Assessing adolescent mental health in war-affected societies: the significance of symptoms

Abstract

Objective: To compare the use of self-report symptom checklists with qualitative methods for assessing adolescent psychological well-being in a war-affected society.

Method: A school-based sample of three hundred and thirty seven 13- to 15-year-olds from two communities on opposite sides of the Bosnian conflict (183 from Gorazde, 154 from Foca) completed the Hopkins Symptoms Checklist and the Harvard Trauma Questionnaire. A gender balanced sub-sample of 40 adolescents was selected on the basis of their combined checklist scores, including equal numbers of high and low scorers from each side. Over the following 6 months this sub-sample was assessed (blind to checklist scores) with qualitative methods that included narrative interviews of child and parent, and participant observation. School marks were taken as a measure of social function.

Results: Qualitative: Some children identified as “less well” by qualitative methods denied having symptoms. Some children identified as “well” had symptoms with no pathological significance for them. The lifeline revealed that feeling “less well” could be more related to post-war circumstances than war events. Quantitative: The two symptom checklist items have shown good internal consistency and discriminant validity. However, comparison with the overall well being revealed that still in 9/40 of cases the reported presence or absence of symptoms did not correspond to the well being of the child. Items of the two questionnaires did not discriminate reliably between children identified as “well” and “less well” by other means.

Conclusions: Self-report checklists may be useful as a public health measure to assess the prevalence of psychological distress in war affected areas, but they are not an adequate means of clinical screening. Checklists used in combination with other qualitative approaches make it possible to identify those in need and avoid unnecessary pathologizing.

Résumé

Objectif: Établir une comparaison avec l’usage d’instruments qui font une auto-collecte de symptômes et l’usage de méthodes qualitatives pour mesurer le bien-être psychologique d’adolescents vivant dans une société touchée par la guerre.

Méthode: Dans deux collectivités ennemies du conflit bosnien, on a recueilli un échantillon de 337 jeunes âgés de 13 à 15 ans (183 de Garazde et 154 de Foca). Ils ont complété le Hopkins Symptoms Checklist et le Harvard Trauma Questionnaire. Basé sur leur score combiné, on a composé un sous-échantillon de 40 adolescents et adolescentes représentant proportionnellement les deux sexes de l’échantillon original et comprenant un nombre égal de participants ayant des scores inférieurs et des scores élevés. Durant les six mois qui ont suivi, on a évalué ce petit groupe au moyen de méthodes qualitatives telles que des entrevues avec l’enfant et le parent et les observations des évaluateurs. Le rendement scolaire servit de mesure pour le fonctionnement social.

Résultats: Mesures qualitatives: Certains des enfants perçus comme étant “moins sains” selon les méthodes qualitatives ont nié avoir des symptômes. Certains enfants perçus comme “sains” manifestaient des symptômes qu’ils considéraient sans importance pathologique. Il semble que se sentir “moins sain” serait l’effet des circonstances d’après-guerre plutôt que de la guerre elle-même. Mesures quantitatives: Deux items de la liste de symptômes s’avèrent capables d’assurer une cohérence interne et une validité discriminante. Toutefois, lorsqu’on compare l’ensemble du bien-être, on note que pour 9 des 40 cas la présence ou l’absence de symptômes ne correspond pas au bien-être de l’enfant. Des éléments des deux questionnaires n’ont pas su discriminer de façon fiable parmi les enfants qui par d’autres mesures se sont avérés “sains” ou “moins sains.”

Conclusions: Les listes qui auto-cotent peuvent servir de mesure en santé publique pour connaı̂tre la prévalence de la détresse psychologique dans des régions touchées par la guerre, mais pour des fins cliniques, elle ne sont pas des mesures adéquates. Les utiliser de paire avec d’autres mesures qualitatives permet de déceler les personnes en besoin et d’éviter de les étiqueter de “pathologiques.”

