis/St. Paul metropolitan area and who were at least 18 years old were eligible. Exclusion criteria included psychological inability to participate or residence in a household where a relative had already been interviewed. Unrelated persons living in the same household were eligible.
Census data were not available for our study because most Oromo and Somali refugees arrived after the 1990 census, and interviewing for this sample (which took place between July 14, 1999, and September 3, 2001) began before the Year 2000 census. Even if census data had been available, the census has historically undercounted refugees and minorities.61 State data were kept only for initial resettlement of refugees, and community organizations and agencies lacked complete data.
Because a random sampling frame was unavailable, a combination of nonrandom sampling approaches62–64 and lengthy recruitment (more than 25 months) was used. Sampling approaches included targeting persons associated with community organizations and geographic locations (62%) and sampling by linkage (38%). The participation rate was 97.1% of all invited.
Lacking the ability to assess true representativeness of the sample, we compared the sample demographics with newly available outside data in order to identify any differences between our sample and the underlying populations.65 The outside data were used to estimate community size, geographic distribution, and demographic feaures and included public school enrollment reports, birth records, and state refugee resettlement data. No large differences were found between the sample and the underlying populations, suggesting that the sample was representative. The outside data allowed us to estimate the Minneapolis/St. Paul populations as 641 Oromos and 6538 Somalis aged 18 years or older. Consequently, the Oromo sample (n = 512) may have represented 80% of potential participants, and the Somali sample (n = 622) may have represented only 8%.
The United Nation’s (UN) definition of torture66,67 formed the basis for classifying participants as tortured. Key components of this definition include physical or psychological pain and suffering, intentionally inflicted for any reason, based on discrimination and perpetrated by persons acting officially. Although torture can occur in many nonpolitical settings, such as domestic violence or satanic cults, “official” perpetration differentiates the UN definition from others. During administration of the questionnaire, staff clarified the context of traumatic events to ensure compatibility with the UN definition. For Somalis, who have had no formal government since 1991, “official” perpetrators included opposing clans who had taken power.
Participants were classified as torture survivors if they (1) responded in the positive to any of 3 items directly asking whether they had been tortured (Have you been tortured in prison? [Y/N]; Was tortured [marked off on a checklist]; Were you tortured in prison or jail? [Y/N]) and reported experiencing at least 1 identified torture technique item (details available from authors) or (2) reported experiencing 1 of the subsets of torture techniques that investigators considered could be used only during torture sessions (details available from authors), even if participants responded in the negative to all of the questions about torture exposure.
Data Collection, Management, and Analysis
Staff explained the purpose, procedures, risks, and benefits of the study to potential participants and their rights to refuse participation at any time throughout the interview process. Literate persons read the consent form in the language of their choice, and interviewers read it to those who were illiterate. Staff obtained signed or oral informed consent from subjects before they began the interviews.
Two strategies were undertaken to appraise the level of cooperation and credibility of participants. (1) Staff rated participants after completion of interviews using a 1 to 5 Likert scale, with 1 being the most positive. The median rating was 2 (very cooperative or credible). (2) The questionnaire contained 4 internal validity measures identifying participants with response inconsistency on item pairs,68 and, for more than 95% of the sample, more missing items,69 a greater number of unusually virtuous responses,70(i.e., responses designed to assess if respondents presented themselves in an overly favorable manner) and a greater number of extreme responses.71
Those participants identified according to these 4 validity measures were classified as “suspect.” On analysis, the suspect group of participants was found to be associated with torture exposure, Oromo ethnicity, and younger age. Results for all participants were compared with a subset that excluded the suspect records. Because subjective ratings of cooperation and credibility were not different between the suspect and nonsuspect groups, results were presented for the entire sample. Data were double entered, and periodic error checks were made. Summary data for each variable were extensively checked, and out-of-range responses were compared with the original paper questionnaires.
The trauma count for each participant was controlled in order to assess any additional effect of torture beyond other types of trauma. Several specific items to measure social, psychological, and physical problems were incorporated into the questionnaire, combined into summative scales, and refined using an item analysis. These scales, meant to provide only a preliminary indication of problem areas, are presented with the Cronbach αs in Table 1▶.
TABLE 1—
Problem Scales and Cronbach αs: Minneapolis, Minn, July 14, 1999–September 3, 2001
| Social Problems | Psychological Problems/Source of Problems | Physical Problems |
| Does not speak English easily | Had child die | Starvation before leaving homeland |
| Does not read English easily | Had death in family in last 6 months | Currently taking medication |
| Does not have job | Has no good work opportunities in United States | On journey from homeland: |
| Has less than high school education | Feels stress living in United States | Life-threatening illness |
| Is separated from family | Bothered by things done in home country | Life-threatening lack of food |
| Has less money to spend than in homeland | Has trouble sleeping | Life-threatening lack of water |
| Does not anticipate opportunities for work | Has loss of appetite | Life-threatening serious injury |
| Does not have good friends | Hears voices | Life-threatening physical assault |
| Has difficulty caring for monthly expenses | Has thoughts of killing self | Head injuries from torture |
| Feels alone | Has been to doctor for mental health problem | Physical problems resulting from torture |
| Has problems getting job | Has intense memories of torture | Still has physical reactions from torture |
| Family has stayed behind | Has frequent headaches | Has faintness or dizziness |
| Has problems learning English | Has had changes in appetite | |
| Has hard time understanding American life | ||
| Likes home food better than American food | ||
| α = .77 | α = .63 | α = .77 |
Descriptive statistical analysis revealed the distribution of variables and the appropriateness of statistical tests for the data. Stepwise logistic regression was used to compare suspect and nonsuspect groups. Bivariate comparisons between torture groups were conducted using χ2 and t tests. Major outcomes of interest included social, physical, and psychological problems, and total PCL-C scores. For the interval measures, multivariable analysis was accomplished using stepwise multiple linear regression. Regression diagnostics were performed for each model to assess any possible violations of assumptions. Analysis was performed using SPSS version 8 (SPSS Inc, Chicago, Ill) and SAS version 8 (SAS Institute Inc, Cary, NC).
