RESEARCH
Maternal psychosocial well-being in Eritrea: application of participatory methods and tools of investigation and analysis in complex emergency settings
Bien-être psychosocial maternel en Erythrée : application de méthodes et d'outils d'investigation et d'analyse participatifs dans des situations d'urgence complexes
El bienestar psicosocial materno en Eritrea: aplicación de métodos e instrumentos de investigación y análisis de carácter participativo en entornos de emergencia complejos
Astier M. AlmedomI; Berhe TesfamichaelII; Abdu YacobIII; Zaïd DebretsionIV; Kidane TeklehaimanotV; Teshome BeyeneVI; Kira KuhnVI; Zemui AlemuVII
IHenry R. Luce Professor in Science and Humanitarianism, Department of Biology and Friedman School of Nutrition Science and Policy, Tufts University, Boston, MA, USA (email: astier-m.almedom@tufts.edu). Correspondence should be addressed to this author
IIParticipatory Trainer/Facilitator and Social Animator, Water Resources Department, Ministry of Environment, Water and Land, Asmara, Eritrea
IIISanitarian and Participatory Trainer/Facilitator, Ministry of Health, Barentu, Eritrea
IVSenior Researcher and Financial Administrator, Eritrean Relief and Refugee Commission, Keren, Eritrea
VNutrition Unit, Ministry of Health, Asmara, Eritrea
VIField Intern, Ministry of Health, Asmara, Eritrea
VIIDirector, Primary Health Care Division, Ministry of Health, Asmara, Eritrea
ABSTRACT
Objective: To establish the context in which maternal psychosocial well-being is understood in war-affected settings in Eritrea.
Method: Pretested and validated participatory methods and tools of investigation and analysis were employed to allow participants to engage in processes of qualitative data collection, on-site analysis, and interpretation.
Findings: Maternal psychosocial well-being in Eritrea is maintained primarily by traditional systems of social support that are mostly outside the domain of statutory primary care. Traditional birth attendants provide a vital link between the two. Formal training and regular supplies of sterile delivery kits appear to be worthwhile options for health policy and practice in the face of the post-conflict challenges of ruined infrastructure and an overstretched and/or ill-mannered workforce in the maternity health service.
Conclusion: Methodological advances in health research and the dearth of data on maternal psychosocial well-being in complex emergency settings call for scholars and practitioners to collaborate in creative searches for sound evidence on which to base maternity, mental health and social care policy and practice. Participatory methods facilitate the meaningful engagement of key stakeholders and enhance data quality, reliability and usability.
Keywords: Maternal welfare/psychology; Mothers/psychology; Mental health; Anxiety; War; Emergencies; Mental health services; Social support; Eritrea (source: MeSH, NLM).
RÉSUMÉ
Objectif: Définir comment est perçue la notion de bien-être psychosocial maternel dans des environnements touchés par la guerre en Erythrée.
Méthodes: Des méthodes et des outils d'investigation et d'analyse participatifs prétestés et validés ont été utilisés pour permettre aux participants d'effectuer un travail de collecte de données qualitatives, d'analyse sur le terrain et d'interprétation des données.
Résultats: En Erythrée, le bien-être psychosocial des mères est assuré principalement par les systèmes traditionnels de protection sociale qui se situent pour la plupart hors du champ des soins de santé primaires institutionnalisés. Les accoucheuses traditionnelles maintiennent un lien essentiel entre les deux systèmes. La formation systématique de ces dernières et la fourniture régulière de trousses d'accouchement stériles semblent être des options valables, tant sur le plan de la politique de santé que sur le plan pratique, face aux problèmes d'effondrement de l'infrastructure et de surmenage et/ou de mauvais comportement du personnel des services de maternité structurés faisant suite à un conflit.