Resumen

Objetivo: Comparar la utilización de autoinformes de listados de sı́ntomas con la utilización de métodos cualitativos para evaluar el bienestar psicológico adolescente en una sociedad afectada por la guerra.

Método: La muestra fue seleccionada en el ámbito escolar y estaba compuesta por 337 participantes de entre 13 y 15 años pertenecientes a dos comunidades de bandos opuestos del conflicto de Bosnia (183 de Gorazde y 154 de Foca). Todos ellos completaron el Hopkins Symptoms Checklist y el Harvard Trauma Questionnaire. En base a las puntuaciones combinadas en el listado de sı́ntomas, se seleccionó una submuestra, balanceada según el género de los sujetos, compuesta por 40 adolescentes y que incluı́a un número igual de sujetos con puntuaciones altas y bajas de cada bando del conflicto. Después de transcurridos seis meses se procedió a evaluar a esta submuestra (sin conocer sus puntuaciones en el listado de sı́ntomas) con métodos cualitativos que incluı́an entrevistas narrativas del niño/a y del padre y observación participante. Las notas del colegio fueron tomadas como medidas del funcionamiento social.

Resultados: Cualitativos: Algunos niños/as identificados como “menos bien” por los métodos cualitiativos negaron la presencia de sı́ntomas. Algunos niños/as identificados como “bien” tuvieron sı́ntomas que no tenı́an significación patológica para ellos. El sentimiento subjetivo de estar “menos bien” puede estar más relacionado con las circunstancias posteriores a la guerra que con los acontecimientos de la propia guerra. Cuantitativos: Los ı́tems de los dos listados de sı́ntomas han mostrado buena consistencia interna y buena validez discriminante. Sin embargo, la comparación con el bienestar global reveló que todavı́a en 9 de 40 casos en que se notificó presencia o ausencia de sı́ntomas no se obtuvo una correspondencia con el bienestar del niño/a. Los ı́tems de los dos cuestionarios no discriminaron de manera fiable entre los niños identificados como “bien” o “menos bien” por otros medios.

Conclusiones: Los listados de autoinforme pueden ser útiles como medidas de salud pública para evaluar la prevalencia de malestar psicológico en zonas afectadas por la guerra, pero no son un medio adecuado para el “screening” clı́nico. Los listados de sı́ntomas utilizados en combinación con otras aproximaciones cualitativas hacen posible identificar a aquellos sujetos en situación de necesidad y evitan una innecesaria patologización.

Introduction

Courtesy of the global news networks, the crying child and agonized mother are no longer confined to their devastated homelands, but have entered our living rooms, turning all of us into witnesses to their pain and suffering. Five months after air strikes in Kosovo, there were at least ten agencies running psychiatric and psychological support services for children, with 17 projects in the Prishtina municipality alone (UNICEF, 1999). Many such programs use self-report symptom checklists as a way of identifying those in need, but how significant are symptoms? Research evidence from conflict areas around the world shows that between 50 and 80% of children do not have many psychological symptoms (Cairns & Dawes, 1996, p. 130; Saigh, Green, & Korol, 1996). So what is the best method of identifying those in need? This paper compares the usefulness of a symptom-based approach, with qualitative methods such as narrative, lifelines, and participant observation, when assessing the psychological well-being of children in a war-affected society.

There are obvious advantages to self-report symptom checklists, especially when working with limited resources in difficult and insecure conditions. They appear to be simple and quick to administer, and easy to understand and interpret. Non-medical personnel can quickly be trained to use them.