RESULTS
The final sample of 1134 included 622 Somalis and 512 Oromos, 605 men and 529 women. Table 2▶describes characteristics of the ethnic/gender subgroups. On average, participants reported experiencing 21 of 61 possible nontorture traumatic events. All but 6 participants reported experiencing traumatic experiences.
TABLE 2—
Characteristics of the Sample by Ethnic/Gender Group
| Total (n = 1134) | Oromo Men (n = 282) | Oromo Women (n = 230) | Somali Men (n = 323) | Somali Women (n = 299) | F Statistic | PValue | |
| Age, y, mean (SD) | 35.1 (13.9) | 31 (12.4) | 34.8 (13.9) | 37.4 (15.9) | 36.8 (14.7) | 11.63,1123 | <.001 |
| Age at time of leaving home, mean (SD) | 27.5 (14.5) | 24.6 (12.1) | 27.8 (14.4) | 28.9 (16.0) | 28.4 (14.7) | 5.03,1123 | .002 |
| Number of traumatic events endorsed, mean (SD) | 20.8 (11) | 25.3 (10.4) | 25.2 (8.2) | 14.1 (8.5) | 20.2 (11.8) | 83.53,1130 | <.001 |
| Years between leaving home and arriving in United States, mean (SD) | 4.1 (3.2) | 3.3 (3.0) | 3.1 (2.8) | 4.9 (3.2) | 4.8 (3.3) | 25.913,1100 | <.001 |
| Years in United States, mean (SD) | 3.4 (3.6) | 3.1 (4.2) | 3.9 (4.4) | 3.4 (3.2) | 3.2 (2.8) | 1.93,1108 | .12 |
| Marital status, no. (%) | |||||||
| Married, living with partner | 319 (29) | 68 (25) | 93 (41) | 69 (22) | 89 (30) | ||
| Separated by immigration | 264 (24) | 78 (29) | 34 (15) | 86 (27) | 66 (23) | ||
| Single | 342 (31) | 122 (45) | 58 (25) | 115 (37) | 47 (16) | ||
| Separated, divorced, widowed | 180 (16) | 2 (1) | 44 (19) | 43 (14) | 91 (31) | 156.39 | <.001 |
| Education, no. (%) | |||||||
| No formal education | 167 (15) | 14 (5) | 62 (27) | 14 (4) | 77 (26) | ||
| Less than high school | 418 (37) | 104 (38) | 103 (45) | 117 (36.5) | 94 (31) | ||
| High school diploma | 292 (26) | 58 (21) | 31 (14) | 115 (36) | 88 (30) | ||
| Post high school | 245 (22) | 99 (36) | 33 (14) | 75 (23.5) | 38 (13) | 167.29 | <.001 |
| Speaks English, no. (%) | 633 (56) | 198 (70) | 101 (44) | 224 (69) | 110 (37) | 104.83 | <.001 |
| Receives government aid, no. (%) | 83 (7) | 3 (1) | 28 (12) | 12 (4) | 40 (13) | 46.63 | <.001 |
| Employed, no. (%) | 577 (51) | 198 (72) | 103 (45) | 188 (58) | 88 (30) | 114.13 | <.001 |
| Owns home, no. (%) | 64 (6) | 25 (9) | 26 (11) | 6 (2) | 7 (2) | 34.23 | <.001 |
| Has no permanent address, no. (%) | 35 (3) | 5 (2) | 8 (4) | 9 (3) | 13 (4) | 3.43 | .33 |
| Muslim, no. (%) | 990 (87) | 204 (72) | 185 (80) | 313 (97) | 288 (96) | 115.53 | <.001 |
| Christian, no. (%) | 105 (9) | 62 (22) | 43 (19) | 0 | 0 | 142.23 | <.001 |
Problem Scales and PCL-C
Table 3▶ displays results of the analyses assessing variables associated with the problem scales and PCL-C. Trauma count and exposure to torture were significant at P < .0001 for all these measures. In addition, the number of social problems was greater among women, Somalis, and those who had decreased their religious practices since immigrating. English language fluency, employment, high school graduation, marriage, home ownership, and longer residency in the United States were associated with fewer social problems. Male gender and Somali ethnicity were associated with more physical problems, whereas employment, marriage, and high school graduation were associated with fewer physical problems.
