Psychosocial Problems of War-Affected Youth in Northern Uganda: A Qualitative Study

Abstract

Multiple studies have found that children from a variety of cultures who have been affected by war are at increased risk for a range of psychosocial problems. However, most studies are based on Western concepts and assume that these are locally applicable. Very few have investigated how psychosocial problems are perceived by the affected communities, families and the young people themselves. Understanding local perceptions is important to ensuring that local priorities are addressed, and addressed in ways that are likely to be acceptable and effective. In this study we used a previously developed rapid ethnographic assessment method to explore local perceptions of psychosocial problems among issues among children and adults from the Acholi ethnic group that have been displaced by the war in northern Uganda. We conducted 45 free list interviews (N= 30 10-17 year olds, N=15 adults) and 57 key informant interviews (N=32 10-17 year olds, N=25 adults) Our purpose was twofold: 1) To test whether this rapid ethnographic assessment approach previously used among adults would be suitable for use with children; and 2) To use this approach to understand the psychosocial problems affecting local children from their own viewpoint and that of their caretakers, in order to inform subsequent assessment and intervention efforts by organizations serving this population. The rapid assessment approach appears to have worked well for interviewing caretakers and children aged 10-17 years. Several locally defined syndromes were described: two tam/par/kumu (depression and dysthymia-like syndromes), ma lwor (an anxiety-like syndrome), and a category of conduct problems referred to as kwo maraco/gin lugero. The descriptions of these local syndromes are similar to Western mood, anxiety and conduct disorders, but contain some culture-specific elements.

Introduction

The war in northern Uganda is one of the most persistent and deadliest Complex Humanitarian Emergencies (CHEs) in the world today (Relief Web, 2006). Over 1.8 million people, most of them Acholi, have been internally displaced during the more than 20 year conflict. The camps housing internally displaced people (IDPs) in northern Uganda are extremely overcrowded and characterized by constant insecurity and risks to health (ibid.). It is estimated that more than 25,000 children have been abducted to serve in the Lord's Resistance Army (LRA) (OCHA, 2006). Insecurity is such that thousands of child “night commuters” from outlying villages move into the larger towns each evening in order to avoid night time raids and abductions in their home villages.

Research on children affected by armed conflict and displacement, such as that in northern Uganda, indicates increased risk for a range of mental health problems (Barenbaum, 2004; Lustig et. al. 2004; Stichick, 2001). In particular, research has emphasized the many ways in which exposure to war-related traumatic events are likely to contribute to subsequent mental distress, and in some cases, longer-term psychopathology (ibid). A number of studies have found anxiety disorders, particularly Posttraumatic Stress Disorder (PTSD), occurring in very high rates in war-affected children (Kinzie et al. 1986; Weine 1995; Sack, 1999). Prior research in northern Uganda has indicated that war-affected youth exhibit both clinically-significant symptoms of PTSD (Derluyn, et. al. 2004) as well as problems related to depression, anxiety and hostility (MacMullin & Loughry, 2004). However, the Derluyn (2004) study has been heavily criticized by other mental health professionals for both its imposition of the Western diagnosis of post-traumatic stress disorder as the starting point for understanding these youth, as well as concern as to whether asking participants to recall atrocities was ethically justified given that such traumatic exposures in northern Uganda have been well-documented (McKay & Wessels, 2004). In contrast, although an ongoing survey of war-affected youth in eight subcounties of northern Uganda indicated that young men aged 14-30 had experienced, on average, nine serious violence exposures, participants in the survey also reported lower than expected levels of emotional distress using adapted Western mental health measures (SWAY, 2006). One quarter of the young men interviewed reported experiencing moderate to high emotional distress on a modified version of the Harvard Trauma Questionnaire. Educational and occupational functioning were unrelated to levels of distress. A majority of this sample (90%, N=791) also reported high social functioning and low levels of aggression.

