Disability Associated With Psychiatric Comorbidity and Health Status in Bosnian Refugees Living in Croatia FREE

The recent war in Bosnia and Herzegovina that ended with the signing of the Dayton Peace Accord in 1995 has had a catastrophic effect on local populations residing in this region. For example, it is estimated that the war caused more than 250,000 deaths, created more than 2 million refugees and internally displaced persons, and wounded 200,000 in Bosnia and Herzegovina alone.^1- 2 Recent epidemiologic studies have revealed that the psychiatric morbidity associated with mass violence in civilian and refugee populations is elevated when compared with nontraumatized communities.³ In our previous report on Cambodian refugees living on the Thai border, approximately 68% of refugees displayed symptoms of major depression and 37% showed symptoms associated with the diagnosis of posttraumatic stress disorder (PTSD).⁴ These results were anticipated by clinical studies of Cambodian refugees resettled in the United States that revealed high rates of depression and PTSD.^5- 8 Currently, clinical reports of Bosnian refugees in treatment show similar findings for depressive symptoms with rates ranging from 14% to 21%^9- 10 and for PTSD symptoms with rates ranging from 18% to 53%.^9- 12 Because no epidemiologic studies of refugee communities from the Balkans have been reported to date, the degree of impact this recent war has had on the psychiatric morbidity of Bosnian refugees is unknown.

The relationship between psychiatric distress and disability in communities devastated by mass violence has received little attention, despite the widespread knowledge that mental illness is a leading cause of disability worldwide.¹³ In traditional populations, psychiatric disorders, especially depression, are associated with physical impairment,¹⁴ social role impairment,¹⁵ and loss of productivity.¹⁶ Cross-national investigations of primary health care patients have revealed that high rates of disability associated with common psychiatric disorders, such as depression, occur independently of the presence of medical illnesses.¹⁷ While these studies did not focus on nations in conflict, they suggest that the relationship between psychiatric illness and disability in postconflict societies should be at least as strong as in nontraumatized communities and developing nations.

In this study, conducted in February and March, 1996, we surveyed a population of Bosnian refugees living in a refugee camp in Croatia to assess the following: (1) degree of disability associated with trauma and other risk factors, (2) relationship between psychiatric symptoms (depression and PTSD) and disability, and (3) relationship between chronic medical illnesses and disability.

In 1992 the Croatian government established a number of refugee camps, including the Varaždin camp in northeastern Croatia. At approximately the same time, a private social service agency, Ruke, based in Croatia, began providing counseling and other social services to the Varaždin camp residents. In the fall of 1995, Ruke and the Harvard Program in Refugee Trauma conducted a key informant survey of 100 human service and Croatian government workers serving the Varaždin camp.¹⁸ That survey (available from the authors on request) provided the background data for the design and implementation of the current study.

We conducted a census of the 573 families, totaling 1275 people, living in the Varaždin camp as of February 1, 1996. One adult respondent 18 years of age or older was randomly selected from each family from a list of all adult family members. Of the target sample of 573 adults, 534 (93%) agreed to participate in the study. Nonrespondents were somewhat older (mean [SD] age, 58.6 [17.9] years) than respondents (50.4 [16.14] years; (t₅₆₅=− 2.83; P=.005) but did not differ significantly by sex (χ²₁=1.36; P=.24).

Interviews were conducted by Ruke's professional staff of local psychiatrists and by paraprofessionals who were Bosnian refugees living in the Varaždin camp and trained by Ruke staff members. The interviews were conducted in Bosnian and took an average of 90 minutes. Before the interviews, each respondent was read, in Bosnian, a letter of informed consent to sign that indicated the voluntary nature of participation and guaranteed strict confidentiality of responses. Informed consent procedures and study design were approved by the Human Subjects Committee of the Harvard School of Public Health. Interviewers received 4 days of training, including review of the questionnaire and practice interviews with refugees (Varaždin camp residents who met the criteria for selection but were not enrolled in the study sample). Quality assurance procedures assessed interrater reliability until there was 100% concordance among the interviewers.

