Quality of life after postconflict displacement in Ethiopia ... - Springer Link

Soc Psychiatry Psychiatr Epidemiol (2011) 46:585–593 DOI 10.1007/s00127-010-0223-1

ORIGINAL PAPER

Quality of life after postconflict displacement in Ethiopia: comparing placement in a community setting with that in shelters Mesfin Araya • Jayanti Chotai • Ivan H. Komproe Joop T. V. M. de Jong

•

Received: 5 January 2008 / Accepted: 30 March 2010 / Published online: 11 April 2010 Ó Springer-Verlag 2010

Abstract Background The resilience of post-war displaced persons is not only influenced partly by the nature of premigration trauma, but also by postmigration psychosocial circumstances and living conditions. A lengthy civil war leading to Eritrea separating from Ethiopia and becoming an independent state in 1991 resulted in many displaced persons. Method A random sample of 749 displaced women living in the shelters in the Ethiopian capital Addis Ababa was compared with a random sample of 110 displaced women living in the community setting of Debre Zeit, 50 km away from Addis Ababa, regarding their quality of life, mental distress, sociodemographics, living conditions, perceived social support, and coping strategies, 6 years after displacement. Results Subjects from Debre Zeit reported significantly higher quality of life and better living conditions. However, mental distress did not differ significantly between the groups.

Also, Debre Zeit subjects contained a higher proportion born in Ethiopia, a higher proportion married, reported higher traumatic life events, employed more task-oriented coping, and perceived higher social support. Factors that accounted for the difference in quality of life between the shelters and Debre Zeit groups in three of the four quality of life domains of WHOQOL-BREF (physical health, psychological, environment), included protection from insects/rodents and other living conditions. However, to account for the difference in the fourth domain (social relationships), psychosocial factors also contributed significantly. Conclusion Placement and rehabilitation in a community setting seems better than in the shelters. If this possibility is not available, measures to improve specific living conditions in the shelters are likely to lead to a considerable increase in quality of life. Keywords Resilience Ethiopia Postmigration Shelters Community setting

M. Araya Department of Psychiatry, Addis Ababa University, P.O. Box 9986, Addis Ababa, Ethiopia M. Araya J. Chotai (&) Division of Psychiatry, Department of Clinical Sciences, University of Umea˚, 901 85 Umea˚, Sweden e-mail: [email protected] I. H. Komproe J. T. V. M. de Jong Department of Research and Development, HealthNet TPO, Amsterdam, The Netherlands J. T. V. M. de Jong Department of Psychiatry, Boston University School of Medicine, Boston, USA I. H. Komproe Faculty of Social and Behavioural Sciences, Utrecht University, Utrecht, The Netherlands

Introduction There are about 23.7 million internally displaced persons (IDPs) worldwide in 52 different countries today, a large proportion of them in Africa, living amidst war and persecution. They have little legal or physical protection and a very uncertain future, since they are not covered by international laws regarding refugees [35]. Although there are a few studies regarding the health status, quality of life, living conditions, and rehabilitation efforts concerning refugees who have been afforded asylum in developed countries, fewer such studies are available concerning IDPs still living in their generally low-income countries [2, 3, 9–11, 21, 22]. As internally

123

586

displaced, they often live in temporary dwellings or shelters which offer poor living conditions. Moreover, these persons have usually gone through severe traumatic life events that cause psychological distress and that are detrimental to their mental as well as physical health. On the whole, they usually have a relatively poor quality of life [3, 9, 21, 26]. The resilience of persons who have suffered severe trauma or disasters is obviously influenced by the nature of the premigration trauma experienced by them, but it is also influenced by the postmigration psychosocial circumstances and living conditions [5, 18, 25, 26]. Postmigration treatment received by the victims, and the organizational structure of the postmigration dwellings, play an important part in this. In a study on Vietnamese refugees resettled in the US, different premigration and postmigration factors predicted different psychological outcomes [5]. A study on Middle East refugees in Sweden found that the core symptoms of post-traumatic stress disorder (PTSD) were related to the trauma factors before arriving to Sweden whereas the common mental disorders in these refugees were related to resettlement stressors [18]. Similarly, for IDPs in Nepal, it was found that having experienced traumatic events were associated with PTSD, whereas social and demographic factors were associated with anxiety symptomatology [32]. Several psychosocial factors have been shown in the literature to speed up the recovery or resilience of posttrauma victims, including sense of belongingness, social recognition, perception of control over one’s own life, sense of predictability and safety in daily life, respect to personal dignity, and optimism regarding the future [25]. The risk factors for psychological distress in IDPs affected by violent conflicts are found to be somewhat different among members of the household with different household roles (mothers, fathers, children); a study in Indonesia found that psychological distress was highest among mothers and lowest among fathers, and that poverty-related risk factors were significant for mothers whereas long-term illness were significant for fathers [34]. Besides the psychosocial factors, also the living conditions in the resettlement situations are important [3]. A multisite epidemiological study for IDPs in Ethiopia, Algeria, Gaza, and Cambodia found that the living conditions related to adjustment disorders were bad shelter conditions, forced social isolation and lack of food [13]. One study on the victims of earthquakes in El Salvador found that when allocation to shelters was done with respect to the arriving family’s community of origin, it gave better rehabilitation outcome than when the families were allocated to shelters in order of their arrival [25]. Another extensive study has shown that the degree of improvement of refugee children in resettlement was not a