TABLE 3—
Multivariable Analysis of Characteristics and Problems
| Social Problems | Physical Problems | Psychological Problems | Total PCL-C Score | |||||
| β (SE) | PValue | β (SE) | PValue | β (SE) | PValue | β (SE) | PValue | |
| Gender: female | 1.54 (0.13) | <.0001 | (1.0 (0.13) | <.0001 | 0.16 (0.10) | .0899 | (3.35 (0.88) | .0001 |
| Ethnicity: Oromo | (0.71 (0.14) | <.0001 | (0.78 (0.14) | <.0001 | 0.02 (0.10) | .8650 | 2.29 (0.94) | .0149 |
| Age | (0.01 (0.03) | .7023 | (0.02 (0.03) | .5347 | (0.02 (0.02) | .3209 | 0.04 (0.19) | .8336 |
| Married or living with partner | (0.39 (0.13) | .0036 | (0.40 (0.14) | .0035 | (0.18 (0.10) | .0764 | (1.66 (0.91) | .0685 |
| High school graduate | (1.94 (0.13) | <.0001 | (0.33 (0.14) | .0159 | 0.01 (0.10) | .8877 | (0.85 (0.93) | .3615 |
| Speaks English | (2.01 (0.14) | <.0001 | (0.22 (0.15) | .1289 | (0.04 (0.11) | .6848 | −0.57 (0.98) | .5606 |
| Has job | (1.27 (0.13) | <.0001 | (0.44 (0.14) | .001 | (0.18 (0.10) | .0709 | (0.55 (0.90) | .5436 |
| Owns home | (0.84 (0.26) | .0013 | (0.07 (0.27) | .8027 | (0.29 (0.19) | .1381 | 0.95 (1.85) | .6048 |
| Age at which left home country | 0.04 (0.03) | .0979 | 0.02 (0.03) | .4239 | 0.05 (0.02) | .0189 | (0.01 (0.19) | .9752 |
| Years between leaving home and arriving in United States | 0.06 (0.03) | .0472 | 0.02 (0.03) | .5977 | 0.02 (0.02) | .3423 | (0.35 (0.22) | .1046 |
| Years in United States | 0.01 (0.03) | .6347 | 0.03 (0.03) | .2778 | 0.06 (0.23) | .0163 | (0.20 (0.23) | .3795 |
| Religious practices increased | (0.21 (0.18) | .2570 | 0.18 (0.19) | .3144 | 0.08 (0.14) | .5740 | 3.19 (1.28) | .0131 |
| Religious practices decreased | 0.53 (0.24) | .0293 | 0.25 (0.25) | .3398 | 0.76 (0.18) | <.0001 | 4.67 (1.67) | .0051 |
| Trauma count | 0.08 (0.01) | <.0001 | 0.15 (0.01) | <.0001 | 0.09 (0.01) | <.0001 | 0.52 (0.05) | <.0001 |
| Exposure to torture | (0.59 (0.14) | <.0001 | 0.60 (0.15) | <.0001 | 0.61 (0.11) | <.0001 | 7.18 (0.98) | <.0001 |
| Adjusted R2 | 0.7170 | 0.5091 | 0.4962 | 0.4122 | ||||
Note. PCL-C = Posttraumatic Stress Syndrome Checklist.
None of the assessed factors were associated with fewer psychological problems. More psychological problems were associated with leaving home at an older age, longer residency in the United States, and decreased religious practices since immigrating. Female gender was inversely associated with PCL-C score, whereas Oromo ethnicity and any change in religious practices since immigrating were associated with increased PCL-C scores.
Associations with fewer problems on at least 2 problem scales/PCL-C included employment, high school graduation, marriage, and continued religious practices, whereas more problems were associated with decreased religious practices, high trauma count, and torture exposure. Despite high scores on the scales, less than 1% of the participants requested or accepted referral to mental health services.
Prevalence of Torture and Characteristics of Survivors
Prevalence rates for each torture category were calculated (Table 4▶). Of the 1134 participants, 44% met criteria for torture exposure and 56% did not. Criteria for classification are described in the Methods section. Only 15% (n = 92) met criteria solely by reporting that they experienced 1 or more of the techniques occurring only during torture (false negatives). Conversely, 40 participants were excluded from the torture group because, although they gave a positive answer to a torture question, they did not report experiencing any torture techniques (false positives).72 Had we accepted, as the only criterion, an affirmative response when asking participants whether they had been tortured, 132 (11.6%) would possibly have been misclassified according to our definition. Fifty torture survivors (10%) reported experiencing only physical or only psychological techniques, whereas the other 90% reported experiencing both.