The survey questionnaire was adapted for this study from the key informant survey.¹⁸ Translation was done by the Ruke team to construct the final questionnaire by means of the standard methods of cross-cultural research.^19- 20 The study's questionnaire focused on risk factors and outcomes associated with the respondent's refugee experience. Risk factors evaluated included demographics (sex, age, marital status, and education), length of time in the camp, family size, health (self-perception of health, handicaps, and chronic conditions), economic and social resources, work status, prisoner-of-war (POW) status, traumatic life experiences before and after leaving home, and torture experience. Outcomes included functional disability, the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) symptoms of major depression and PTSD, health status, and work. The anticipated repatriation needs of the respondents were also assessed.

The trauma history was derived from the Bosnian version of the Harvard Trauma Questionnaire (HTQ).²¹ Measures of cumulative trauma were constructed from responses to questions about trauma events, before and after leaving home (the refugee experience), related to the conflict. Affirmative responses were summed for 38 trauma events and 19 torture events. The torture experiences conformed to international definitions of torture.²²

Two interview schedules were used to measure psychological symptoms: a 25-item section of the Hopkins Symptom Checklist (HSCL-25)²³ and a 30-item section of the HTQ.²¹ The HSCL-25 includes a 15-item scale of depressive symptoms, while the HTQ contains a scale made up of 16 of the 17 diagnostic criteria for PTSD in DSM-III-R and DSM-IV. For both schedules, respondents were asked whether they were bothered by each symptom "not at all" (1), "a little" (2), "quite a bit" (3), or "extremely" (4) in the past week. In our sample, Cronbach α,²⁴ a measure of internal-consistency reliability, was estimated at .89 for both the depression scale and the PTSD scales.

Both instruments have been widely translated and used in a number of studies among diverse cultural groups^25- 30 and validated against clinical diagnoses.^{21,23,31- 32} Since the cutoff points for psychiatric diagnosis have not yet been established in a Bosnian refugee population, an algorithm method was selected that replicated the DSM-IV criteria for major depression (HSCL-25) and PTSD (HTQ).^33- 34 The algorithm method correlated well with standard cutoff points for the HSCL-25 and HTQ for PTSD and depression, respectively (results available from authors on request), and was more conservative in classifying respondents as symptomatic.

Functional disability was measured using the physical functioning scale of the Medical Outcomes Study 20-Item Short-Form Health Survey (MOS SF-20), which has been tested for its reliability and validity in a large number of disease settings, countries, and languages.^35- 37 Our survey questionnaire also contained other measures of disability, including a World Health Organization (WHO) physical functioning scale,³⁸ a measure of perceived energy level,²⁵ and a measure of self-perceived health.³⁹ The physical functioning scale of the MOS SF-20 correlated highly with the WHO scale (r=0.78, P<.001), and moderately high with energy level (r=− 0.62, P<.001) and self-perceived health (r=−0.58, P<.001). Cronbach α was estimated at .80 for the 6-item physical functioning scale of the MOS SF-20.

Since cutoff points have not been established for this population for the HSCL-25 and HTQ PTSD scales, another method of establishing caseness was used.⁴⁰ To be classified as symptomatic for depression a respondent required a positive response (score of 3 or 4) on either depressed mood or diminished interest/pleasure, and at least 4 of the following 6 DSM-IV criterion A symptoms: significant weight loss or change in appetite, insomnia or hypersomnia, fatigue or loss of energy, feeling of worthlessness, diminished ability to think or concentrate, and recurrent thoughts of death. The criterion A symptom, "observable psychomotor agitation or retardation" was omitted because of the inability to conduct physical examinations in this study.

For PTSD the algorithm for caseness included a positive response (3 or 4 on the HTQ) on at least 1 of the 4 reexperiencing symptoms (criterion B), at least 3 of the 7 avoidance and numbing symptoms (criterion C), and at least 2 of the 5 arousal symptoms (criterion D). Criterion A (exposure to traumatic event) was deemed to have been met by all respondents due to their flight from home and refugee status.