123

Soc Psychiatry Psychiatr Epidemiol (2011) 46:585–593

function of the quantity of services received from clinicians and mental health workers, likely due to the complexity of service delivery arising from a vast diversity in terms of ethnicity, language and age [6]. This suggests that the placement of the displaced persons in a community setting with a possibility of individual adaptation in the delivery of help would likely be better than placement in large-scale shelters. According to the estimates of the Ethiopian Disaster Prevention and Preparedness Commission, about 1 million persons were displaced in Ethiopia as a result of famine and war by the time of the establishment of an interim government in Eritrea in 1991 after a lengthy civil war. Thus, there are currently about 55,000 IDPs in the Ethiopian capital Addis Ababa who were forced to leave Eritrea. These people have suffered difficult traumatic experiences over the years. They had left everything behind, had lost friends and relatives, and had gone through exhaustion, lack of water, disease, or had been under cross fire [1–3]. A majority of these displaced persons ended up in shelters in the Ethiopian capital Addis Ababa. A minority of the displaced persons ended up in a small town 50 km southeast of Addis Ababa, called Debre Zeit. These latter persons were given varying types of smaller shelters within the local community, and the displaced individuals or families lived scattered in that town. We have previously reported detailed results regarding the internally displaced living in the shelters of Addis Ababa, but we have not reported any results earlier regarding the internally displaced who had ended up in the community setting of Debre Zeit. In the present study, therefore, we compare the group of displaced persons living in the shelters in Addis Ababa with the group living in the community setting of Debre Zeit, with respect to their quality of life and their mental health, as well as their coping strategies, social support and living conditions. Our goal is to identify the characteristics that differ between these two groups and also contribute to the observed differences between these two groups with regards to quality of life.

Methods Participants Ethics approval for this study was obtained from the Ethiopian Science and Technology Agency as well as from the Addis Ababa Disaster Preparedness and Prevention Bureau. The subjects of the present study were postconflict displaced persons after a lengthy civil war that led to Eritrea’s independent government in 1991. They had suffered


tremendous hardships during their travel to Addis Ababa (exhaustion, lack of food and water, disease, combat situation, torture, witnessed death of family members), where they arrived around 1991–1992. The average travelling time was estimated as 6.7 (±SD 8.4) months. This study was conducted during 1997–1998. A majority of these displaced persons in Ethiopia ended up in shelters in the capital Addis Ababa. When this study was commenced in 1997, around 70,000 lived in over 17 shelters, many of them in tents, 6–7 people sharing 4–5 m2 sized partitions. Some shelters (Kaliti) consisted of former warehouses and similar structures with mud-walls and tinroofs; up to 26 households had to share the four walls in these shelters. Other shelters (Kore) were made of small but detached (around 2 9 2 m) structures made of Bamboo stem walls cemented by mud, and the roofs covered by canvas and plastic. These participants who were living in the shelters have been described elsewhere [10, 11, 13] and have also been reported in detail [2, 3]. They comprise a random sample of totally 1,200 subjects (749 women) between the ages 18 and 60 years. A minority of the displaced persons ended up in a small town 50 km south east of Addis Ababa, called Debre Zeit. We have not reported any results earlier regarding these internally displaced who had ended up in Debre Zeit. Some of them had fled along the eastern route through Djibouti and decided to settle here before reaching Addis Ababa. Others had chosen Debre Zeit because they had acquaintances or distant relatives there. They were given varying types of smaller shelters within the local community, and the displaced individuals or families lived scattered in that town. A major goal of the collaborative research project was to compare the IDPs living in temporary shelters in four different countries: Ethiopia, Algeria, Gaza, and Cambodia [10, 11, 13]. However, as a minor goal, it was decided in Ethiopia to similarly select a sample from the internally displaced in the community setting of Debre Zeit, and the sample size was arbitrarily set as 10% of the sample size from the shelters. So the participants from the community setting of Debre Zeit comprised totally 120 subjects selected similarly, between the ages 18 and 60 years. It turned out that most (110) of the 120 subjects selected from Debre Zeit were women. Since earlier studies reported on the subjects from the shelters of Addis Ababa have found gender differences in various characteristics investigated [2, 3], it was decided to compare only women from the shelters with the women from Debre Zeit. All the 749 IDPs from the shelters and 110 IDPs from Debre Zeit who were approached agreed to participate, and their responses were anlaysed in univariate analyses. However, because of missing records on some of the variables, 15 respondents were excluded (12 subjects from

