Abstract
Objectives. This study assessed patterns of displacement and human rights abuses among Kosovar refugees in Macedonia and Albania.
Methods. Between April 19 and May 3, 1999, 1180 ethnic Albanian refugees living in 31 refugee camps and collective centers in Macedonia and Albania were interviewed.
Results. The majority (68%) of participants reported that their families were directly expelled from their homes by Serb forces. Overall, 50% of participants saw Serb police or soldiers burning the houses of others, 16% saw Serb police or soldiers burn their own home, and 14% witnessed Serb police or soldiers killing someone. Large percentages of participants saw destroyed mosques, schools, or medical facilities. Thirty-one percent of respondents reported human rights abuses committed against their household members, including beatings, killings, torture, forced separation and disappearances, gunshot wounds, and sexual assault.
Conclusions. The present findings confirm that Serb forces engaged in a systematic and brutal campaign to forcibly expel the ethnic Albanian population of Kosovo. In the course of these mass deportations, Serb forces committed widespread abuses of human rights against ethnic Albanians.
The Kosovo crisis resulted in the largest population displacement in Europe since World War II. Accurately documenting specific human rights abuses and estimating the extent of human rights abuses during such complex humanitarian emergencies are essential to informing political interventions in the region, guiding humanitarian relief efforts, and holding perpetrators accountable for their actions. Unfortunately, estimates of human rights abuses during such crises are generally based on unscientific anecdotal reports. Although difficult and potentially dangerous, obtaining representative firsthand data is crucial.
Epidemiologic methods have been used in human rights work,1–5 but they have not been used to establish the incidence and nature of human rights abuses occurring during large and complex humanitarian emergencies. In this article, we attempt to provide a scientific, population-based estimation of the magnitude of human rights abuses in the Kosovo crisis.
In 1980, the death of Josip Broz Tito, Yugoslavia's leader since World War II, heralded a decade of instability and increasing tensions in the Federal Republic of Yugoslavia and the province of Kosovo. Protests by ethnic Albanians resulted in arrests and prison sentences. Human rights abuses and discrimination against the Albanian majority in Kosovo increased steadily after revocation of Kosovo's autonomous status within Yugoslavia in 1989.
Increasing repression of ethnic Albanians by the Federal Republic of Yugoslavia, headed by President Slobodan Milosevic, and the failure of peaceful resistance efforts led to the emergence of an armed insurgency, the Kosovo Liberation Army (KLA), and the beginning of armed conflicts with Serb forces in 1998.6 Serb forces engaged in an indiscriminate military campaign against the KLA and Albanian civilians beginning in March 1998 that resulted in the mass deportation of more than 800 000 ethnic Albanians from their homeland and the displacement of almost all of the 500 000 to 600 000 individuals who remained in Kosovo.7
On March 24, 1999, the North Atlantic Treaty Organization (NATO) launched air strikes against Serbian military positions. In the weeks leading up to the bombing, the Yugoslav government had violated nearly every point of its October 1998 cease-fire agreement with NATO.
Kosovo had a population of approximately 2 million before the outbreak of armed conflicts in March 1998. By mid-April 1999, the United Nations High Commissioner for Refugees estimated that almost half a million refugees had crossed the border from Kosovo to Macedonia and Albania.8,9 This flow of refugees resulted in immense financial and political strains on neighboring countries. Indeed, some actions in Macedonia, including forced displacement of refugees and placement of camps near border areas, violated refugee rights and protections.10–12
The purpose of this study was to document patterns of forced displacement and human rights abuses among Kosovar refugees living in Macedonia and Albania. The hundreds of thousands of potential witnesses to war crimes and human rights abuses in Macedonian and Albanian refugee camps provided an important opportunity for gathering data on the incidence and nature of abuses. This study, excerpted from a more extensive report,13 represents the first epidemiologic assessment of human rights abuses among Kosovar refugees in Albania and Macedonia, the 2 countries that hosted the most refugees.
At the start of the study, there were an estimated 132 000 refugees in Macedonia (58% with local families and 42% in camps) and 365 000 in Albania (65% with local families and 35% in collective centers and tented camps).8,9 Refugees living in tented camps and collective centers were included in the study population. There was not enough information on the number of sites, their locations, and the accessibility of the host-family population to permit these variables' inclusion in a sampling frame. In Macedonia, interviews were conducted in all 6 refugee camps, which were located in the north, near Skopje. In Albania, interviews were conducted at 25 different sites: 7 tented camps in Kukes, 9 camps and 7 collective centers in the Tirana/Durres/Elbasan region, 1 camp in Korce, and 1 collective center in Lozhan. In total, we sampled 31 refugee camps and collective centers in Macedonia and Albania.