Conclusion: Compte tenu des progrès méthodologiques de la recherche en santé et du manque la pénurie de données sur le bien-être psychosocial maternel dans des situations d'urgence complexes, il serait utile que les chercheurs et les praticiens fassent preuve de créativité et unissent leurs efforts pour rassembler des données factuelles fiables sur lesquelles pourraient s'appuyer la politique et la pratique en matière de santé maternelle, de santé mentale et de protection sociale. Le recours à des méthodes participatives facilite un véritable engagement des principaux acteurs et permet d'améliorer la qualité, la fiabilité et la commodité d'utilisation des données.
Mots clés: Protection maternelle/psychologie; Mère/psychologie; Santé mentale; Angoisse; Guerre; Urgences; Service santé mentale; Soutien social; Erythrée (source: MeSH, INSERM).
RESUMEN
Objetivo: Establecer el contexto en que se interpreta el bienestar psicosocial materno en los entornos afectados por conflictos bélicos en Eritrea.
Métodos: Se emplearon instrumentos de investigación y análisis y métodos de carácter participativo, debidamente preensayados y validados, para permitir a los participantes colaborar en los procesos de recopilación de datos cualitativos, análisis in situ e interpretación.
Resultados: El bienestar psicosocial materno en Eritrea se mantiene principalmente gracias a sistemas tradicionales de apoyo social que en su mayoría no forman parte de la atención primaria obligatoria. Las parteras tradicionales constituyen un eslabón crucial entre esas dos esferas. La capacitación formal y el suministro regular de kits estériles para partos parecen ser opciones válidas para la política y práctica sanitarias ante los retos posconflicto que suponen una infraestructura arruinada y una fuerza laboral desbordada y/o poco delicada en los servicios de maternidad.
Conclusión: Considerando los avances metodológicos de las investigaciones sanitarias y la escasez de datos sobre el bienestar psicosocial materno en los entornos de emergencia complejos, es preciso que los especialistas y los profesionales colaboren para buscar creativamente pruebas sólidas en las que basar la acción normativa y práctica en los servicios de maternidad, salud mental y atención social. Los métodos participativos ayudan a lograr un compromiso provechoso por parte de interesados directos clave y mejoran la calidad, fiabilidad y usabilidad de los datos.
Palabras clave: Bienestar materno/psicología; Madres/psicología; Salud mental; Ansiedad; Guerra; Urgencias médicas; Servicios de salud mental; Apoyo social; Eritrea (fuente: DeCS, BIREME).
Introduction
There has recently been a growth of interest in mental health research, particularly in relation to the improvement of policy and practice across sectors (14). Most studies have involved descriptive data on mental disorders at the population level and have used large-scale survey methods, notably in work on the global burden of disease and disability-adjusted life years (5). Policy and practice in the field of mental health have traditionally been based on evidence obtained from such disease-centred studies. There is, however, growing interest in person-centred approaches to mental health policy, particularly in connection with such factors as stigma and social exclusion (6, 7). Interdisciplinary salutogenic (as opposed to pathogenic) perspectives on human ecology, coping mechanisms, strategies of adaptation, stress management and resilience have become increasingly relevant (8, 9).
Mental health studies in settings of complex humanitarian emergency fall into either the category of epidemiological psychiatry (10) or that of psychosocial psychiatry, often incorporating anthropology and/or sociology (11, 12). The present study focuses on the psychosocial well-being of women during pregnancy, childbirth, and the postpartum period in complex emergency settings in Eritrea associated with the 19982000 war with Ethiopia. Women, especially mothers of infants and young children, bear the brunt of economic, social, cultural and health burdens in societies affected by war. Moreover, the morale and mental well-being of mothers are important determinants of infant health in wartime (13). Consequently, this study is also of relevance to infant and child health policy and practice in complex emergency settings.