At this time it is not clear why studies of war-affected children show varying results. It may be partly due to conducting assessments at different time periods during or after a conflict, to variability in types of traumatic exposure or to the influence of culture on how children react. There may also be methodological reasons for the differences. Most studies use quantitative instruments based on Western mental health concepts such as those defined in the Diagnostic and Statistical Manual of the American Psychiatric Association/ DSM-IV-TR (APA, 2000) or the International Classification of Diseases/ICD-10 (World Health Organization, 2006). These choices are usually based on prior assumptions as to what the major mental health problems will be and how they will manifest locally. Yet, as indicated in the criticisms of the Derylun et al. study (2004), it is not always known how appropriate such Western concepts and diagnostic criteria are in these settings. It may be that concepts central to diagnosing disorders such as PTSD and depressive disorders in American and European psychiatry (and by extension, instruments based on these concepts) are not easily compatible with local concepts.

The purpose of this qualitative study was both to test a method of rapid ethnographic assessment designed to explore local mental health concepts, and to use this method to explore the local applicability of Western psychosocial concepts among a population of Acholi war-affected children. In this way we wanted to determine, based on data, whether Western concepts were both applicable and important to the local situation rather than just assuming that this was so. Our study question was “What do children and their caretakers consider to be the most important psychosocial problems affecting Acholi 10-17 year olds in IDP camps and how do they understand these problems?” Our investigation included exploring the local names of these problems, how they manifest, what are the causes and what is commonly done to address them. Comparison of the resulting data with Western concepts of mental health problems and good functioning was later used both to select and create appropriate instruments for later quantitative assessments of these problems and of related dysfunction. The approach is the same as one previously used among adult populations (Bolton, 2001). However, this study is the first time we have used this approach among children and their caretakers.

The study formed the first step in a collaborative process between Boston University (where the first and last authors were based at the time) and World Vision (WV) and War Child Holland (WCH) which uses both qualitative and quantitative assessment methods to improve program design, monitoring, and evaluation. A report of the entire process (which has since been completed) is available from the authors on request. Both WV and WCH are organizations implementing assistance programs in northern Uganda, including the sites of this study.

Methods

Procedures

Interviews were conducted in July and August of 2004. As in previous assessments among adults, two qualitative interviewing methods were used: Free Listing (FL) followed by Key Informant (KI) Interviewing. Interviews were conducted in the Acholi Luo language by 10 local Luo-speaking Acholi interviewers and two non-Acholi non-Luo speakers who worked with the assistance of Luo-speaking translators. Interviewers also spoke English and were trained and supervised by the authors. Interviewing was done in pairs with one person serving as the lead interviewer and the second person serving as primary note taker and “quality control”. The latter role included feedback to the lead interviewer as to their rapport with the participant, their use of non-leading interviewing methods, and their use of probes to fully explore the locally-described psychosocial constructs. Members of each team could alternate their roles or not, according to their preference. Despite initial concerns, interviewers reported that the presence of two interviewers did not appear to affect the openness of interviewee responses.

Problem and Child Task/Roles Free Listing Exercise

Free List (FL) participants (N=15 adults and N=31 10-17 year olds) were asked two questions, each of which was designed to elicit responses in the form of a list. All responses were recorded verbatim on a standard FL record sheet in the respondent's own language. The first FL was based on the question “What are the problems of children in this camp?” Interviewers probed for as many problems as possible and also asked for a short description of each problem. Interviewers recorded in the respondents' own words the names of problems and the short description given to each one. At the end of the FL interview (but before the respondent had left), interviewers reviewed the problem names and descriptions for reference to problems of thinking, feeling, or relationships. These were regarded as psychosocial issues. Interviewers were instructed to ask respondents for (and record) the names and contact information for community members who were particularly knowledgeable about each of these psychosocial problems either because of their connections within the community or because people consulted them about the problem. Here ‘community member’ excluded persons working in the community but not from the community, such as health workers or other persons with outside training. These knowledgeable persons formed the initial group of informants for the second qualitative method.

The second FL used the question “What do boys/girls do?” This question explored the routine activities and tasks that children engage in. As with the “problems” free list, interviewers probed for as many activities/tasks as possible and recorded short descriptions of each one. Girls were asked only about girls and boys only about boys. Caretakers were questioned according to the gender of their child. The resulting data are not presented in this article but were used to form a scale of functional limitations which was used in later phases of this research.