We sampled households via a modified form of systematic probability sampling. The target sample size was 1000, 500 in Macedonia and 500 in Albania; refugees in Macedonia were oversampled by a factor of about 2.75. In Macedonia, the number of interviews conducted in each camp was based on the proportion of the estimated number of refugees in that camp in comparison with the total estimated number of 55 440 refugees in all 6 refugee camps.
In each camp, a map of tent placement was established, and tents were counted. The number of households in each camp was estimated and the number of interviews to be conducted in that camp determined. A sampling interval (n) was calculated by dividing the number of households in the camp by the number of interviews to be conducted in the camp. A starting household was determined by random number generation, and each nth household was interviewed until the entire camp had been surveyed.
In Albania, the estimated 127 750 refugees were housed in more than 120 camps and collective centers, many of which were small and geographically isolated.14 We sampled the 25 sites estimated to have more than 1000 refugees. Camps and collective centers were mapped as described earlier. Because of logistic constraints involved in sampling every site and the lack of information on the overall distribution of refugees in Albania, we sought to adequately represent refugees settled in smaller camps by modifying the sampling interval according to camp population. We surveyed every fourth household in camps hosting 1000 to 2000 refugees, every sixth household in camps hosting 2000 to 3500 refugees, and every eighth household in camps with more than 3500 refugees.
Investigators interviewed 1 person per household. Household members were asked whether they were willing to be interviewed. If they agreed, they were asked to nominate a household representative who could provide the most accurate account of the experiences of the entire household over the past year. Although adults were preferred, a small number of individuals younger than 18 years participated. Interviews with minors were conducted with the permission of their parents. In addition, we actively encouraged participation by women. If a household representative was not available during our first attempt, the household was visited a second time.
Given the urgency of the problems we studied, the project was conceived, designed, and implemented within a period of only 2 weeks. Under such circumstances, evaluation by an institutional review board was not possible. Therefore, the research proposal was reviewed and approved by an ad hoc group of individuals with expertise in clinical medicine, public health, and bioethics. Participants were informed of the sensitive nature of the interview content, assured that refusing to take part would not jeopardize their access to aid or safety, and told that they could stop at any time. Respondents could choose to remain anonymous or provide their names.
The questionnaire contained 50 questions that assessed patterns of forced migration among the individuals interviewed and their household members and human rights abuses both witnessed and experienced. The content was based on case information collected during the first 2 weeks of April 1999. The questionnaire was written in English and translated into Albanian. It was pilot tested among 10 refugees in Macedonia and modified to improve clarity and response options. Given the urgency of the project, survey development did not include provisions such as reverse-scored items and measures of social desirability.
The survey sought information in the following areas: (1) demographic characteristics, (2) time frame and reasons for leaving home (and other interim locations) and subsequent arrival in the refugee camp or center, (3) witnessing of human rights abuses, and (4) refugee household members' experience of human rights abuses. The complete content of the survey has been published elsewhere.13 Abuses included forced separation or disappearance, beating, shooting, killing, torture, and sexual assault, including rape. Torture was defined according to the United Nations Convention Against Torture,15 and beatings were considered single episodes of beating of limited duration and intensity.
There has been considerable research on the measurement of traumatic experiences16–18; however, the present survey was designed specifically to assess experiences of human rights abuses to serve the interests of justice and accountability. The selection of specific human rights abuses in the study was based on a presurvey assessment (including interviews with refugees and humanitarian assistance providers, human rights reports, and pilot testing of the survey instrument) and relevant international human rights and humanitarian law.
All interviews were conducted during the 2-week period between April 19 and May 3, 1999, in or near respondents' tents or other places of residence. Usually, other household members, relatives, or friends were present. Each interview lasted approximately 30 minutes. Participants did not receive material compensation.
Interviews were conducted by research team members with translators or by local Albanian-speaking research assistants after several days of training and supervision. Because the interviewers developed a comprehensive understanding of each survey instrument item, the face-to-face interview approach offered the opportunity to clarify and explain survey items to participants. Several measures were taken to minimize any variance originating in the process of clarifying and explaining survey items. The research assistants' training included standardized techniques for clarifying and explaining specific questions. Also, daily supervision of data collection included dissemination of clarifying strategies that emerged in the course of individual interviews. All questionnaires were reviewed for completeness and for correctness of recording after the interview. The data were analyzed via standard analytic computer programs19 designed for descriptive data analysis.