Eritrea has a population of about 4 million and has frontiers with Djibouti, Ethiopia, and the Sudan. In the highlands, staple crops and cash crops are grown (taff, wheat, barley, maize and, in the green belt of Semenawi Bahri, coffee) and there is dairy farming and goat, sheep, and poultry production. In the lowlands and middle altitudes, agro-pastoralism and transhumance are practised. There are nine ethnic groups with marked linguistic and sociocultural differences: predominantly Tigrinya-speaking orthodox Christians in the highlands; mostly Moslem societies consisting of Afar, Bilein, Hidareb, Nara, Rashaida, Saho and Tigre, ethnic groups in mid-altitude and lowland areas; along with Kunama Christians and followers of traditional religion.
The following questions are considered in this paper: What are the local perceptions of available statutory and non-statutory health services? What factors are believed to mitigate the effects of war-induced anxiety and mental distress? How do women and men differ in their perceptions of mental well-being?
Methods
Fieldwork was conducted in December 2001 and January, August and September 2002 in seven locations of Eritrea selected to include a range of geoclimatic zones and modes of subsistence (Table 1). The western lowland and mid-altitude provinces of Zoba Anseba and Zoba Gash-Barka and the southern highland provinces of Zoba Debub were the areas worst affected by the war because of their proximity to the Ethiopian border. Word-of-mouth invitations resulted in the recruitment of 104 women and 124 men to the study. All the ethnic groups except the Afar and Rashaida were represented.
Participatory methods and tools of investigation and on-site analysis were employed. They had been extensively field-tested, evaluated and validated in various countries, including Eritrea (14, 15). The study team carefully questioned and listened to the participants and kept written records. The quality and reliability of the data were established as follows: feedback sessions held at each study site in order to allow the participants to review and interpret, and, if necessary, correct the findings, and to enable the study team to follow up incomplete data; checks on the accuracy of factual or objective (as opposed to subjective) data relating to historical events, involving the use of published material where available and other secondary sources; and data triangulation by methods and sources.
The study was designed in collaboration with the Ministry of Health, with the fieldwork being conducted in accordance with established ethical protocols and procedures for the participatory research methods and tools used (14). The guidelines of the Institutional Review Body of Tufts University were observed. No incentives were offered and no obligations were imposed on the participants. The purpose of the study and all the group activities were explained in the local language or languages by trained and qualified facilitators at the beginning of each meeting. Any topic that the participants were unwilling to discuss was set aside. Women's and men's discussion groups were held in separate locations. All meetings began and were brought to a formal closure by the facilitators. For reasons unrelated to the conduct of the study, two participants left before their discussion groups were scheduled to end.
Results
Community maps and historylines were most appreciated by study participants who saw these as useful records of local knowledge that had not previously been sought and/or recognized. Striking differences emerged between urban or semiurban settings where little or no displacement had occurred and camps for internally displaced persons in respect of the amount of detail in community maps, seasonal calendars and historylines.
Quality and effectiveness of available health services
The participants emphasized the inadequacy of health services in general and of maternity services in particular at all sites except Ghinda (Table 2). They indicated that it was urgently necessary to have emergency transportation services for women in labour, to provide training and a regular supply of sterile delivery kits for traditional birth attendants, and to improve the attitudes and behaviour of health workers.
Childbirth continued to take place mostly in people's homes with the help of traditional birth attendants. In cases of protracted labour, loss of blood and/or delayed expulsion of the placenta, the traditional birth attendants referred their patients and often accompanied them to the nearest medical facility. Fatalities occurred because of the limited availability of transport and ambulance services. By the time a weak, anaemic, and haemorrhaging pregnant woman arrived at a hospital she might already be in shock or coma. The highest estimates of maternal mortality, up to two or three deaths in a particular month, were reported in Gogne, where most women were said to be anaemic during pregnancy and weak during childbirth.