On completion of the free list interviews, the authors reviewed the results in collaboration with World Vision local staff to select which psychosocial problems would be explored in depth. (While War Child supported the study their staff were not involved in the study). This choice is based on which issues (and how many of them) the implementing organization is able to address, and is part of the rapid assessment methodology: Rapid assessment is achieved by focusing not on all the psychosocial problems that are described in the free lists but only on those frequently mentioned problems that are likely to be addressed by the implementer. For example, epilepsy was frequently mentioned but not explored because World Vision did not have the resources to provide effective drug therapy. Based on these discussions with World Vision we agreed to identify no more than five or six related problems which could be addressed by one or two interventions. On this basis, five local syndromes were selected for more detailed investigation by key informant interview (below). The first three problems (two tam, kumu, and par) bear similarity to Western mood and depressive disorders. The fourth syndrome (ma lwor) shares similarities with criteria for anxiety disorders and a fifth local syndrome (kwo maraco/gin lugero) shares some similarities with Western conduct and oppositional defiant disorders, but describes behavioral problems of relevance for children and adolescents in the IDP camps studied.

Key Informant Interviews

Once we selected the psychosocial problems to explore, those persons said by FL informants to be knowledgeable about those problems were approached by the interviewers to be interviewed as Key Informants (KIs). These KIs (N=32 adults and N=25 10-17 year olds) were interviewed in-depth about the selected problems from the free lists. KI interviews began with either an open question such as “Tell me about the problem of… (referring to the selected problem),” or with a description of a hypothetical person with the selected problem(s). The KIs were probed on symptoms and signs, perceived causes, and what people do about the problem (if anything). KIs were also probed for additional psychosocial issues that were not mentioned on the FLs. When a cluster of signs and symptoms was described together, interviewers were instructed to ask “Is there a word for these sorts of problems or for someone who has a lot of these types of problems?” Unlike in the free list interviews, interviewers were instructed to interview the same KI repeatedly to the extent that the person was available and until no new information was presented.

Participants

Study participants were Acholi children (boys and girls) 10-17 years and adults (mostly caretakers) living in the Awer and Unyama camps of Gulu District, northern Uganda. These camps were chosen because they are near to Gulu (important in terms of security) and because World Vision and War Child Holland were developing programs for IDP children and adolescents in both sites. For the free lists we chose a purposive sample of children (and their caretakers) including participants from both camps, both genders, the age range from 10-17 years, those attending school and not attending school, those formerly abducted by the LRA and those never abducted. All children were required to have lived in the camp for at least 3 months. There were N=15 adults (80% female) and N=31 10-17 year olds (47% female) who participated in FL interviews.

KI s were identified by the FL respondents as described above. There were N=32 adults (50% female) and N=25 10-17 year olds (52% female) who participated in KI interviews. KIs included young adults who were well-known in the IDP camps, parents or relatives of young children, traditional healers, and some teachers and health professionals. All spoke Acholi Luo and lived in the IDP camps. Exclusion criteria included persons with a severe cognitive or physical disability who were impaired and unable to answer questions in Luo. The study procedures were reviewed and approved by local camp leaders and the World Vision psychosocial program staff. The study design was approved by the Institutional Review Board (IRB) at Boston University and all interviewees provided informed consent. All staff were trained in research ethics and all precautions were taken to protect participant identities and minimize any adverse effects that might result from the interviewing process.

Data Analysis

At the end of each day of interviews, the interviewers met with World Vision staff and the authors to review the interview experiences and the interview notes, and to transcribe the Luo interview notes into English. This review process included identifying problems in the interviewing process and topics that emerged in the interviews that required further exploration. In such instances, interviewers were directed to return to conduct additional interviews with the FL or KI participants.

FL Data Analysis

Data from the FL interviews were first summarized into lists that included every response and the number of that response was mentioned. These lists were then reviewed by the authors and the interviewers for responses that were conceptually identical (even though the wording was different). Such items were combined into a single item and the number of responses summated. For example, if ‘hunger’ was identified by 10 respondents and ‘lack of food’ by another 5, this would be combined into a single item referred to either as ‘hunger’ or ‘lack of food’ and listed as being reported 15 times. These composite free lists are shown in Table 1.

Table 1.