The physical functioning scale of the MOS SF-20³⁶ was used to measure disability. The 6 items measure limitations in physical functioning due to health and are measured on a 3-point scale (1, limited for more than 3 months; 2, limited for 3 months or less; and 3, not limited at all). The item responses were summed and then transformed linearly to a 0 through 100 scale with 0 indicating the poorest physical functioning and 100 indicating no limitations in physical functioning.³⁶

For the logistic regression analysis it was necessary to create a dichotomous variable from the 100-point functional impairment scale. The cutoff point of 50 was selected as a conservative estimate of functional disability. Stewart et al,³⁶ in a study of a US population using the same MOS SF-20, defined poor physical functioning as being limited on 1 or more items on the 6-item scale, which is equal to a score of 91.7 on the 100-point scale. In our sample, this same cutoff point (91.7) would result in 65% of respondents being classified as functionally impaired; we chose instead a more stringent cutoff point of 50 in order to examine the highly physically impaired individuals who, by most professional and clinical standards, unequivocally would be classified as functionally impaired or disabled.

Bivariate relationships between disability and characteristics of interest were examined using logistic regression. These analyses produced relative odds, that is, the likelihood of being disabled given the presence of a certain characteristic compared with not having the characteristic. Later, multivariate logistic regression was performed to produce odds ratios (ORs) that control for the presence of all the other variables in the model. An estimated pseudo-R² statistic was calculated to assess the degree to which the hypothesized model fit the observed data. We estimated F ratios and χ² values for categories of trauma. All analyses were performed with SAS statistical analysis software, version 6.11.⁴¹

The sample was 41% male and 59% female. The mean age was 50 years. Nearly half of the sample (48%) were married, with 22% widowed. A relatively large proportion of the sample was poorly educated, with 37% not having completed primary school. Self-identified ethnicity was 60% Bosnian Muslim, 23% Bosnian Croat, 11% Croat, and 6% other (Serb, Gypsy, and Bosnian unspecified).

The mean (SD) number of unduplicated trauma events reported by respondents was 6.5 (4.7). All but 31 respondents reported at least 1 event; 18% (95/534) reported experiencing 1 or more of the 19 torture events.

The 5 most frequently experienced trauma events were being present during shelling or grenade attacks (83%), hiding from snipers (75%), hiding outdoors (63%), being confined to home (51%), and being present while one's home was searched (37%). Other frequently reported traumas were being threatened or humiliated (34%), witnessing sexual abuse (33%), lacking food or water (28%), having a family member injured from violence (28%), witnessing violence to family member (28%), and being ill but having no access to medical care (23%). Roughly 17% of the sample reported the death, due to violence, of a family member: death of a child accounted for 3.4% of such deaths and other family members, 13.1%.

The most commonly reported torture events were forced standing (58%); being placed in sack, box, or other small space (36%); mock execution (34%); forced witnessing of torture (31%); and blows to the ears (28%).

Table 1 provides the results of the logistic regression models, with disability as the outcome variable and the demographic and other personal characteristics as the uncontrolled (univariate) predictors. In this sample, the scores on the MOS SF-20 ranged from 0 to 100, with a mean (SD) of 74 (26.6) and a median of 83. When we set a cutoff point of 50 on the physical functioning scale of the MOS SF-20, 25.5% of respondents were classified as disabled. As is shown in Table 1, sex, length of camp time, cumulative trauma, POW status, and experience of torture were not associated with increased risk for disability. Older age, being widowed or never married, lack of education, physical handicaps, and having 2 or more chronic conditions were associated with risk for disability. Having 3 or more family members in camp, earning an income while in camp, and reporting good to excellent health status were associated with a decreased risk for disability.