587

the shelters and 3 from Debre Zeit) from further analyses that involve multivariate binary logistic regression. Instruments As described elsewhere [2, 3], all the instruments were translated into the Ethiopian official language Amharic. All the interviewers were given 3 months training by accredited trainers from Europe, and the instruments were then pre-tested in a pilot study. The instruments were culturally validated and translated in a seven-step procedure described in other publications [10, 11, 13]. For each subject, information about the subject and the subject’s response to the instruments were collected during the same interview. Sociodemographics Sociodemographic information regarding the respondents included gender, age, country of birth, ethnic group, marital status, literacy (reading ability), and religious activity. Trauma Three measures captured childhood trauma, and one measure captured trauma related to displacement [2, 3]. Traumatic load due to family history of psychiatric illness was denoted by ‘‘mental problems in parents or siblings’’ (yes or no), assessed by whether or not a parent or a sibling had mental illness or a parent had alcohol problems. ‘‘Childhood maltreatment’’ was given by the total number of yesresponses (ranging from 0 to 13) obtained through the question ‘When you were growing up, did anyone in your household do some of the following things often to you?’, followed by a series of 13 statements (response yes or no to each), comprising ‘insulting you or swearing at you’, ‘threatened to hit’, ‘pushed, grabbed or shoved you’, and so on, including ‘choked you’ and ‘burned and scalded you’. ‘‘Traumatic childhood life events’’ were evaluated by the total number of affirmative responses (ranging from 0 to 16) among 16 questions derived from Harvard Trauma Questionnaire, Section I [20, 23], and were posed regarding the life period before 12 years of age. ‘‘Traumatic life events related to displacement’’ were evaluated by the same questionnaire, but posed regarding the period beginning 2 years before displacement and up to the time of arrival at the shelters. Mental distress Mental distress was evaluated by the self-report 90-item Symptom Check List (SCL-90-R), which is a general standardized measure of psychopathology [12]. It has been tested [24, 29] and employed in various cultural and

123

588

clinical settings including those concerning trauma victims [17, 27]. The symptom level of each item of the SCL-90-R is rated by the subject on a five-point scale of distress, from ‘‘not at all’’ (score 0) to ‘‘extremely’’ (score 4). The average of the scores of these 90 items, called the global severity index (GSI), indicates an overall degree of mental distress and is employed in this study [3]. The score digits of each item range from 0 to 4, but since a relatively large number of subjects will usually have a score 0 (not at all) for a given item, the mean scores for a large group often obtain values less than 1. Quality of life Quality of life was assessed by an instrument called the WHOQOL-BREF, developed and validated by WHO in several studies [30, 33]. It contains 26 items, the response (from least to most) to each item being on a 5-point rating scale of a particular aspect of quality of life. Besides the first two items of general nature, the remaining 24 items of the instrument are known to factorise into four domains of quality of life, denoted by ‘physical health’ (7 items, domain 1), ‘psychological’ (6 items, domain 2), ‘social relationships’ (3 items, domain 3), and ‘environment’ (8 items, domain 4), respectively [3]. Coping strategies Coping strategies were assessed by a list of ten items adapted for the survey and described elsewhere [2]. To each item, the respondent was asked to respond with ‘‘this is like me’’ or ‘‘this is not like me’’. The instrument roughly captures the three coping strategies often described in the literature as task-oriented, avoidance-oriented, and emotion-oriented coping strategies [14, 15, 28]. Perceived social support Perceived social support was captured in a series of eight statements, partly in line with the Social Provisions Scale [8], with the response ‘‘agree’’ or ‘‘disagree’’ to each statement [2]. These statements roughly capture the components reassurance of worth, reliable alliance, and guidance. The total number of ‘‘agree’’ among these eight items constituted the overall level of perceived social support. Self-reported indicators of living conditions The subjects were asked a number of questions with a ‘‘yes–no’’ response to each, that captured the living conditions of the shelters [3]. These included questions about whether the accommodation was too cramped, a bed/mat/ mattress was available, private facilities were available,

123


toilets or latrines were available, if there was sufficient food and water, if protection against animals and insects was provided, if the organizational support was enough, if the subject’s health was good or very good, or whether the subject had experienced loss of general benefits during the last year. Statistical analyses All the statistical analyses were performed using the SPSS software version 14.0 [31]. To assess the difference in a covariate (characteristic) between Debre Zeit and the shelters, univariate methods were used; chi-square tests were used for categorical covariates and t tests were used for continuous covariates. Significance level of 0.05 was used. Multivariate binary logistic regression allows one to predict a binary discrete outcome (dependent variable), such as group membership, from a set of variables (independent variables) that may be continuous, discrete, dichotomous, or a mix of any of these. Multivariate binary logistic regression was performed to compare the quality of life scores between the two groups, when controlling for specified groups of independent variables. Multivariate binary logistic regression with forward stepwise selection was then employed to identify the independent variables accounting for the difference in quality of life between the groups.