Health service provision in camps for internally displaced persons was generally limited. In the Adi Qeshi camp, which had existed for 3 years, mobile clinics had come and gone at different times and were mostly limited to immunization and nutritional surveillance of under-5-year-olds. Women reported that most of their infants and young children were malnourished because of inadequate rations and delayed weaning. In Hamboka and Awle'a Hahale, most of the residents had been displaced more than once before arriving in the Sen'afe area. Both men and women reported that they had received a better quality of health care in a previous camp at Mai Habar, a lowland, heavily malarial location farther away from the war zone, where a Catholic mission had provided mobile clinics. Their nearest hospital at Sen'afe had been destroyed: all that remained was a row of large tents that had been left behind by Médecins sans Frontières doctors soon after the bombing and shelling had stopped. The participants said that very long waiting times and rude health workers were experienced in the tent hospital. As an example of what this actually meant, it was reported that a young boy became ill after queuing all night in the cold on behalf of his sick grandmother who had to be seen by a doctor. Sen'afe lies over 2500 m above sea level and cases of upper respiratory infections are common.
With the exception of Ghinda Hospital, where nurses and doctors were said to have good relations with patients and traditional birth attendants, the behaviour of health workers was reportedly unacceptable. In Elaberid, the worst example, nurse midwives reportedly treated their patients harshly, routinely saying to women in labour "So you expect to have a baby without pain, do you?". In contrast, traditional birth attendants were reported to have commendable skills with pregnant and newly-delivered mothers but they were not duly recognized by the formal health care system. The Ministry of Health training programme for traditional birth attendants was perceived to be good. Most trained traditional birth attendants continued to have a good rapport with their clients but they lacked material support from the ministry. At Elaberid and Ghinda the question of cash payment for trained traditional birth attendants was debated in women's discussion groups. Some of the traditional birth attendants present found the idea of cash payment problematic because it undermined the humanitarian and prestigious role that earned them trust and respect in their communities. Others did not object to cash payment and said that nakfa 2040 (less than US$ 4) per delivery was too little. All traditional birth attendants emphasized the need for formal training and regular supplies of sterile delivery kits.
Factors mitigating the effects of war-induced anxiety and distress
At the micro level, the factors that were believed to mitigate mental ill-health in women included protective traditional practices during pregnancy and childbirth. It was said that, even in normal times, pregnancy and childbirth were associated with anxiety and mental distress in women. For this reason most Eritrean women are traditionally provided with special care and support during pregnancy, childbirth, and beyond. Most ethnic groups still uphold the tradition of postpartum seclusion and rest for mothers who have just given birth. At this time, mothers are exempted from any work other than breastfeeding, and their families and friends help with domestic tasks and ensure that they eat and rest well. Among the Tigrinya highlanders and other Christians this period may last for 4080 days, while among Muslims a period of up to 6 months or even longer is allowed.
The war limited women's ability to enjoy postpartum seclusion and rest. The absence of husbands, brothers and brothers-in-law and the associated loss of material, moral and emotional support added to the burdens of expectant and lactating women. Women emphasized the value of social support from family, friends and confidantes, including traditional birth attendants. Nevertheless, even during and after the war, most people were said to be in good mental health and cases of severe mental illness were rare, even among mothers who had just given birth.
Women in Gogne concluded that the best "maternity leave" practices were those of the Hidareb ethic group. A Hidareb man takes full responsibility for the care of his wife and new baby during the first 56 months postpartum, making provision for his mother-in-law to live in the family home and look after her daughter and new grandchild. If his wife and baby fail to thrive he dismisses his mother-in-law and brings in a substitute carer. Such dismissals were believed to have little or no social consequences, because a mother-in-law who failed to feed and care for her own daughter would expect to leave and spare her son-in-law the shame and ridicule he would otherwise have to endure in the community.
Women in the Adi Qeshi and Hamboka camps reported that traditional postpartum seclusion and rest were no longer possible because of lack of resources. The case of a young mother, who had been mentally ill during her pregnancy and had deteriorated after delivery, was discussed in Hamboka. This woman could not breastfeed and care for her baby, which was therefore fed on goat's milk by relatives. The woman was taken to stay with her mother, who lived in another village. Women in Gogne and Ghinda mentioned rare but similar cases of maternal mental illness before and after childbirth. They believed the causes to be unwanted pregnancy and lack of social support, particularly among young and unmarried women. A mother's inability to relate to her infant and failure to breastfeed and care for it were the most commonly identified indicators of maternal mental illness.