TABLE 4—
Demographic Characteristics of the Sample by Torture Classification
| Exposed to Torture | Unexposed to Torture | χ2 | P Value | |
| Entire sample (n = 1134) | 502 (44) | 632 (56) | ||
| Ethnicity | 42.61 | <.0001 | ||
| Somali, no. (%) (n = 622) | 221 (36) | 401 (64) | ||
| Oromo, no. (%) (n = 512) | 261 (55) | 231 (45) | ||
| Gender | 0.741 | .3900 | ||
| Male, no. (%) (n = 605) | 272 (45) | 333 (55) | ||
| Female, no. (%) (n = 529) | 228 (43) | 302 (57) | ||
| Married or living with partner, no. (%) | 120 (25) | 199 (32) | 7.791 | .0052 |
| Graduated from high school, no. (%) | 230 (47) | 307 (49) | 0.601 | .4387 |
| Employed, no. (%) | 243 (49) | 334 (53) | 1.551 | .2128 |
| Owns home, no. (%) | 30 (06) | 34 (05) | 0.191 | .6656 |
| Has no permanent address, no. (%) | 19 (04) | 16 (03) | 1.471 | .2255 |
| Current age, mean (SD) | 36.78 (14) | 33.78 (15) | 3.451125 | .0006 |
| Age on leaving home, mean (SD) | 29.50 (14) | 25.90 (15) | 4.191109 | <.0001 |
| Years between leaving home and arriving in United States, mean (SD) | 4.04 (03) | 4.12 (03) | 0.661102 | .5119 |
| Years in United States, mean (SD) | 3.04 (04) | 3.66 (04) | 2.821110 | .0048 |
| Number of traumas endorsed, mean (SD) | 28.07 (10) | 14.94 (08) | 25.021132 | <.0001 |
| Number of social problems, mean (SD) | 7.09 (04) | 6.43 (03) | 3.261132 | .0012 |
| Number of psychological problems, mean (SD) | 3.67 (02) | 1.74 (01) | 19.141132 | <.0001 |
| Number of physical problems, mean (SD) | 4.32 (03) | 1.72 (02) | 18.881132 | <.0001 |
| PCL-C scores, mean (SD) | 42.53 (15) | 27.17 (10) | 18.53919 | <.0001 |
Note. PCL-C = Posttraumatic Stress Syndrome Checklist.
Torture history varied by gender and ethnicity: men (n = 272, 45% of the men) and women (n = 228, 43% of the women) had approximately equal exposure to torture, and more Oromos were exposed to torture (n = 286, 55%) than Somalis (n = 224, 36%). Among ethnic/gender groups, those most often exposed to torture were Oromo men (n = 194, 69%) and Somali women (n = 141, 47%), followed by Oromo women (n = 85, 37%) and Somali men (n = 81, 25%). Gender differences were not statistically significant (Table 3▶), but differences by ethnicity (χ2 = 42.6, df = 1, P < .001) and by ethnic/gender group (χ2 = 126.0, df = 3, P < .001) were significant.
Participants in the tortured and nontortured groups differed in several characteristics and in the degree of reported problems (Table 3▶). Torture survivors were less often married and were older when they left their home countries. No significant between-group differences were found for men or for those employed, more educated, with permanent addresses, or with longer times between leaving their home countries and arriving in the United States.
For both groups, a higher number of traumatic events correlated positively with scores on all problem scales and total PCL-C. The tortured group averaged 13 more traumatic events than the nontortured group (28 vs 15), adding an average of 1 social and 1 psychological problem, 2 physical problems, and 7 points to the PCL-C scale for those exposed to torture. Beyond the general trauma, torture exposure added small but significant increases to the number of psychological and physical problems but not to the number of social problems. For social problems, the estimated adjusted means were 6.46 (SE = 0.11) for those exposed and 7.03 (SE = 0.09) for those unexposed to torture, showing no significant statistical difference.
The most striking correlate of torture exposure was the increase in total PCL-C score. Those exposed to torture averaged an additional 7-point (> 10%) increase in PCL-C score beyond the increase due to trauma alone. Using a cutoff PCL-C score of greater than 50 (range 17 to 85) to indicate suspected PTSD, 123 (25%) of those exposed to torture met this criterion, but only 23 (4%) of the unexposed did.
DISCUSSION
Our study interviewed 1134 East Africans, the largest refugee community sample conducted in a resettlement country. As expected, social problems were common throughout the sample, because the early stages of resettlement for all refugees are characterized by many problems in daily living and adjustment.9,73,74 Factors associated with fewer social problems included speaking English and graduating from high school, which presumably increased the chances for employment and home ownership. Marriage can provide social stability, and longer residency in the United States can allow time for adjustment. High school education, employment, and marriage were associated with fewer physical problems, perhaps because those with fewer physical problems could adjust more successfully. However, none of these factors were associated with fewer psychological problems.
Low scores on at least 2 of the problem scales/PCL-C were associated with employment, high school graduation, marriage, and continuing religious practices since immigration. These factors indicate the importance of education, economic stability, social support, and religion for the well-being of refugees. High scores for trauma and torture, as expected, were associated with more problems.
Torture rehabilitation centers have historically claimed that the more educated men were most often targeted as leaders who would serve as examples to their communities.1 Our study, in contrast, showed that men and those with higher education levels were no more likely to be tortured than women or those with less education. Although more time in transition to the United States, usually in refugee camps, might be expected to increase the risk of trauma, our study did not find any increase for those experiencing longer transitions.
Except for Somali men (25% of whom were tortured), the prevalence of torture exposure in this community sample was significantly higher than the 5% to 35% often reported.7 This high prevalence was found despite the expectation that participants would underreport.8 If these results can be replicated in other refugee communities, 400 00075 is likely an underestimation of the number of torture survivors living in the United States.