Results Tables 1, 2 and 3 show univariate comparisons between placement in Debre Zeit and placement in the shelters with regards to various characteristics. We find that compared to the shelters, the subjects from Debre Zeit contained a significantly higher proportion born in Ethiopia, a higher proportion married, reported higher traumatic life events, employed more task-oriented coping, and perceived higher social support. Also, the subjects from Debre Zeit had significantly better living conditions in all aspects except as regards reporting having good or very good health, where they did not differ significantly from the shelters. Mental distress, as assessed by the GSI, did not differ significantly between Debre Zeit and the shelters. However, the subjects of Debre Zeit reported significantly higher quality of life in all the four domains. For a specified quality of life domain, to investigate which of the characteristics that account for the difference in that domain between the groups Debre Zeit and shelters, we performed multivariate binary logistic regression with group membership as the dependent variable, and controlling for various groups of characteristics in turn, with the results shown in Table 4. For example, beta = 0.136 in

Soc Psychiatry Psychiatr Epidemiol (2011) 46:585–593 Table 1 Univariate comparisons between Debre Zeit and the shelters regarding sociodemographics and trauma background

589

Characteristics

Per cent with yes, or mean score Debre Zeit (n = 110)

Shelters (n = 749)

33.0 years

34.1 years

Sociodemographics Age Country of birth Eritrea

39.1%

50.8%

Ethiopia

60.9%

49.2%

Tigre

55.0%

57.6%

Amhara

37.6%

39.4%

Oromo

7.3%

2.9%

0.066

Marital status

0.000

Married Widowed

58.2% 16.4%

27.0% 25.5%

Separated

13.6%

28.2%

Divorced

1.8%

8.8%

10.0%

10.5%

Literacy

60.9%

54.7%

0.258

Attends religious services regularly

56.5%

56.1%

0.937

Mental problems in parents or siblings

8.2%

10.1%

0.610

Childhood maltreatment

0.71

0.31

0.087

Traumatic childhood life events

0.58

0.25

0.002

Traumatic life events related to displacement

3.39

2.57

0.001

Never married t tests were employed when comparing mean scores between Debre Zeit and shelters, and chisquare tests were used when comparing percentages between these groups. None of the cells in the chi-square tests had expected count less than 5 Table 2 Univariate comparisons between Debre Zeit and the shelters regarding coping strategies, social support and living conditions

0.245 0.024

Ethnic group

a

P valuea

Trauma background

Characteristics


P valuea

Shelters (n = 749)

Coping strategies Task-oriented coping

3.32

2.95

0.018

Avoidance-oriented coping

0.75

0.90

0.084

Emotion-oriented coping

1.14

1.10

0.705

Social support Perceived social support

a

t tests were employed when comparing mean scores between Debre Zeit and shelters, and chisquare tests were used when comparing percentages between these groups. None of the cells in the chi-square tests had expected count less than 5

3.83

2.73

0.000

Living conditions Is your accommodation too cramped?

30.9%

63.8%

0.000

Is there a bed/mat/mattress available?

72.7%

18.8%

0.000

Are there private facilities or private places?

35.5%

3.2%

0.000

Are there toilets/latrines?

64.5%

24.8%

0.000

Is there sufficient food/water?

64.5%

17.5%

0.000

Do you have protection against animals/insects?

78.2%

10.0%

0.000 0.000

Is the support from the organizations enough for you?

31.8%

10.7%

Is your health good or very good during your stay?

51.8%

55.7%

0.473

During the last year, have you experienced loss of general benefits?