At the macro level the most important protective factor for internally displaced persons in camps was reported to be statutory social support in the form of assisted relocation. A special effort was made to allow village units to remain together even after displacement. For example, people in the Adi Qeshi camp, named after a cluster of evacuated villages, reported that much of their anxiety was alleviated by the manner in which their flight to safety was facilitated. Lorries, buses and trucks had been provided primarily for the vulnerable, i.e. pregnant and nursing women, children, the sick and the elderly, while able people had left their villages on foot when Ethiopian attack was imminent. Villagers and their administrators and leaders were taken to a camp equipped with tents, latrines, water points and reservoirs. The administrators and leaders were thus able to maintain contact with the people for whom they were responsible, and the surviving men and women were not required to prove their identities and family sizes to officials or aid workers whom they did not know. The village administrators worked closely with the Eritrean Relief and Refugee Commission in order to expedite the distribution of food and blankets and to facilitate the immunization of children. This helped to maintain pre-existing social ties and support networks. Moreover, people were kept informed about events outside their camp through the national radio station, Dimtsi Hafash, and through regular meetings convened by the administrators. Most participants believed that a sense of being in touch with their own community and the country as a whole protected them from Chinquet (mental oppression), 
(thinking too much) and Ihihta (sighing). However, some expressed weariness about the length of time they had lived in camps. In Adi Qeshi, women explained that even if the peace agreement and border demarcation went smoothly they might not be able to return to their homes until their villages were cleared of land mines.
Gender and mental well-being
Overall, men and women were equally prepared to participate in group discussions of mental well-being in general terms. In the Adi Qeshi camp, women reported that previous visitors had repeatedly asked many questions about violence and trauma they might have experienced. Although many women had suffered, most had coped reasonably well "thanks to Qidisti Mariam" (Saint Mary), who "helps women during the difficult periods of pregnancy and childbirth". As time passed, some women had become frustrated about their prolonged dependence on hand-outs and food aid and the loss of Qisanet (serenity, peace of mind). These women preferred not to discuss the matter further.
Marked differences were found between men's and women's reported psychosocial well-being in the community and their priorities for action when peace returned. In Elaberid, Ghinda and Gogne, men discussed the increasing prevalence of restlessness during daytime and sleeplessness at night caused by the prolonged conflict and uncertainty about the prospects of lasting peace. In the Adi Qeshi camp, men discussed their inability to come to terms with the loss of lives and assets. They described overwhelming feelings of anger and a desire to take revenge. In the Hamboka camp, men were mainly concerned about their reputations. They believed that they were being blamed for starting the war by failing to resolve disputes with neighbours on the other side of the border.
Discussion
A sociocultural basis for research and humanitarian interventions has recently been advocated in order to bring about psychosocial well-being among the survivors of disasters (9, 11, 16). WHO supports the view that preconceived and/or crude notions of mass trauma (as manifested by post-traumatic stress disorder) and social dysfunction should not drive humanitarian action (18, 19). Nongovernmental organizations such as the Save the Children Fund have also advocated caution in this matter (20). Nevertheless, it remains rare for psychosocial projects to be based on assessments of local sociocultural contexts and/or expressed needs. Many needs assessment exercises concentrate on requirements for what funding and operational agencies have to offer rather than on needs articulated by victims (17, 18). The present study has set out to demonstrate that understanding the context and expressed needs and priorities of survivors of complex emergencies is both essential and achievable.