Main Problem Themes Emerging from Free Listing Exercises (N=45)^*

Theme	Number reporting N (%)
Lack food	34 (74)
Lack clothing	31 (67)
Lack school fees, uniforms, books, etc.	30 (65)
Insecurity/fear of abduction	18 (39)
Diseases (sexually transmitted, due to poor hygiene, malaria)	14 (30)
Poor hygiene (latrines, bathing, soap etc)	13 (28)
Lack parents	9 (20)
Lack of safe housing/shelter	9 (20)
Males disrupting females/girls staying with soldiers/rape	9 (20)
Lack money (general)	8 (17)
Dropping out of school	7 (15)
Stubborn, don't listen to parents	7 (15)
Fighting	6 (13)
Rude or spoilt (children)	6 (13)
Lack of bedding	6 (13)
Getting married/having children young	5 (11)
Boredom, disinterest, idle	5 (11)
Children sleeping in town	4 (9)
Stealing	4 (9)
Lack of medicines	4 (9)
Lack clean drinking water	3 (7)
Disabled	3 (7)
Roaming camp	3 (7)
Fleeing rebels	3 (7)
Being sent away with relatives	3 (7)
Children who return after abduction are "weird"/difficult	3 (7)

^*Problems mentioned by three or more respondents.

KI Data Analysis

Analysis was limited to syndromes described by multiple KIs as being “common” in the population of Acholi IDP children and adolescents. The authors daily reviewed the description of each of the locally-described syndromes as they appeared in the interview records. In this way the team developed a composite description of each syndrome. In most cases, only symptoms agreed to by at least three or more KIs were retained in the final syndrome descriptions. In some cases, symptoms that represented an important theoretical or contextual issue were retained even if they did not receive multiple endorsements. For example, the symptom of not caring whether one lives or dies was only endorsed twice was retained for describing two tam given its clinical significance in this context. Several symptoms such as loss of appetite, feeling weak, and not sleeping were shared between one or more syndromes.

Results

Free List Data

The free listing exercise was conducted among 45 adult and child respondents. Only three adult males were interviewed because we emphasized interviewing child caretakers and few male caretakers were available during the day time when interviews were conducted. The summary of results (Table 1) reveal a wide range of psychosocial and other problems affecting children in the Gulu and Awer IDP camps. These problems ranged from lack of food, lack of clothing and lack of school fees to concern about the continued insecurity or fear of abduction, rape, children dropping out of school, fighting, being stubborn or disobedient and idleness/disinterest. Selected FL problems related to thinking, feeling or relationships (our functional definition of psychosocial problems) were used as the foci of the in-depth KI interviews.

Key Informant Interviews

Local Syndromes

Upon analysis of the free listing and KI interview data, an original list of seven local syndromes or problem clusters were identified. The symptoms and features of five of these locally-described syndromes are listed in Table 2. These include:

Table 2.

Signs and Symptoms for the 5 Local Mental Health Syndromes

Two tam	Kumu	Par	Kwo Maraco	Ma lwor
Lots of thoughts	Loss of appetite	Lots of thoughts	Fights	Clings to elders
Constant worries	Pain in the heart	Wants to be alone	Uses bad	Thinks has no future
Body pain	Sits with cheek in palm	Easily annoyed	language	Constantly running
Brain isn't functioning		Holds head	Is disrespectful	Does not like noise
Think self is of no use	Cries when alone	Loses concentration in class	Misbehaves	Fast heart rate
Thinks about suicide	Does not sleep at night	Drinks alcohol	Drinks alcohol	Fears being alone
Talks about problems		Thinks about suicide	Loses interest in school	Loss of appetite
Sits alone	Talks about problems	Doesn't greet people		Wants to be alone
Loses interest in school		Sits alone	Disinterested	Does not sleep at night
	Lies down all the time	Lots of worries	Deceitful
Headaches		Does not think straight	A rough person	Drinks alcohol
Feels sad	Has lots of worries	Cannot do anything to help themselves	Uses drugs	Doesn't greet people
Does not care if lives or dies	Headaches		Disobedient	Thinks people are chasing him/her
	Feels cold	Does not trust
Thinks of bad things	Weak	Mutters to self
Doesn't feel like talking	Does not feel like talking	Insults friends
		Disobedient
Forgetful	Disobedient	Weak
Weak		Cries continuously
Cries continuously