54.5%

78.0%

0.000

the row ‘‘none’’ and column ‘‘domain 1’’ is obtained as the regression coefficient of ‘‘domain 1’’ in a regression analysis where domain 1 is the only independent variable

entered, and the binary dependent variable all through is group membership. Beta = 0.119 on the row ‘‘sociodemographics’’ and column ‘‘domain 1’’ is obtained as the

123

590


Table 3 Univariate comparisons between Debre Zeit and the shelters regarding mental distress and quality of life Characteristics


P valuea

Shelters (n = 749)

Mental distress GSI of the SCL-90-R

0.22

0.20

0.629

Quality of life Domain 1 (physical)

13.6

12.9

0.004

Domain 2 (psychological)

14.0

13.2

0.001

Domain 3 (social relations)

9.1

7.5

0.000

Domain 4 (environment)

9.5

8.3

0.000

a

t tests were employed when comparing mean scores between Debre Zeit and shelters, and chi-square tests were used when comparing percentages between these groups. None of the cells in the chi-square tests had expected count less than 5

regression coefficient of ‘‘domain 1’’ when the independent variables entered in multivariate binary logistic regression comprise ‘‘domain 1’’ as well as all of the 6 variables belonging to the group ‘‘sociodemographics’’ in Table 1. Beta = 0.184 on the row ‘‘trauma background’’ and column ‘‘domain 2’’ is obtained as the regression coefficient of ‘‘domain 2’’ when the independent variables entered in multivariate binary logistic regression comprise ‘‘domain 2’’ as well as all the 4 variables belonging to the group ‘‘trauma background’’ from Table 1, and so on. We see in Table 4 that controlling for all the characteristics except for the living conditions did not remove the statistical significance of better quality of life for the Debre Zeit subjects. However, after controlling for the living

conditions alone, there was no longer any significant difference between the groups in the three quality of life domain 1 (P = 0.565), domain 2 (P = 0.361), and domain 4 (P = 0.768), but the difference remained for domain 3 (P = 0.010). Furthermore, controlling for all the characteristics including the living conditions resulted in lack of significant difference in quality of life between the groups for all the quality of life domains. We performed detailed analyses to identify the individual characteristics that account for the difference between the groups in each domain of quality of life. For a given quality of life domain, we performed stepwise multivariate binary logistic regressions, employing group membership as the dependent variable all through as follows. Among the independent variables, we first entered this given quality of life domain, and then employed the (likelihood ratio) stepwise forward selection method to select from all the characteristics (excluding the remaining three quality of life domains). At each stage of the forward selection, we examined the significance level of the difference between the groups for this given quality of life domain. The results are given in Table 5. For example, beta = 0.058 on row ‘‘Do you have protection against animals/insects?’’ and column ‘‘domain 1’’ is obtained as the regression coefficient of ‘‘domain 1’’ when the independent variables entered are ‘‘domain 1’’ and the forward selected variable ‘‘Do you have protection against animals/ insects?’’. Beta = 0.180 on row ‘‘Are there private facilities or private places?’’ and column ‘‘domain 3’’ is obtained as the regression coefficient of ‘‘domain 3’’ when the independent variables entered are ‘‘domain 3’’ as well as the forward selected variables ‘‘Do you have protection

Table 4 Comparisons between Debre Zeit and the shelters for each the quality of life domain, when controlling in turn for various groups of characteristics by multivariate binary logistic regression Group of characteristics that is controlled fora

Regression coefficient of specified domain of Quality of life (WHOQOL-BREF) Domain 1 (physical) Beta

P value




Beta

P value

Beta

P value

Beta

P value

None

0.136

0.003

0.134

0.001

0.241

0.000

0.363

Sociodemographics

0.119

0.014

0.128

0.005

0.188

0.000

0.362

0.000

Trauma background

0.210

0.000

0.184

0.000

0.263

0.000

0.423

0.000

Coping strategies

0.158

0.001

0.153

0.001

0.255

0.000

0.388

0.000

Perceived social support

0.100

0.030

0.091

0.036

0.207

0.000

0.318

0.000

-0.040

0.565

-0.058

0.361

0.142

0.010

-0.026

0.768

Mental distress

0.189

0.000

0.134

0.001

0.252

0.000

0.400

0.000

All the above except living conditions

0.182

0.003

0.188

0.001

0.193

0.000

0.420

0.000

-0.009

0.922

-0.032

0.712

0.124

0.103

0.041

0.716

Living conditions

All the above including living conditions

0.000

Positive beta values indicate that Debre Zeit has higher quality of life, negative beta values indicate that the shelters have higher quality of life a

Tables 1, 2 and 3 show which individual characteristics that are included in each group

123


591

Table 5 Multivariate binary logistic regression comparing quality of life domains among Debre Zeit versus shelters, controlling for characteristics added by stepwise selection until the difference is no longer statistically significant Characteristics that are selected and cumulatively controlled for in forward stepwise selection

Regression coefficient of specified domain of Quality of life (WHOQOL-BREF) Domain 1 (physical) Beta

P value




Beta

Beta

Beta

P value

P value

P value

None

0.136

0.003

0.134

0.001

0.241

0.000

0.363

0.000

Do you have protection against animals/insects?