Women were as likely to experience torture as men, an exposure rate not previously reported. Because civilians are increasingly affected by modern warfare and terrorism, it is not surprising that women frequently experience torture.76,77
Significant differences in torture exposure by ethnicity and ethnic/gender group were found. Oromos were tortured more often than Somalis, whereas Oromo men and Somali women were the ethnic/gender groups most often tortured. A number of possible explanations can be posited. The very high rates in the Oromo community may reflect long-standing interethnic conflicts. Somali women were more often tortured than Somali men. Anecdotally, Somali men were either killed in their home country or able to escape unharmed, whereas women and children had a more difficult time leaving the country.
Our results suggest that torture’s effect may be additive to other forms of trauma. Torture survivors were more likely than other refugees to experience physical and psychological problems, even after we analyzed the group differences using nontorture trauma as a covariate. The most striking correlate of torture exposure in our study was the increase in total PTSD symptoms. Of those tortured, 25% had suspected PTSD compared with only 4% of those not tortured. These rates are comparable to those cited in earlier studies. In addition, PTSD also showed the greatest additive effect of torture. In another study, Silove et al.19 found an additive effect of PTSD in Tamil torture survivors in Australia after they accounted for other traumatic events. Although the published literature is contradictory, our findings also support a possible dose–response relationship between torture and PTSD.
Several limitations were inherent in the study design. Because we could not use true random sampling methods, our prevalence results are estimates. However, our analyses indicate that an estimated 80% of the Oromo community, nearly the entire Oromo “village,” may have been sampled. With such a comprehensive sample of Oromo, the extremely high torture prevalence of Oromo men (69%) is plausible. The study variables were measured at the same time, a limitation shared by all cross-sectional studies. Therefore, inferring causality between torture history and problems was not possible. Although the present analyses strongly support associations between torture history and key problem areas, our 2 follow-up studies will further elucidate these findings. The final clinician-administered diagnostic phase of our research will also assess the duration and frequency of torture experiences in greater depth.
Our study of 2 communities represents only a small proportion of refugees and displaced persons but contributes to the enormous gap in understanding of the effects of torture and trauma on refugee populations worldwide. The capacity to care for survivors falls far short of the need. This study suggests that the shortfall is even greater than previously thought. The 31 torture rehabilitation centers in the United States and the more than 200 worldwide78 cannot treat all torture survivors.
From a public health perspective, our findings warrant screening refugees for a history of torture at least among East African and women refugees. However, less than 1% of our highly traumatized sample either requested or followed up a referral to Western mental health services. This highlights the need to investigate the reasons for underutilization and to adequately address the needs of torture survivors living not only in Minnesota but throughout the world.
Acknowledgments
This research was funded by the National Institutes of Health (National Institute of Mental Health) grants 5R01-MH59579 and 1R01-MH59570 (project officer Farris Tuma, ScD.)
The authors thank Dr Alan Lifson for his contributions to the initial development of this study and Dr Russell Luepker for his insights into management of the research project. The authors acknowledge the valuable contributions in advising, translating, recruiting, and interviewing by our Somali community coordinators, Nadifa Osman, Dr Abdullahi Gas, Dr Osman Ahmed, Dr Abdulqadir Omar, and Dr Mohamed Wadi, and by our Oromo community coordinators, Israel Gobena, Johara Mohammed, and Elizabeth Namarra. We are grateful to our research assistants, Michelle Peterson, Katherine Grimm, MPH, and Dr Sheng Zhong for data management.
Human Participant Protection
The University of Minnesota institutional review board and human subjects protection program and the HealthPartners Foundation institutional review board approved the study before interviewing began and subsequently approved protocol changes, consent forms, and annual progress reports.
The University of Minnesota institutional review board and human subjects protection program and the HealthPartners Foundation institutional review board approved the study before interviewing began and subsequently approved protocol changes, consent forms, and annual progress reports.
Notes
Contributors
All authors contributed to the conceptualization of ideas, interpretation of findings, and review of the article. J. Jaranson (principal investigator) conceived the study, drafted the article, and supervised operations and personnel. J. Butcher contributed to design and developed the questionnaire. L. Halcon contributed the epidemiological perspective for the research. D. R. Johnson contributed to design and provided insights from his work at the Center for Victims of Torture. C. Robertson contributed to design from the public health nursing perspective. K. Savik supervised the analyses of the data and the research assistants. M. Spring was project coordinator, supervised the interview staff, ensured the quality of the data, and provided a cross-cultural perspective. J. Westermeyer contributed to design with his extensive background in cross-cultural psychiatry.
All authors contributed to the conceptualization of ideas, interpretation of findings, and review of the article. J. Jaranson (principal investigator) conceived the study, drafted the article, and supervised operations and personnel. J. Butcher contributed to design and developed the questionnaire. L. Halcon contributed the epidemiological perspective for the research. D. R. Johnson contributed to design and provided insights from his work at the Center for Victims of Torture. C. Robertson contributed to design from the public health nursing perspective. K. Savik supervised the analyses of the data and the research assistants. M. Spring was project coordinator, supervised the interview staff, ensured the quality of the data, and provided a cross-cultural perspective. J. Westermeyer contributed to design with his extensive background in cross-cultural psychiatry.