To generate population estimates of human rights abuses among the 1.8 million ethnic Albanians,9 we calculated weighted overall proportions by using the inverse of the sampling fraction in the interview locations. According to United Nations estimates, there were 132 000 refugees in Macedonia and 365 000 refugees in Albania.8,9 We calculated 95% confidence intervals (CIs) for weighted estimates by multiplying the standard error of the proportion by 1.96 and applying the resulting value to the population estimates.
Of the 1209 refugee households selected to participate in the study, 1180 (98%) did so. Of the 29 household representatives who did not participate, 8 refused, 19 were not available during either of the 2 attempted visits, and 2 interviews could not be completed because the respondents were overcome by emotion. Of the 1180 refugee households surveyed, 511 (43%) were located in refugee camps or centers in Macedonia, and 669 (57%) were located in refugee camps or centers in Albania (Table 1). The ages of the household representatives interviewed ranged from 14 to 85 years, with a mean of 40 years (SD = 14). Two thirds of the respondents were men. Nearly all respondents identified themselves as Albanian and Muslim.
The majority (937; 79%) of household representatives interviewed reported that their families were displaced by Serb forces within the month before the interview (Table 1). Median time spent in the refugee camp or center was 2 weeks (range: 1–8 weeks). Most (68%) of the representatives interviewed reported that their primary reason for leaving Kosovo was direct expulsion by Serb forces. Among the 806 refugees expelled by Serb forces, circumstances included Serb police or soldiers coming to participants' homes (426; 53%), Serb bombing (298; 37%), Serb police or soldiers harming people (46; 6%), and Serb police or soldiers destroying people's property (36; 4%). In addition, nearly one quarter (23%) of the respondents reported that their families left Kosovo because they feared Serb forces. None of the Kosovar refugees surveyed reported that they left their homes in Kosovo because of NATO bombing.
More than one third (35%) of the respondents either witnessed Serb police or soldiers killing someone or saw bodies of people they believed were killed by Serb police or soldiers (Table 2). More than half (60%) of the respondents observed Serb forces removing or destroying personal identification documents. Overall, 88% of participants either saw Serb police or soldiers burning the houses of others or saw the remains of destroyed homes. Furthermore, 28% saw Serb police or soldiers burn the respondent's own home or saw the burned remains of the home. Nearly half (48%) of the sample witnessed Serb police or soldiers destroying property. In addition, 49% reported that Serb police or soldiers had demanded money or valuables from them.
Nearly half (47%) of the refugees saw destroyed mosques, and 39% saw schools that had been destroyed. Overall, respondents reported that 156 different mosques and other places of worship and 242 schools had been destroyed in Kosovo. Participants also witnessed abuses of medical neutrality. Almost one quarter (23%) had seen medical facilities that had been destroyed, 15% witnessed Serb police or soldiers using medical facilities for military purposes, and 12% witnessed Serb police or soldiers forcing medical workers or patients from medical facilities. Overall, participants reported the destruction of 100 different medical clinics, pharmacies, and hospitals.
The 1180 household representatives reported on the experiences of 11 458 household members, including themselves and those who lived with them before their displacement (mean number of members per household: 9.7). Of the 9304 household members for whom data on age and sex were available, 2385 (26%) were adult men, 2689 (29%) were adult women, and 4230 (45%) were children younger than 18 years. Overall, 362 (31%) participants reported at least 1 case of abuse among their household members.
Human rights abuses of household members by Serb forces were reported for April 1998 through April 1999 (Figure 1). Before March 1999, the average rate of reported abuses by Serb forces was 4.4 per week. This rate increased more than 4-fold in the 3 weeks before the NATO bombings and then increased nearly 14-fold over baseline after NATO bombings began on March 24, 1999. Overall, the respondents reported that the majority (59%) of abuses occurred in March and April 1999.
Members of these refugee households suffered a total of 598 incidents of physical abuse by Serb forces (Table 3). The types of abuse included beatings, killings, torture, forced separation and disappearances, threats at gunpoint, gunshot wounds, and sexual assault. These abuses occurred in 23 of the 29 municipalities of Kosovo. By extrapolating the number of abuses reported by participants in our sample to the total refugee population, we estimated that approximately 91 000 (95% CI = 83 700, 98 100) ethnic Albanians suffered such human rights abuses by Serb forces and that as many as 8000 (95% CI = 5800, 10 200) may have been killed.
Our findings indicate that Serb forces engaged in a systematic and brutal campaign to forcibly expel the ethnic Albanian population of Kosovo throughout the province. In the course of these mass deportations, and over the previous year in Kosovo, Serb forces committed widespread abuses of human rights against ethnic Albanians, including killings, beatings, torture, forced separation and disappearances, shootings, sexual assault, looting and destruction of property, and abuses of medical neutrality.