0.058

0.277

0.071

0.176

0.188

0.000

0.226

0.001

Are there private facilities or private places?

0.180

0.000

0.166

0.020

Is there a bed/mat/mattress available?

0.158

0.002

0.076

0.312

Traumatic life events related to displacement

0.185

0.001

Is there sufficient food/water?

0.152

0.006

Avoidance-oriented coping

0.171

0.003

Are there toilets/latrines?

0.160

0.006

Country of birth

0.159

0.007

Marital status Ethnic group

0.131 0.119

0.041 0.067

Positive beta values indicate that Debre Zeit has higher quality of life, negative beta values indicate that the shelters have higher quality of life

against animals/insects?’’ and ‘‘Are there private facilities or private places?’’; and so on. As reported in Table 5, we found that for domain 1 and domain 2, already the first item that was selected, namely ‘‘Do you have protection against animals/insects?’’, removed the significance of the difference between the groups for these domains. For domain 4, two further items of living conditions needed to be selected before the significant difference between the groups was removed for this domain. For domain 3, still 7 more items were needed to be selected (2 more regarding living conditions and 5 from the rest) before the significant difference between the groups was removed for this domain.

Discussion All the subjects of this study were women because most of the participants from Debre Zeit (110 out of 120) were women. Thus, we included only women from Debre Zeit and from the shelters since earlier studies have found gender differences in the various characteristics investigated [2, 3]. In the same vein, all the results of this study would not necessarily be valid if we were to compare men in Debre Zeit with men in the shelters. Our study shows that displaced women who were placed in the community setting of Debre Zeit had a better quality of life in all the four domains, when compared to those placed in the shelters of Addis Ababa. However, mental distress as measured by GSI of the SCL-90-R did not differ significantly between the two groups, although the subjects of Debre Zeit had reported

significantly higher trauma, and higher trauma has earlier been shown to be associated with higher mental distress [3]. This may be due to Debre Zeit subjects having significantly higher task-oriented coping. An earlier study has shown that higher task-oriented coping among women was associated with significantly lower mental distress, and that task-oriented coping was a moderator by attenuating the effect of trauma on mental distress [3]. An earlier study has also shown that the effect of trauma in reducing quality of life was to a large extent mediated through mental distress and only partly as a direct effect [3]. Since mental distress between the two groups in the present study did not differ significantly, the higher trauma reported by Debre Zeit subjects was not likely to have been a major influencing factor for the difference between the two groups as regards overall quality of life. Thus, the various items of living conditions, which were significantly different between Debre Zeit and the shelters, are expected to play a relatively important role in accounting for the difference in quality of life between the groups. This is what we actually found (Table 4), in that only when controlling also for the items of living conditions did the significant difference in quality of life domains between the two groups lose its significance. For three of the domains, living conditions alone could account for the difference between the groups. For domain 1 (physical) and for domain 2 (psychological) of quality of life, the difference between the groups was accounted for by the single living condition associated with protection against animals/insects, since the significant difference lost its significance when we controlled for this living condition (Table 5). The subjects in Debre Zeit were obviously more

123

592

protected from rodents like mice, and insects like cockroaches and mosquitoes, and this was an important factor for increasing the quality of life. Regarding the quality of life domain 4 (environment), the significant difference between Debre Zeit and the shelters lost its significance when two further living conditions were controlled for, namely higher availability of private facilities or private places, and having a bed/mat/ mattress, and these were available to a significantly larger extent to the Debre Zeit subjects. However, regarding the quality of life domain 3 (social relations), several further circumstances contributed to the higher scores obtained among the subjects from Debre Zeit compared to those among the shelters. These included, besides the above living conditions, items like traumatic life events, avoidance-oriented coping, country of birth, marital status, and the ethnic group the subject belonged to. This suggests that a partially direct effect of trauma in reducing quality of life found in an earlier study [3], was likely to be mainly within domain 3 (social relationships) of quality of life. In an earlier study, it was also found that higher perceived social support was associated with higher quality of life [3], but in the present study the observed significant difference in perceived social support between the two groups did not turn out to be a significant contributing factor to explain the difference in quality of life between the two groups. Since marital status, ethnic belonging and country of birth, however, were involved in explaining the difference in domain 3 (social relationships) between the groups (Table 5), it is likely that perceived social support is a proxy for marital status, ethnic belonging or country of birth. It is interesting that the living condition that constitutes protection from animals (rodents) and insects, turned out to be of major importance for quality of life. A study from refugee camps in West Africa showed that the infestation of rodents, and consequently the risk of Lassa fever, was highly increased for the residents of dwellings with poor quality housing and poor external hygiene [7]. Several studies have shown that the quality of the shelters (or dwellings) contributes significantly to the well-being of the internally displaced or the refugees [4]. In a country like Ethiopia, where malaria is endemic, measures that give protection from this disease are expected to be important. These measures could include mosquito proofing of night shelters which reduce the exposure to mosquito biting [19], or employing insecticide-treated plastic tarpaulins that kill high proportions of mosquitoes [16]. In summary, placement and rehabilitation of displaced persons in the context of a community setting thus seem to be better than in the shelters. It is likely that the placement of the displaced persons in a community setting offers