In Table 2 the odds of being disabled according to psychiatric status is assessed. In this sample 39.2% of respondents reported symptoms of depression (18.6% reported depression alone) and 26.3% reported symptoms of PTSD (5.6% reported PTSD alone); 20.6% reported symptoms comorbid for depression and PTSD. The odds for disability increased progressively with the addition of psychiatric symptoms. Respondents who met the criteria for depressive symptoms were 3.75 times more likely to report disability than asymptomatic respondents (95% confidence interval [CI], 2.23-6.31). In this analysis, the association between disability and PTSD alone became insignificant. Those who reported symptoms for both depression and PTSD were 5 times more likely to report disability than those who reported no psychiatric symptoms (95% CI, 3.05-8.26).

Table 3 presents the results of the multivariate logistic regression, controlling for sex, marital status, education, length of time in refugee camp, number of family members in camp, income earning in camp, POW status, medical conditions, and torture. In this analysis, the association between disability and depression became insignificant. However, the association between disability and symptoms comorbid for both depression and PTSD symptoms remained (OR=2.06; 95% CI, 1.10-3.86).

Analysis of cumulative trauma revealed that at 3 to 5 traumatic events the probability of disability was 2.36 times the probability for respondents with 0 to 2 events (95% CI, 1.14-4.86). Surprisingly, the fitted odds of impairment for persons with 6 or more events is not statistically different than the odds of impairment for persons with 0 to 2 events (P=.88).

Perceived health, physical handicap, and age continued to be important variables in explaining the variation in disability, although the magnitude of the relationship was no longer as great as in the previous univariate analyses. Persons aged 65 years and older were 3.6 times as likely to be disabled as young adults aged 18 to 34 years (95% CI, 1.03-12.53).

In this study, 94% of refugee respondents reported having been exposed to war-related violence with a mean (SD) experience of 6.5 (4.7) events. The most common trauma events were warfare and refugee experiences, including shelling and grenade explosions, exposure to sniper fire, and hiding. Torture was experienced by 18% of respondents. Even with conservative measurement, almost 40% of respondents reported symptoms of depression and 26% reported symptoms of PTSD. Of the refugee respondents, 25.5% reported having a significant disability.

Our data suggest that the following risk factors are associated with disability in this population: psychiatric symptoms, trauma, age, handicaps, and perceived health status. The impact of trauma, such as torture^22,45- 46 and the refugee experience,^26,47- 48 on psychiatric symptoms has received considerable attention by previous investigators. While cumulative trauma was initially associated with disability in this population, after 5 trauma events the association was no longer significant. Our measure of trauma is based on an unweighted self-report of having experienced a possible range of trauma and/or torture events. Since the psychological and biological responses to specific types of traumas experienced by refugees are poorly understood, our instruments have avoided a priori weightings of individual events. In previous studies of Cambodian refugees, our data have revealed significant dose-effect relationships between cumulative trauma and psychiatric symptoms.^4,47

The apparent lack of a consistent dose-effect response in the present study might be due to the higher cumulative trauma experienced by younger soldiers who, through their training, were better prepared to cope with traumatic events.⁴⁹ Therefore, we hypothesize that respondents with 6 or more trauma events may be different from less traumatized respondents in ways that are not captured by other control variables. Specifically, they may represent a disproportionate number of combatants. While direct inquiry about combat status was considered too sensitive for this sample of Bosnians living in the territory of Croatia, a country that had been at war with Bosnia and Herzegovina approximately 12 months before the interviews, an analysis of the characteristics of respondents in the different trauma categories was conducted. It indicated that those with 6 or more events were more likely to be male, younger, have POW experience, and have been tortured than their counterparts reporting fewer trauma events). This suggests that persons with 6 or more events disproportionately represent combatants. Military training and the belief that one is protecting country and family may have a protective effect, so that even though the number of traumas may be higher, the psychiatric symptoms are not more elevated than in less traumatized refugee respondents.^45,50- 51