Peer Reviewed
References
1. Jaranson JM. The science and politics of rehabilitating torture survivors: an overview. In: Jaranson JM, Popkin, MK, eds. Caring for Victims of Torture. Washington, DC: American Psychiatric Press; 1998:15–40.
2. Putman DB, Noor MC. The Somalis: Their History and Culture. Washington, DC: The Refugee Service Center, Center for Applied Linguistics; 1993. CAL Refugee Fact Sheet 9.
3. Trueman T. Genocide against the Oromo people of Ethiopia? Western influence. Malvern, UK: Oromia Support Group. Occasional paper: June 7, 2001.
4. Country Profile: Ethiopia, Eritrea, Somalia, Djibouti. London, England: Economist Intelligence Unit; 1994–1995.
5. Annual Human Rights Reports, 1997. Washington DC: US State Department: Bureau of Democracy, Human Rights and Labor; 1998.
6. Basoglu M, Jaranson JM, Mollica R, Kastrup M. Torture and mental health: a research overview. In: Gerrity E, Keane TM, Tuma F, eds. The Mental Health Consequences of Torture. New York, NY: Kluwer Academic/Plenum Publishers; 2001:35–62.
7. Baker R. Psychological consequences for tortured refugees seeking asylum and refugee status in Europe. In: Basoglu M, ed. Torture and Its Consequences: Current Treatment Approaches. Cambridge, England: Cambridge University Press; 1992:83–101.
8. Jaranson JM, Kinzie JD, Friedman M, et al. Assessment, diagnosis, and intervention. In: Gerrity E, Keane TM, Tuma F, eds. The Mental Health Consequences of Torture. New York, NY: Kluwer Academic/Plenum Publishers; 2001:249–275.
9. Jaranson JM, Martin SF, Ekblad S. Refugee mental health: issues for the new millennium. In: Manderscheid RW, Henderson MJ, eds. Mental Health, United States, 2000. Rockville, Md: US Department of Health and Human Services; 2001:120–133.
10. Jaranson JM, Popkin MK, eds. Caring for Victims of Torture. Washington, DC: American Psychiatric Press; 1998.
11. Basoglu M, ed. Torture and Its Consequences: Current Treatment Approaches. Cambridge, England: Cambridge University Press; 1992.
12. Lavik NJ, Hauff E, Skrondal A, et al. Mental disorder among refugees and the impact of persecution and exile: some findings from an out-patient population. Br J Psychiatry. 1996;169:726–732. [PubMed] [Google Scholar]
13. Moore LJ, Boehlein JK. Posttraumatic stress disorder, depression, and somatic symptoms in US Mien patients. J Nerv Ment Dis. 1991;179:728–733. [PubMed] [Google Scholar]
14. Boehnlein JK, Kinzie JD. Refugee trauma. Transcultural Psychiatr Res Rev. 1995;32:223–252. [Google Scholar]
15. Mollica RF, Wyshak G, Lavelle J. The psychosocial impact of war trauma and torture on Southeast Asian refugees. Am J Psychiatry. 1987;144:1567–1572. [PubMed] [Google Scholar]
16. Ekblad S, Roth G. Diagnosing posttraumatic stress disorder in multicultural patients in a Stockholm psychiatric clinic. J Nerv Ment Dis. 1997;185:102–107. [PubMed] [Google Scholar]
17. Westermeyer J, Bouafuely M, Neider J, et al. Somatization among refugees: an epidemiological study. Psychosomatics. 1989;30:34–43. [PubMed] [Google Scholar]
18. Beiser M, Fleming JA. Measuring psychiatric disorder among Southeast Asian refugees. Psychol Med.1986;16:627–639. [PubMed] [Google Scholar]
19. Silove D, Steel Z, McGorry P, et al. The impact of torture on post-traumatic stress symptoms in war-affected Tamil refugees and immigrants. Compr Psychiatry. 2002;43:49–55. [PubMed] [Google Scholar]
20. Steel Z, Silove D, Bird K, et al. Pathways from war trauma to posttraumatic stress symptoms among Tamil asylum seekers, refugees, and immigrants. J Trauma Stress. 1999;12:421–435. [PubMed] [Google Scholar]
21. De Jong JTVM, Komproe IH, Van Ommeren M, et al. Lifetime events and posttraumatic stress disorder in 4 postconflict settings. JAMA. 2001;286:555–562. [PubMed] [Google Scholar]
22. Modvig J, Pagaduan-Lopez J, Rodenburg J, Salud CM, Cabigon RV, Panelo CI. Torture and trauma in post-conflict East-Timor. Lancet. 2000;356:1763. [PubMed] [Google Scholar]
23. Hinton WL, Chen YCJ, Du N. DSM-III-R disorders in Vietnamese refugees: prevalence and correlates. J Nerv Ment Dis. 1993;181(2):113–122. [PubMed] [Google Scholar]
24. Carlson EB, Rosser-Hogan R. Cross-cultural response to trauma: a study of traumatic experiences and posttraumatic symptoms in Cambodia refugees. J Trauma Stress. 1994;7(1):43–58. [PubMed] [Google Scholar]
25. Harding RK, Looney JG. Problems of Southeast Asian children in a refugee camp. Am J Psychiatry.1977;134:407–411. [PubMed] [Google Scholar]
26. Sughandabhirom B. Experiences in a first asylum country: Thailand. In: Williams CL, Westermeyer J, eds. Refugee Mental Health in Resettlement Countries. New York, NY: Hemisphere Publishing Corp; 1986:81–96.