123


better possibilities for individual adaptation in the process of resilience compared to placement in large-scale shelters. Community settings in our study seem more conducive to task-oriented coping, higher perceived social support, and a favourable marital life. Moreover, community settings seem to offer much better living conditions, and these account for much of their significantly higher quality of life. If it is not possible to find placement opportunities in community settings, our study suggests that measures to improve specific living conditions in the shelters are likely to lead to a considerable increase in quality of life. These measures particularly include protection from animals (rodents) and insects, but also include accommodating for private facilities and offering a bed or a mattress to sleep on. Such measures may yield high benefits for relatively low costs and efforts. A limitation of our study is that the displaced persons were not randomly allocated between placement in the shelters of Addis Ababa or in the community of Debre Zeit. We do not fully know the selection processes that led some individuals to end up in Debre Zeit and others to end up in the shelters of Addis Ababa. Another limitation is that whereas the shelters of Addis Ababa were aggregated within specific areas, the housing of those in the community of Debre Zeit was scattered. This may have led to a difference in the representativeness of the samples from these two settings. The general procedure to select the samples was to use multistep random sampling procedures by first selecting the addresses of the temporal housing from the lists of resident addresses, and then one person was randomly selected from each selected address. In practice, we cannot exclude some systematic differences in the outcome of the selection procedures in these two settings. It cannot be excluded that the above two limitations may have influenced our results and conclusions to some extent.

References 1. Aptekar L, Giel R (2002) Walks in Kaliti, life in a destitute shelter for the displaced. In: de Jong J (ed) Trauma, war, and violence: public mental health in socio-cultural context. Kluwer Academic/Plenum Publishers, New York, pp 337–366 2. Araya M, Chotai J, Komproe IH, de Joop JTVM (2007) Gender differences in traumatic life events, coping strategies, perceived social support and sociodemographics among postconflict displaced persons in Ethiopia. Soc Psychiatry Psychiatr Epidemiol 42:307–315 3. Araya M, Chotai J, Komproe IH, de Joop JTVM (2007) Effect of trauma on quality of life as mediated by mental distress and moderated by coping and social support among postconflict displaced Ethiopians. Qual Life Res 16:915–927 4. Ashmore J, Babister E, Corsellis T, Fowler J, Kelman I, McRobie A, Manfield P, Spence R, Vitale A, Battilana R, Crawford K


5.

6.

7.

8.

9.

10.

11. 12.

13.

14. 15.

16.

17.

18.

19.

(2003) Diversity and adaptation of shelters in transitional settlements for IDPs in Afghanistan. Disasters 27:273–287 Birman D, Tran N (2008) Psychological distress and adjustment of Vietnamese refugees in the United States: association with preand postmigration factors. Am J Orthopsychiatry 78:109–120 Birman D, Beehler S, Harris EM, Everson ML, Batia K, Liautaud J, Frazier S, Atkins M, Blanton S, Buwalda J, Fogg L, Cappella E (2008) International family, adult, and child enhancement services (FACES): a community-based comprehensive services model for refugee children in resettlement. Am J Orthopsychiatry 78:121–132 Bonner PC, Schmidt WP, Belmain SR, Oshin B, Baglole D, Borchert M (2007) Poor housing quality increases risk of rodent infestation and Lassa fever in refugee camps of Sierra Leone. Am J Trop Med Hyg 77:169–175 Cutrona CE, Russell D (1987) The provisions of social relationships and adaptation to stress. In: Jones WH, Perlman D (eds) Advances in personal relationships, vol 1. JAI Press, Greenwich, CT, pp 37–68 de Jong JTVM (2002) Public mental health, traumatic stress and human rights violations in low-income countries: a culturally appropriate model in times of conflict, disaster and peace. In: de Jong J (ed) Trauma, war, and violence: public mental health in socio-cultural context. Kluwer Academic/Plenum Publishers, New York, pp 1–91 de Jong JT, Komproe IH, van Ommeren M, el Masri M, Araya M, Khaled N, van de Put W, Somasundaram D (2001) Lifetime events and posttraumatic stress disorder in 4 postconflict settings. JAMA 286:555–562 de Jong JT, Komproe IH, van Ommeren M (2003) Common mental disorders in postconflict settings. Lancet 361:2128–2130 Derogatis LR (1994) SCL-90-R: symptom checklist-90-R. Administration, scoring and procedures manual. National Computer Systems, Minneapolis, MN Dobricki M, Komproe IH, de Jong JTVM, Maercker A (2010) Adjustment disorders after severe life-events in four postconflict settings. Soc Psychiatry Psychiatr Epidemiol 45:39–46 Endler NS, Parker JDA (1990) Multidimensional assessment of coping: a critical evaluation. J Pers Soc Psychol 58:844–854 Endler NS, Parker JDA (1994) Assessment of multidimensional coping: task, emotion, and avoidance strategies. Psychol Assess 6:50–60 Graham K, Rehman H, Ahmad M, Kamal M, Khan I, Rowland M (2004) Tents pre-treated with insecticide for malaria control in refugee camps: an entomological evaluation. Malar J 3:25 Lev-Wiesel R, Amir M (2000) Posttraumatic stress disorder symptoms, psychological distress, personal resources, and quality of life in four groups of Holocaust child survivors. Fam Process 39:445–459 Lindencrona F, Ekblad S, Hauff E (2008) Mental health of recently resettled refugees from the Middle East in Sweden: the impact of pre-resettlement trauma, resettlement stress and capacity to handle stress. Soc Psychiatry Psychiatr Epidemiol 43:121–131 Medlock JM, Aryemo M, Bean J (2007) Impact of mosquito proofing of night shelters in refugee camps in Kitgum, northern Uganda. Trop Med Int Health 12:370–376