Two tam was described as a problem of having “lots of thoughts”. In Western diagnostic terms, it has features of mood disorders and anxiety. Two tam shares symptoms with the DSM IV definition of Dysthymia [300.4] including low self-esteem, poor concentration and feelings of hopelessness. It also includes some of the DSM IV criteria for Major Depressive Disorder, [296.3] - depressed mood (being sad or tearful), diminished interest in activities, fatigue, feelings of worthlessness or excessive guilt, diminished ability to think/concentrate, and recurrent thoughts of death or suicide, as well as somatic complaints without medical cause such as headaches and pain all over the body. The symptom of not feeling like talking may indicate social withdrawal or social impairment due to depression. Other symptoms of two tam - talking about problems, and anxiety symptoms including having lots of thoughts and constant worries, are more suggestive of related anxiety.

Kumu, like two tam, has features of mood disorder and anxiety. Local informants described it as a problem of experiencing extreme and persistent grief or sadness. Sitting with one's cheek in the palm and sitting alone were said to be highly characteristic of kumu. Both of these symptoms were frequently demonstrated by key informants to demonstrate what a child with kumu is like. DSM IV criteria for Major Depressive Disorder, [296.3] are reflected in the symptoms of depressed mood, crying, decreased appetite, insomnia and fatigue or loss of energy nearly every day. Somatic complaints were again observed as possible expressions of sadness including feeling cold, having headaches and feeling a lot of pain in the heart, as well as lying down all the time and feeling weak. Disobedience and not feeling like talking may reflect irritability which is a feature of depression among adolescents (DSM IV-TR, ibid.). As with two tam, kumu includes talking about having problems as well as anxious features like having lots of worries.

Par is a third local syndrome with mood and anxiety features. Local informants referred to par as a problem of having many worries. It shares a number of symptoms with the DSM IV description of Dysthymia [300.4], including depressed mood, poor concentration, and feelings of hopelessness. Some features of Major Depressive Disorder, [296.3] are present including diminished ability to concentrate as well as recurrent thoughts of death or suicide. There are a number of antisocial features of par that may be reflective of adolescent expression of sadness through irritability (i.e. disobedience, insulting friends, being easily annoyed) or through withdrawal and preoccupation (wanting to be alone, not greeting people, and muttering to oneself). The more anxious features of par include having lots of thoughts and lots of worries. Par has been identified in one other report on war-affected Acholi people in Pader district issued by MSF Holland (MSF, 2004b).

Ma Lwor is a local term for an anxiety-like problem of children and young people. Ma lwor shares symptoms with DSM IV-TR criteria for Generalized Anxiety Disorder [300.02] including sleep disturbance and excessive anxiety as well as symptoms of increased arousal and restlessness such as a fast heart rate, not liking noise, and constantly running around. Not liking noise, thinking people are chasing you, and clinging to elders may reflect hyper-arousal and re-experiencing, which are part of the Post Traumatic Stress Disorder Diagnosis/PTSD [309.81]. On the other hand, thinking that one has no future, wanting to be alone, loss of appetite, drinking alcohol, and not greeting people are mood disorder symptoms suggesting that Ma lwor, like two tam, kumu, and par, is an anxiety-depression disorder (but with mainly anxiety features).

Kwo maraco/gin lugero are two local terms used to describe young people who have a bad lifestyle (kwo maraco) or are “being rude” (gin lugero). Kwo maraco/gin lugero shares some symptoms with DSM IV-TR criteria for oppositional Defiant Disorder [313.81] including actively defying or refusing to comply with adult requests. Kwo maraco/gin lugero also shares some symptoms with Conduct Disorder [312.82] including violating major age-appropriate societal norms or rules including aggressive behavior, such as fighting, that causes or threatens physical harm to others. Further similarities with conduct disorder include deceitfulness, using bad language, drinking alcohol and using drugs (in this case marijuana and the local drug, jai). Being a rough person, disrespectful and being disinterested, particularly losing interest in school, all appeared as locally-distinct expressions of kwo maraco/gin lugero and represent violations of rules and social norms in this setting.

Two other syndromes were also identified. However, unlike the five syndromes above, these were not investigated in detail during the qualitative study. This was either because they did not appear to be common, appeared to reflect environmental stressors rather than true mental health disorders, or because World Vision and War Child did not have the resources to address them in a subsequent intervention.