The majority of the survey respondents were expelled from their homes by Serb forces. Participants indicated that forced expulsion was committed by several groups, frequently acting together, including the Yugoslav army, special police, and paramilitary forces and armed Serb civilians.20 Although the regime of Slobodan Milosevic consistently alleged that the ethnic Albanians fled because of NATO bombing, none of the refugees in our study reported leaving their homes in Kosovo for this reason.
This study documented numerous reports of Serbian forces destroying and looting property owned by Albanians throughout Kosovo. More than one quarter of our respondents lost their homes, and more than three quarters saw the homes of others that had been destroyed. Much of this destruction took place in the context of the forced expulsions and appeared to represent a “scorched earth policy” designed to ensure that ethnic Albanians would not return to Kosovo. Specifically, Serb forces engaged in acts that represented an attempt to destroy the social and cultural identities of ethnic Albanians in Kosovo. Cultural and religious symbols, such as mosques, were common targets of violence. Nearly half of the respondents saw places of worship that had been destroyed during their forced migration.
Study participants reported numerous accounts of Serb forces violating medical neutrality, including destroying medical facilities, using medical facilities for military purposes, and removing medical workers and patients from medical facilities. Such abuses effectively incapacitated Kosovo's health care institutions, eliminated its health professionals, and deprived ethnic Albanians of health care. According to international humanitarian law, when combatants engage in armed conflict, they must consider medical personnel, their facilities and vehicles, and their patients neutral and thus immune from attack.21,22
Kosovar refugees provided numerous and detailed accounts of human rights abuses that they and their household members experienced between April 1998 and April 1999. Generally, participants indicated that they were targeted for abuse on the basis of their ethnic and cultural identities. Individuals suspected of having arms or connections with the KLA were targeted for torture. Several accounts of human rights abuses in Kosovo have included reports of rape among ethnic Albanian women by Serb forces.20,23 The frequency of sexual violence and rape reported in this study may underestimate the actual incidence because of cultural stigma and the lack of privacy in most of the interviews. Furthermore, accounts of rape in times of war have been exaggerated or suppressed for political reasons.24
This study documented numerous reports of killing of ethnic Albanian civilians by Serb forces. These killings were part of a pattern of creating fear and intimidation. According to respondents, many of the killings by Serb forces were committed in public places, and witnesses were prevented from removing the bodies for days, leaving other Albanians to contemplate the possibility of a similar fate.
The estimates of human rights abuses and killings provide important insights into the possible extent of abuses suffered by ethnic Albanians. Of note, our estimate of 8000 killed among the population of all ethnic Albanians is similar to the 10 000 estimated by the US Department of State.20 However, the number of killings alone does not convey the nature and extent of human rights abuses perpetrated against ethnic Albanians. Furthermore, accountability for war crimes does not depend on identifying a specific number of people killed. Each intentional killing of a civilian is a war crime, and forced expulsions of civilians represent grave breaches of the Geneva Conventions. There is only an initial understanding of the full extent of abuses suffered by ethnic Albanians in Kosovo.
Our research relied on self-reports, and thus the lack of privacy and confidentiality involved in the interview setting may have led to underestimates of the extent of abuse. Also, the extent of abuse may have been overestimated if refugees judged that it was in their material, political, or psychological interest to exaggerate claims of abuse and loss. Data collected from proxy informants reporting for other members of the household were probably even less reliable, given that these experiences and observations were secondhand.
We used data from refugees in Albania and Macedonia to extrapolate the frequency of abuses to all ethnic Albanians. Because of limitations in sampling caused by the refugee camp environment, these generalizations should be treated with caution. In particular, refugees who resettled among host families or abroad and who may have experienced different human rights abuses were not included in our sample. In addition, refugees who experienced physical and psychological abuses that prevented them from fleeing Kosovo were not included.
Actual frequencies of physical abuse probably varied by the timing and circumstances of displacement. For example, refugees who resettled with host families in Albania and Macedonia may have had the financial and material means to minimize their risk for abuse. Alternately, internally displaced ethnic Albanians were likely to have the greatest exposure to Serb forces, and thus they were likely to have experienced higher rates of abuse.