593 20. Mollica RF, Caspi-Yavin Y, Bollini P et al (1992) The Harvard Trauma Questionnaire—validating a cross-cultural instrument for measuring torture, trauma, and posttraumatic stress disorder in Indochinese refugees. J Nerv Ment Dis 180:111–116 21. Mollica RF, Cui X, McInnes K et al (2002) Science-based policy for psychosocial interventions in refugee camps: a Cambodian example. J Nerv Ment Dis 190:158–166 22. Mollica RF, Lopes Cardozo B, Osofsky HJ et al (2004) Mental health in complex emergencies. Lancet 364:2058–2067 23. Mollica F, McInnes K, Pham T et al (1998) The dose–effect relationships between torture and psychiatric symptoms in Vietnamese ex-political detainees and a comparison group. J Nerv Ment Dis 186:543–553 24. Olsen LR, Mortensen EL, Bech P (2006) Mental distress in the Danish general population. Acta Psychiatr Scand 113:477–484 25. Perez-Sales P, Cervellon P, Vazquez C et al (2005) Post-traumatic factors and resilience: the role of shelter management and survivors’ attitudes after the earthquakes in El Salvador (2001). J Community Appl Soc Psychol 15:368–382 26. Porter M, Haslam N (2005) Predisplacement and postdisplacement factors associated with mental health of refugees and internally displaced persons: a meta- analysis. JAMA 294:602– 612 27. Punama¨ki RL, Komproe I, Qouta S et al (2005) The deterioration and mobilization effects of trauma on social support: Childhood maltreatment and adulthood military violence in a Palestinian community sample. Child Abuse Negl 29:351–373 28. Rafnsson FD, Smari J, Windle M et al (2006) Factor structure and psychometric characteristics of the Icelandic version of the Coping Inventory for Stressful Situations (CISS). Personal Individ Differ 40:1247–1258 29. Schmitz N, Hartkamp N, Kiuse J et al (2000) The Symptom Check-List-90-R (SCL-90-R): a German study. Qual Life Res 9:185–193 30. Skevington SM, Lotfy M, O’Connell O (2004) The World Health Organization’s WHOQOL-BREF quality of life assessment: psychometric properties and results of the international field trial. A report from the WHOQOL Group. Qual Life Res 13:299–310 31. SPSS Inc (2005) SPSS Base 14.0 and AMOS 6.0 for Windows. Chicago, IL 32. Thapa SB, Hauff E (2005) Psychological distress among displaced persons during an armed conflict in Nepal. Soc Psychiatry Psychiatr Epidemiol 40:672–679 33. Trompenaars FJ, Masthoff ED, Van Heck GL et al (2005) Content validity, construct validity, and reliability of the WHOQOLBREF in a population of Dutch adult psychiatric outpatients. Qual Life Res 14:151–160 34. Turnip SS, Hauff E (2007) Household roles, poverty and psychological distress in internally displaced persons affected by violent conflicts in Indonesia. Soc Psychiatry Psychiatr Epidemiol 42:997–1004 35. UNHCR (2007) Internally displaced persons: questions and answers. Available at http://www.unhcr.org/basics/BASICS/ 405ef8c64.pdf. Retrieved January 2007

123