The situation in Eritrea does not precisely fit the conventional definition of a complex humanitarian emergency. This term was coined to distinguish so-called natural disasters from those caused by political turmoil, conflict and war (21). However, most if not all disasters are complex, be they natural or man-made. Eritrea has experienced both kinds of disaster. During the armed struggle for independence from 196191, primary and secondary health care systems were developed under complex and chronic emergency conditions in response to pressing problems among both civilians, mostly in rural areas, and members of the armed liberation movement. A symbiotic relationship existed between the civilian and armed groups. The civilians depended on the armed groups for material social support, including health care provision, while the armed groups depended on the civilians for moral support, protection and emotional social support. Both relied on mutual and timely social support of the cognitive type (22). There is much evidence of individual and collective resilience and resourcefulness in the face of war and calamity (23, 24). However, this does not mean that people who succumb to mental illness should be overlooked. The mental well-being and morale of mothers with infants and young children should be attended to, as women bear the burden of responsibility of maintaining their families, bringing up future citizens and preserving the social fabric, health and well-being of their communities in emergency conditions (13).
The health and mental well-being of mothers and infants are protected and promoted at the micro level by social support in the form of postpartum seclusion and rest for mothers, and at the macro level by social support in the form of government- assisted relocation. On the other hand, prolonged and multiple or serial displacements present risk. Traditional birth attendants have a major role in the promotion of maternal mental well-being. The formal recognition and integration of trained traditional birth attendants in mainstream maternity services may be viable and sustainable in post-conflict primary heath care in Eritrea. The problem of harsh and ill-mannered health workers is not unique to the settings described in this paper. Other studies have found that nurse midwives may be abusive to their patients for various reasons, ranging from professional insecurity to a notion of patient inferiority (25). Such abuse discourages women from seeking medical care during pregnancy and childbirth (26). Traditional birth attendants could conceivably be recruited to train nurse midwives in the care of patients. Of course, many nurse midwives and other health workers care for their patients with professional integrity and compassion. Unfortunately, news of bad practice travels quickly and may have a more lasting impact on perceptions.
Concern over the quality and reliability of qualitative data can safely be allayed. Participant checking and cross-checking with other reliable sources have confirmed the accuracy of lay people's knowledge of their own social and political history and geography. For example, the data from Gogne's historyline concur with published historical accounts of the ethnic diversity, political allegiances and conflict in the region (27, 28). Women's accounts of shortcomings in maternity health services were corroborated by visiting the health facilities mentioned and by reviewing independent sources (29). Participatory qualitative research of the type described here can evidently help to address the problems of unreliable and unusable data relating to disasters (30). 
Acknowledgements
We thank the following: the participants, who gave generously of their time and knowledge; Ayesha Sa'id and Sa'edia Suleiman, interpreters of the Kunama and Saho languages, respectively; Sister Kidisty Habte, Sister Azenegash Ghebresellasie and Dr Tesfaldet Teclemichael for helpful insights; and two anonymous reviewers for constructive criticisms of an earlier draft of this paper. The Eritrean Ministry of Health and USAID/Technical Assistance and Support Cooperation provided logistical assistance.
Funding: The study was funded by Tufts University Henry R. Luce Professorship in Science and Humanitarianism. Additional fieldwork was supported by the United States Centers for Disease Control and Prevention, National Center for Environmental Health, International Emergency and Refugee Health Branch, Requisition No. IER08/01.
Conflicts of interest: None declared.
References
1. Kessler RC. Psychiatric epidemiology: selected recent advances and future directions. Bulletin of the World Health Organization 2000;78:464-74.
2. Brundtland GH. Mental health in the 21st century. Bulletin of the World Health Organization 2000;78:411.
3. Gureje O, Alem A. Mental health policy development in Africa. Bulletin of the World Health Organization 2000;78:475-82.
4. Jenkins R. Making psychiatric epidemiology useful: the contribution of epidemiology to government policy. Acta Psychiatrica Scandinavica 2001;103:2-14.
5. Vos T, Mathers CD. Burden of mental disorders: Australian and global burden of disease studies. Bulletin of the World Health Organization 2000;78:427-38.