The first of these was a locally-described psychosis called cen said to be caused by spirit possession. Cen includes symptoms of extreme aggressiveness, auditory and visual hallucinations and severely impaired reality testing. This disorder was described as being “haunted by the spirits of those you have killed” and referred in particular to former child combatants who had participated in killing. As such, cen was not considered to be a common disorder among the general IDP youth population. (This was later supported by an ongoing survey of war-affected males in other districts of northern Uganda in which 5% of the sample endorsed being afflicted by cen per self-report (SWAY, 2006)). At the time of this study, although residential care at the World Vision center for former child soldiers was an option, there was no available community treatment model or clinical expertise within World Vision to appropriately serve children identified with this type of violent or psychotic symptoms.

The second syndrome referred to fear of rebel attack and included anxiety and self-protection behaviors due to the ongoing state of insecurity and chronic threat of attack by LRA rebels in the IDP camps. This cluster of problems was not considered a mental health ‘disorder’ since such responses as “fear of leaving the camp to gather firewood” or “fear of rebel attack” are as likely to be normal responses to the abnormal circumstance of life in a war zone.

Existing Interventions

When asked what local people normally do to help children facing these sorts of problems, particularly two tam, kumu, par, and ma lwor, local respondents indicated that it is common to “sit with” a child individually or in groups to provide comfort and help a child talk about their problems. For the conduct-like problems of gin lugero/tic marac, there was less of a clear sense of what could be done. Multiple key informants indicated that the best solution was for service organizations like World Vision and War Child to come in and set up programs to help such children. These responses indicated little confidence in naturally-existing resources or practices for addressing conduct-related problems. These findings may also reflect the degree to which social and cultural structures that traditionally provided for the protection and guidance of children have become eroded in this setting of persistent displacement, insecurity and the despair and hopelessness that results.

Discussion

This study had two purposes. The first purpose was to test a qualitative research methodology for use among children. On reviewing the literature we could find no record of the use of the combination of methods described here among children and adolescents. We did find one study that used free listing in combination with focused group interviews to study attitudes about reproductive health in teenage boys in the United States (Marcell et. al, 2003), but we could not find studies that used free listing among children and adolescents in developing countries. Yet this approach appeared to work well among children aged 10-17 years of age in this particular population, as well as with their caretakers. The interviewing methods were well accepted by the respondents in that interviewers reported little difficulty in engaging respondents. Records of child interviews using this methodology were of comparable detail and length to those among adults; both the caretakers interviewed in this study and in previous studies of adults using the same approach (Bolton, 2001; Wilk and Bolton, 2002).

Our second purpose was to explore local concepts of mental illness. This study identified seven psychosocial syndromes among children aged 10-17 years living in displaced persons camps in northern Uganda. Of these, we explored five syndromes in detail because these five appeared to be common, severe, reflective of mental health problems (rather than appropriate responses to a difficult situation) and potentially treatable using resources available to existing service providers. The types of problems revealed are not surprising in this population of children exposed to displacement, loss due to war as well and the day-to-day adversity that characterizes life in the overcrowded IDP camps of Gulu District.. In Western terms, these five syndromes can be divided into two broad categories: anxiety/depression-like disorders (two tam, kumu and par which have more depression-like features, and ma lwor which has more anxiety features) and conduct problems (kwo maraco/gin lugero). Many of the symptoms of these local syndromes are identical to those of Western disorders of mood, anxiety and conduct problems. For mood disorders, symptoms shared with Western diagnoses included depressed mood, diminished interest in activities, fatigue, feelings of worthlessness or excessive guilt, inability to diminished ability to think/concentrate, and recurrent thoughts of death or suicide. For anxiety problems, shared symptoms included increased arousal and restlessness including having a fast heart rate and constantly running around. For conduct problems similar symptoms included aggressive behavior that causes or threatens physical harm to others such as fighting as well as associated symptoms of deceitfulness, using bad language, drinking alcohol and using drugs. Culturally specific symptoms also emerged. For example, “sitting kumu” (sitting while holding one's cheek in their hand) and not greeting people were described as symptoms of the locally derived mood disorder kumu. Not greeting people was also an important sign of both par and ma lwor. In the Acholi culture, to not extend a kind greeting to others you encounter is offensive and an important indicator of psychopathology. Such important local idioms of distress would not be captured without investigating these issues using qualitative methods. Yet these symptoms may well be important manifestations of mental health symptoms. In Western terms, cheek in palm is a likely idiom for sadness and perhaps lack of hope or energy, whereas not greeting people may reflect preoccupation or withdrawal.