However there are a number of problems. Firstly, many checklists are focused on a single diagnosis: Post-Traumatic Stress Disorder (PTSD) (DeJong, Mulhern, Ford, Van der Kam, & Kleber, 2000; Kuterovac, Dyregrov, & Stuvland, 1994; Thabet & Vostanis, 1999; UNICEF, 1996). The assumption being that after exposure to a traumatic event, this will be the most common problem. Yet mental distress may manifest itself in a wide variety of ways in war-affected areas. Secondly, even when used in research conditions in a Western hospital, the scales lack specificity and sensitivity. A self-assessment screening battery recommended by Yule and Udwin (1991), which consists of the Impact of Events scale (IES), Beck Depression Inventory (BDI) and Revised Children’s Manifest Anxiety scale (RCMA), was used to screen children for PTSD in an accident and emergency department following everyday trauma. It failed to identify 14% of those of those with PTSD, as identified by clinical interview. Arbitrary cutoff scores may have meant that some children with slightly less “symptoms” but significant distress were missed (Stallard, Velleman, & Baldwin, 1999). Cutoff scores will vary depending on the population under study (Joseph, Williams, & Yule, 1991). Most of the commonly used scales have not been properly validated for the countries and conflict situations in which they are used (AACAP, 1998, Richman, 1993). Sack looked at the psychometric properties of the IES in a sample of Cambodian youth. He found that of those who screened positive, only 30% were true cases, and of those who screened negative 89% were non-cases (Sack, Seeley, Him, & Clarke, 1998). While this may not matter so much when using the scale for epidemiological research, it does become significant when the purpose is to identify individuals for treatment.

High levels of symptoms do not necessarily equal psychiatric disability. They may reflect a norm for that population, or a temporary adjustment to the stresses of war (Westermeyer, 2000). Goldstein looked for war related distress symptoms in a sample of 364 displaced children in Bosnia and found that 94% of the children met DSM IV criteria for PTSD. As there was shelling of the city on 35 days in the 2-month period of data collection, he suggested that many of the symptoms were adaptive in that context (Goldstein, Wampler, & Wise, 1997).

On the other hand the absence of symptoms is not necessarily an indication of psychological well being. Avoidant symptoms are by definition hard to measure (Cassidy & Kobak, 1988; Green, 1991) and are best assessed by observing the child. Paradoxically some suggest that higher symptom reports in the aftermath of trauma might indicate better psychological health in the form of openness, self-awareness, trust and processing (Davies & Flannery, 1998). Eisenbruch has shown that the absence of sadness and apparently successful assimilation into a host country may not indicate good health in child Cambodian refugees. Just as the continuing presence of “nostalgia” may be a crucial component of sustaining identity and continuity in a new environment, rather than a symptom of pathological grief (Eisenbruch, 1990).

There is an increasing amount of research showing that it is the way that individuals interpret their experience, and the context in which it occurs, that influences whether or not psychological symptoms are perceived as distressing (Kleinman, 1987). Hume and Summerfield (1994) studying ex-combatants in Nicaragua found poor sleep, noise intolerance, nightmares, hypersensitivity and somatic complaints in the majority, but it was only those unable to maintain social functioning who regarded these symptoms as problematic.

A focus on quantifiable variables such as the presence or absence of psychopathological symptoms means that the young person’s subjective understanding, which is rooted in their life history and social world, cannot be fully explored. So the meaning of a particular symptom to that person, in that cultural context, cannot be understood. War is a collective experience and the primary impact on individuals is through witnessing the destruction of their social worlds embodying their history, identity and living values (Summerfield, 1999). Changes in trust, world view, moral outlook, sense of security, and sense of connection to others may have a profound impact on well-being but not manifest themselves as “symptoms” that can be counted, but rather as “ways of living” that must be observed and discussed.

There have been frequent calls for more qualitative and ethnographic work to address these issues (Cairns, 1994; Cairns & Dawes, 1996; Jensen & Shaw, 1993; Kleinman, 1987, Richman, 1993). Yet there are relatively few studies of war affected children that combine qualitative and quantitative work (Farwell, 1999, Miller, 1996; Weine et al., 1995).

The primary researcher has been engaged in an ethnographic study of Bosnian adolescents’ understandings of war in two towns on opposite sides of the conflict. Self-report symptom checklists were given to young people, and their subjective world view was explored through in-depth interviewing and participant observation. This provided an opportunity to compare the usefulness of symptom checklists with other qualitative means of assessing psychological well being.