27. Mollica RF, Donelan K, Tor S, et al. The effect of trauma and confinement on functional health and mental health status of Cambodians living in Thailand-Cambodia border camps. JAMA. 1993;270:581–586. [PubMed] [Google Scholar]
28. Mollica RF, Poole C, Tor S. Symptoms, functioning, and health problems in massively traumatized populations: the legacy of Cambodian tragedy. In: Dohrenwend BP, ed. Adversity, Stress, and Psychopathology. New York, NY: Oxford University Press; 1998:34–51.
29. Shrestha NM, Sharma B, Van Ommeren M, et al. Impact of torture on refugees displaced within the developing world: symptomatology among Bhutanese refugees in Nepal. JAMA. 1998;280:443–448. [PubMed] [Google Scholar]
30. Van Ommeren M, de Jong JT, Sharma B, et al. Psychiatric disorders among tortured Bhutanese refugees in Nepal. Arch Gen Psychiatry. 2001;58:475–482. [PubMed] [Google Scholar]
31. Iacopino V, Frank, MW, Bauer HM, et al. A population-based assessment of human rights abuses committed against ethnic Albanian refugees from Kosovo. Am J Public Health. 2001;91:2013–2018. [PMC free article] [PubMed] [Google Scholar]
32. Tang SM, Fox SH. Traumatic experiences and the mental health of Senegalese refugees. J Nerv Ment Dis. 2001;189:507–512. [PubMed] [Google Scholar]
33. Holtz TH. Refugee trauma versus torture trauma: a retrospective controlled cohort study of Tibetan refugees. J Nerv Ment Dis. 1998;186:24–34. [PubMed] [Google Scholar]
34. Allden K, Poole C, Chantavanich S, et al. Burmese political dissidents in Thailand: trauma and survival among young adults in exile. Am J Public Health. 1996;86:1561–1569. [PMC free article] [PubMed] [Google Scholar]
35. Paker M, Paker O, Yuksel S. Psychological effects of torture: an empirical study of tortured and non-tortured non-political prisoners. In: Basoglu M, ed. Torture and Its Consequences: Current Treatment Approaches. Cambridge, England: Cambridge University Press; 1992:72–82.
36. Basoglu M, Paker M, Paker O, et al. Psychological effects of torture: a comparison of tortured with non-tortured political activists in Turkey. Am J Psychiatry. 1994;151:76–81. [PubMed] [Google Scholar]
37. Thonneau P, Gratton J, Desrosiers G. Health profile of applicants for refugee status (admitted into Quebec between August 1985 and April 1986). Can J Public Health. 1990;81:182–186. [PubMed] [Google Scholar]
38. Chung RC, Kagawa-Singer M. Predictors of psychological distress among Southeast Asian refugees. Soc Sci Med. 1993;36:631–639. [PubMed] [Google Scholar]
39. Kinzie JD. The psychiatric effects of massive trauma on Cambodian refugees. In: Wilson JP, Harel Z, Kahana B, eds. Human Adaptation of Extreme Stress: From the Holocaust to Vietnam. New York, NY: Plenum Press; 1988:305–317.
40. Carlson EB, Rosser-Hogan R. Trauma experiences, posttraumatic stress, dissociation, and depression in Cambodian refugees. Am J Psychiatry. 1991;148:1548–1551. [PubMed] [Google Scholar]
41. Wing JK, Babor T, Brugha T, et al. SCAN: Schedules for Clinical Assessment in Neuropsychiatry. Arch Gen Psych. 1990;47:589–593. [PubMed] [Google Scholar]
42. Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition. Washington, DC; American Psychiatric Association; 1994.
43. International Classification of Diseases, 10th Revision. Geneva, Switzerland: World Health Organization; 1980.
44. Marsella AJ. Measurement issues. In: Gerrity E, Keane TM, Tuma F, eds. The Mental Health Consequences of Torture. New York, NY: Kluwer Academic/Plenum Publishers; 2001:277–290.
45. Mollica RF, Caspi-Yavin Y. Measuring torture and torture-related symptoms. Psychol Assess.1991;3:581–587. [Google Scholar]
46. Willis GB, Gonzalez A. Methodological issues in the use of survey questionnaires to assess the health effects of torture. J Nerv Ment Dis. 1998;186:283–289. [PubMed] [Google Scholar]
47. Hollifield M, Warner TD, Lian N, et al. Measuring trauma and health status in refugees: a critical review. JAMA. 2002;288:611–621. [PubMed] [Google Scholar]
48. Weathers FW, Litz BT, Herman DS, et al. The PTSD Checklist (PCL): reliability, validity, and diagnostic utility. Paper presented at: 9th Annual Meeting of the International Society for Traumatic Stress Studies; October 24–27, 1993; San Antonio, Tex.