Although the locally-described syndromes of two tam, kumu and par are similar, there are significant differences. For example, suicidal ideation or feeling that life is not worth living is a feature of two tam. Par contains the symptom of thinking of suicide, but has more antisocial features such as disobedience, drinking alcohol and not greeting people. Kumu in comparison consists of less severe items, bearing more similarity to dysthymia characterized by persistent depressed mood than to Major Depression.

While mood and anxiety problems emerged as important issues in the study, conduct problems were equally salient. Conduct problems were frequently mentioned in the free list interviews and included sexual violence or high risk of sexual activity (20%), dropping out of school, (15%) being stubborn (15%), rude or spoiled (13%) and fighting (13%). These conduct problems deserve attention in that they also represent risky behaviors. They should also be seen in the context of a larger social breakdown which was vividly described in the free list and key informant interviews: adults struggling with their own despair, loss and subsequent mental health problems as well as alcohol abuse and extreme poverty. As a result, young people in the camp are not monitored closely and there is a great deal of opportunity for young people to be involved in more negative survival strategies such as trading sex for food, shelter or protection as well as antisocial behavior representing the other side of this exploitative equation such as sexual violence.

Our intent subsequent to this study was to use the resulting data to develop and test instruments to quantitatively assess the problems identified here. While there are many similarities between the local syndromes that emerged from this study and Western mood, anxiety and conduct disorders, we also found enough differences in terms of how symptoms are expressed to suggest that simply translating existing Western instruments to assess emotional and behavioral problems in this population would be insufficient. As a result, we later used the qualitative data to develop and validate a locally-relevant scale of emotional and behavioral problems reflecting these five syndromes. This same instrument was eventually used to screen young people into a trial of mental health interventions for the problems that emerged from the qualitative study. A description of these subsequent activities, culminating in an effectiveness study, are available from the authors on request.

Limitations

The methods presented here involve the use of only two qualitative methods of data collection, despite a wide variety of methods available. Our intent in using only two methods was to develop a rapid approach that would be feasible for use by service providers in planning their programs: By limiting the approach to 2 methods we were able to complete training and data collection in 2 weeks. We have found in similar situations that the two methods we chose - free listing and key informant interviewing - provide a good balance in terms of an overview of issues and more detailed investigation of certain key issues. Under circumstances of increased time and resources, additional methods would provide increased data triangulation.

Our methods relied on hand-written field notes written in Acholi by the pairs of interviewers and later translated into English by these same interviewers. We do not use tape recorders because of the privacy concerns (the possibility of recognizing voices if the tapes are lost) and the length of time and resources it takes to transcribe interviews. While the use of interviewing pairs (who both take notes and combine these notes into a single record) is meant to improve accuracy, this is unlikely to reach the level of accuracy of transcribed recordings. Also, while the interviewers are bilingual, they are not professional translators. Therefore, the quality of their translation, even when working together, is likely to be less than that of a professional.

Finally, the decision to focus the key informant interviews only on the five problems that were likely to be addressed by the service provider constitutes a limitation in the data. For programmatic reasons we decided not to explore the local disorder cen. However, this is clearly a severe disorder that would warrant further research. Such information would be critical for informing additional models of intervention, particularly should current peace negotiations result in an end to the current situation of insecurity.

Conclusion

The research presented here demonstrates that a rapid ethnographic assessment method can be applied to describe and explore local perceptions of mental health problems among Acholi children displaced by war in northern Uganda. The approach proved suitable for use with children and adults and provided useful information on the psychosocial problems affecting Acholi children and adolescents in the IDP camps from a local perspective. Several locally defined syndromes similar to Western mood, anxiety and conduct disorders were identified along with a number of culture-specific symptoms. The research addressed an immediate mental health services need by providing contextual information on how local Acholi IDP youth and their caregivers think about mental health problems. The findings will inform subsequent assessment and intervention efforts to improve services for these war-affected youth.