6. The world health report 2001 mental health: new understanding, new hope Geneva: World Health Organization; 2001.
7. Sayce L. From psychiatric patient to citizen: overcoming discrimination and social exclusion. Basingstoke: Macmillan; 2000. New York: St. Martin's Press; 2000.
8. Antonovsky A. Unraveling the mystery of health: how people manage stress and stay well. San Francisco: Jossey-Bass; 1987.
9. Ahearn FL. Psychosocial wellness: methodological approaches to the study of refugees. In: Ahearn FL, editor. Psychosocial wellness of refugees: issues in qualitative and quantitative research. New York: Berghahn Books; 2000. p. 3-33.
10. Lopes Cardozo B, Vergara A, Agani F, Gotway CA. Mental health, social functioning, and attitudes of Kosovar Albanians following the war in Kosovo. JAMA 2000;284:569-77.
11. Summerfield D. War and mental health: a brief overview. BMJ 2000;321:232-5.
12. Summerfield D. The invention of post-traumatic stress disorder and the social usefulness of a psychiatric category. BMJ 2001;322:95-8.
13. Almedom AM. Mother's morale and infant health in Ethiopia. In: Boyce AJ, Reynolds V, editors. Human populations: diversity and adaptation. Oxford: Oxford University Press; 1995. p. 138-54.
14. Almedom AM, Blumenthal U, Manderson L. Hygiene evaluation procedures: Approaches and methods for assessing water- and sanitation-related hygiene practices. Boston (MA): International Nutrition Foundation for Developing Countries; 1997.
15. Almedom AM, Esak L, Gebrehiwot E, Habtemichael B, Isaac G, Saleh M, et al. Hygiene evaluation in Zobas Ma'ekel and Debub, Eritrea. London and Asmara: Report to UNICEF and the London School of Hygiene and Tropical Medicine; 1998.
16. Summerfield D. The social experiences of war and some issues for the humanitarian field. In: Bracken P, and Petty C, editors. Rethinking the trauma of war. London: Free Association Books; 1998.
17. Summerfield D. A critique of seven assumptions behind psychological trauma programmes in war-affected areas. Social Science and Medicine 1999;48:1449-62.
18. Pupavac V. Afghanistan: The risk of international psychosocial risk management. Health in Emergencies 2002;12:1-2.
19. Pupavac V. Therapeutic Governance: Psycho-social intervention and trauma risk management. Disasters 2001;25:358-72.
20. Bracken P, Petty C, editors. Rethinking the trauma of war. London: Free Association Books; 1998.
21. Duffield M. Complex emergencies and the crisis of developmentalism. IDS Bulletin 1994; 25:37-45.
22. Jacobson DE. Types and timing of social support. Journal of Health and Social Behavior 1986;27:250-64.
23. Gottesman L. To fight and learn. Lawrenceville (NJ) and Asmara, Eritrea: Red Sea Press; 1998.
24. Connell D. Against all odds: a chronicle of the Eritrean revolution (with a new afterward on the postwar transition). Lawrenceville (NJ) and Asmara, Eritrea: Red Sea Press; 1997.
25. Jewkes R, Abrahams M, Mvo Z. Why do nurses abuse patients? Reflections from South African obstetric services. Social Science and Medicine 1998;47:1781-95.
26. Abrahams M, Jewkes R, Mvo Z. Health care-seeking practices of pregnant women and the role of the midwife in Cape Town, South Africa. Journal of Midwifery and Women's Health 2001;46:240-7.
27. Naty A. Memories of the Kunama of Eritrea towards Italian Colonialism. Africa 2001;LVI:573-89.
28. Trevaskis GKN. Eritrea: a colony in transition. London: Oxford University Press; 1960.
29. Kwast B. Report on a consultancy visit to review the maternal health program in Eritrea (Draft). Asmara, Eritrea: USAID Technical Assistance and Support Contract (TASC); 2001.
30. World disasters report: focus on public health. Geneva: International Federation of Red Cross and Red Crescent Societies; 2000.