Demographic characteristics of the disabled indicate that age is associated with disability in this population.⁵² In previous refugee investigations, the relative youth of the refugee population, especially in developing nations, has not provided an opportunity for the important association of age to be demonstrated.²⁵ In the Bosnian context, hundreds of thousands of elderly have been displaced. These results raise questions about the physical and psychological abilities of older people to adapt to their ongoing traumatizing life situation after repatriation.^11,52 The association between disability and being widowed or never married,¹² lack of education, and absence of family members is consistent with previous studies revealing the lack of resiliency of these groups when coping with extreme environments.^53- 54 The negative impact of poor physical health status and chronic medical conditions, including handicap, on disability is evident. While there appears to be a beneficial effect on disability status associated with economic activity, the actual value of working is unknown since it is confounded with other variables such as age and education.

This study documents high comorbidity of the 2 major psychiatric disorders investigated in this Bosnian refugee population, major depression and PTSD. This finding has been well established in clinical reports of resettled refugee patients,^6,8 other traumatized special clinical populations such as POWs,⁵⁵ in community studies of PTSD in the general population,^56- 58 and in veteran groups.^59- 60 Our results support the ubiquitousness of comorbid psychiatric illness in traumatized refugees. This conclusion led us to explore the hypothesis that psychiatric comorbidity was associated with higher levels of disability than that found in asymptomatic refugees or those with single diagnoses. Our analyses revealed that disability levels were markedly increased among Bosnian refugees comorbid with symptoms of PTSD and depression. Furthermore, our analyses reveal that psychiatric illness is not an epiphenomenon of medical illness in relationship to disability. Our multivariate analysis reveals that self-reported symptoms of PTSD and depression are associated with disability independent of poor physical health. These results in a refugee population are similar to the results of the WHO collaborative study of primary health care patients in nonconflict societies.¹⁷

A number of limitations need to be noted. The classification of respondents into diagnostic categories was affected by the lack of validation with clinical diagnoses that has been achieved with the screening instruments in other refugee populations.^31,50 Because neither clinical interviews nor structured clinical interviews were included in the study it is unclear to what extent the rates of self-reported symptoms of PTSD and depression cited here would match clinical diagnoses. For this reason, rather than using a predetermined cutoff point for the instruments, a multidimensional classification algorithm, based on DSM-IV, was used. Our follow-up study, now in progress, includes structured clinical interview for the DSM-IV diagnoses to validate cutoff points and case definitions.

Currently there exists no criterion standard to validate disability measures in refugee populations. For example, we do not know the baseline functional status of Bosnia and Herzegovina citizens before the war, and we do not know how functioning changes as individuals adapt to their new refugee situation. Because of these limitations we used multiple measures of disability, which, in our study, were shown to be highly intercorrelated. We also selected a conservative cutoff point on the physical functioning scale of the MOS SF-20 relative to US-based norms established by the Medical Outcomes Study.⁶¹ What remains undetermined, however, is how many more respondents would have been classified as disabled if less stringent criteria were used based on a criterion standard or structured clinical rating scale.

The majority of the variables are based on self-reported data such as measures of trauma, physical functioning, and health status. Physical examinations were not conducted. The only health-related measure that could be independently verified was physical handicap status. The accuracy of reporting of trauma events by refugees has been extensively reviewed in previous reports.^27,62 The validity and reliability of the disability measure used in our analyses have been evaluated against objective assessments supporting the feasibility of using self-report disability scales in refugee populations such as the one we describe.³⁷

Furthermore, the 3 disability measures used in this study are not highly correlated with psychiatric symptoms, suggesting that disability and psychiatric symptoms are separate constructs. Von Korff et al³⁷ arrived at similar conclusions while investigating the association of the MOS survey short form disability questions to depressive symptoms as measured by the General Health Questionnaire.⁶³

Finally, we are in the process of conducting a prospective study of this sample of Bosnian refugees returning home because cross-sectional data cannot show how risk factors such as torture, trauma events, and demographic status (eg, age) interact with psychiatric symptoms and disability, respectively, over time.