49. Blanchard EB, Jones-Alexander J, Buckley TC, et al. Psychometric properties of the PTSD checklist (PCL). Behav Res Ther. 1996;34:669–673. [PubMed] [Google Scholar]
50. Hall ET. The Silent Language. Garden City, NY: Doubleday; 1959.
51. Brislin R. Back-translation for cross-cultural research. J Cross-Cultural Psychol. 1970;1:185–216. [Google Scholar]
52. Butcher JN, Garcia R. Cross-national application of psychological tests. Personnel Guidance.1978;56:472–475. [Google Scholar]
53. Hulin CL. A psychometric theory of evaluations of item and scale translations: fidelity across languages. J Cross-Cultural Psychol. 1987;18:115–142. [Google Scholar]
54. Sabin JE. Translating despair. Am J Psychiatry. 1975;132:197–199. [PubMed] [Google Scholar]
55. Kinzie JD, Manson SM, Vinh DT, et al. Development and validation of a Vietnamese language depression rating scale. Am J Psychiatry. 1982;139:1276–1281. [PubMed] [Google Scholar]
56. Robins LN, Wing J, Wittchen HU, et al. The Composite International Diagnostic Interview: an epidemiologic instrument suitable for use in conjunction with different diagnostic systems and in different cultures. Arch Gen Psychiatry. 1988;45:1069–1077. [PubMed] [Google Scholar]
57. Sartorius N. Making of a common language for psychiatry: development of the classification of mental, behavioral and developmental disorders in the 10th revision of the ICD. WPA Bull. 1989;1(1):3–6. [Google Scholar]
58. Bravo M, Woodbury-Farina FM, Canino GJ, et al. The Spanish translation and cultural adaptation of the Diagnostic Interview Schedule (DISC) in Puerto Rico. Cult Med Psychiatry. 1993;17:329–344. [PubMed] [Google Scholar]
59. Westermeyer J, Janca A. Language, culture, and psychopathology. Transcultural Psychiatry.1997;34:291–311. [Google Scholar]
60. Ebigo PO. Development of a culture-specific (Nigeria) screening scale of somatic complaints indicating psychiatric disturbance. Cult Med Psychiatry. 1982;6:29–43. [PubMed] [Google Scholar]
61. El Nasser H. Cities, minorities lose in census undercount. USA Today. 28March2001: A1, A6.
62. Waters JK, Biernaski P. Targeted sampling: options for the study of hidden populations. Soc Probl.1989;36:416–430. [Google Scholar]
63. Lepkowski JM. Sampling the Difficult-to-Sample. Nutritional Assessment and Intervention: Interface of Science and Policy. Washington, DC: Federation of American Societies for Experimental Biology; 1991.
64. Frank O, Snijders T. Estimating the size of hidden populations using snowball sampling. J Official Statistics. 1994;10(1):53–67. [Google Scholar]
65. Spring M, Westermeyer J, Halcon L, et al. Sampling in difficult to access refugee and immigrant communities. J Nerv Ment Dis. 2003:191: 813–819. [PubMed] [Google Scholar]
66. Convention Against Torture and Other Cruel, Inhuman, or Degrading Treatment or Punishment. New York, NY: Office of Public Information, United Nations; 1984.
67. UN Convention Against Torture and Other Cruel, Inhuman, or Degrading Treatment or Punishment: Methods of Combating Torture. Geneva, Switzerland: United Nations Centre for Human Rights; 1989:17.
68. Tellegen A. The analysis of consistency in personality assessment. J Pers. 1988;56:621–663. [Google Scholar]
69. Berry DTR, Adams JJ, Smith GT, et al. MMPI-2 clinical scales and 2-point code types: impact of varying levels of omitted items. Psychol Assess. 1997;9:158–160. [Google Scholar]
70. Hathaway SR, McKinley JC. The Minnesota Multiphasic Personality Schedule. Minneapolis, Minn: University of Minnesota Press; 1943.
71. Arbisi P, Ben-Porath YS. An MMPI-2 infrequency scale for use with psychopathological populations: The Infrequency-Psychopathology Scale, F (p). Psychol Assess. 1995;7:424–431. [Google Scholar]
72. Montgomery E, Foldspang A. Criterion-related validity of screening for exposure to torture. Dan Med Bull. 1994;41:500–591. [PubMed] [Google Scholar]
73. Exile Kunz E. and resettlement: refugee theory. Int Migr Rev. 1981;15:42–51. [Google Scholar]
74. Kinzie JD, Jaranson JM. Refugees and asylum-seekers. In: Gerrity E, Keane TM, Tuma F, eds. The Mental Health Consequences of Torture. New York, NY: Kluwer Academic/Plenum Publishers; 2001:111–120.
75. Jaranson JM. Government-sanctioned torture: status of the rehabilitation movement. Transcultural Psychiatr Res Rev. 1995;32:253–286. [Google Scholar]
76. Voices. New York, NY: Womens’ Commission for Women and Children; 1993.
77. Robertson C. Patterns of survival for Bosnian rural mothers during the war and post-war period. Minneapolis, Minn: School of Nursing, University of Minnesota; 2000.
78. Rehabilitation of Torture Victims: Centres and Programmes Worldwide. Copenhagen, Denmark: International Rehabilitation Council for Torture Victims; 2000.
Articles from American Journal of Public Health are provided here courtesy of American Public Health Association
No comments:
Post a Comment