Despite these limitations, it is clear from this study that until Serb forces departed Kosovo, to be an ethnic Albanian in Kosovo was to be vulnerable to forced expulsion, theft, destruction of property, beating, killing, torture, forced separation from family members, shooting, and sexual violations for no reason other than one's ethnic identity.
| Household Representatives, No. (%) | |
|---|---|
| Location of refugee camp | |
| Albania | 669 (57) |
| Macedonia | 511 (43) |
| Male | 776 (66) |
| Albanian ethnicity | 1159 (98) |
| Muslim | 1151 (98) |
| Time since displacement from home | |
| 2 weeks or less | 279 (24) |
| 3 weeks | 261 (22) |
| 4 weeks | 397 (34) |
| 5 weeks or more | 240 (20) |
| Primary reason for displacement from home | |
| Direct expulsion by Serb forces | 806 (68) |
| Fear of Serb forces | 267 (23) |
| Other reason | 81 (7) |
| No reason given | 26 (2) |
| Type of Human Rights Abuse Witnessed | Household Representatives, No. (%) |
|---|---|
| Killings | |
| Serb police or soldiers killing anyone | 161 (14) |
| Dead bodies | 242 (21) |
| Destruction of personal property | |
| Serb police or soldiers destroying identification documents | 710 (60) |
| Serb police or soldiers burning the homes of others | 587 (50) |
| Remains of destroyed homes of others | 457 (39) |
| Serb police or soldiers burning own home | 186 (16) |
| Remains of own home destroyed | 150 (13) |
| Serb police or soldiers looting or destroying property | 567 (48) |
| Remains of looting or destroyed property | 282 (24) |
| Serb police or soldiers demanding money or valuables | 580 (49) |
| Social and cultural destruction | |
| Places of worship that had been destroyed | 557 (47) |
| Schools that had been destroyed | 456 (39) |
| Abuses of medical neutrality | |
| Medical facilities that had been destroyed | 269 (23) |
| Serb police or soldiers using medical facilities for military purposes | 179 (15) |
| Serb police or soldiers forcing staff or patients from medical facilities | 147 (12) |
| Human Rights Abuse | Household Members, No. (%) |
|---|---|
| Beating | 372 (3.25) |
| Killing | 59 (0.52) |
| Torture | 44 (0.38) |
| Forced separation and disappearance | 33 (0.29) |
| Threatened at gunpoint | 31 (0.27) |
| Gunshot wound | 16 (0.14) |
| Sexual assault | 4 (0.04) |
| Othera | 39 (0.34) |
| Total | 598 (5.22) |
aIncluding car bomb threats, forcing individuals to run, and robbery.
FIGURE 1— Number of abuses per week reported by respondents among household members (n = 11 458): Macedonia and Albania, 1998–1999.
This research was supported by grants from the Blaustein Foundation, the Open Society Institute, the John Merck Fund, the Ford Foundation, and the Edna McConnell Clark Foundation.
A group of individuals with expertise in clinical medicine, public health, bioethics, and international human rights research reviewed and approved this study. All research involving human subjects was conducted in accord with the Declaration of Helsinki, as revised in 1983. Every effort was made to ensure the protection and confidentiality of the participants and to reduce any potential adverse consequences. Verbal informed consent was obtained from all participants, and parental consent was obtained for all participants younger than 18 years.
We are grateful to Linda Cushman, PhD, and Roger Vaughan, DrPH, Center for Population and Family Health, Joseph L. Mailman School of Public Health, Columbia University, for their assistance in the design of the survey instrument and data analysis. Patricia Dawson, Linda Federico-O'Murchu, Aaron Foeste, Shannon Lightner, Carol McMahon, Martine Nokes, Jordana Sander, and Elias Wolfberg provided invaluable research and data entry assistance.
We would like to extend sincere gratitude as well to Luan Jaha, MD, for his tireless efforts on behalf of Physicians for Human Rights over the past year. We thank Susannah Sirkin and Barbara Ayotte of Physicians for Human Rights for reviewing the manuscript. Furthermore, we thank the research interviewers who assisted in data collection (Sebiha Ahmeti, Alban Bakija, Vedat Beshiri, Hasan Bishlimi, Vulnet Dulatahi, Learta Gunga, Lulzim Hasani, Faton Islami, Uran Ismaili, Eriona Minga, Agim Poshka, Ardian Skikuli, and Yllka Spahiu) and the leaders and field staff of Kinderberg, the Organization for Security and Cooperation in Europe, the Mercy Corps, Oxfam, the International Rescue Committee, the World Food Program, El Hilal, the International Medical Corps, Médecins Sans Frontières, the United Nations High Commission for Refugees, the World Health Organization, Tirana Military Hospital, Catholic Relief Services, UNICEF, and the International Committee of the Red Cross. We would also like to thank Benjamin Qylafku.
We are especially grateful to all of the Kosovar refugees who participated in this study despite their suffering.
