Assessment conducted by Epicentre at the request of Doctors Without Borders/Médecins Sans Frontières (MSF), in collaboration with the Institut Français de Veille Sanitaire

William A Perea

Main Indicators N %

  1. Description of the sample
    • Kosovar households sampled 205 100%
    • Total sample size 1510
    • Average N° of household members = 7.4
    • Ratio Male:Female (15-54 age-group) = 1:1.6
    • Households coming from villages 161 78%
  2. Details of the exodus
    • Directly/indirectly deported by armed forces 138 70%
    • Duration of exodus < 5 days 186 91%
    • Households arriving by walking (70 answers) 8/70 9%
    • Arrived in Kukes > 3 weeks ago 133 65%
    • Households with at least one member absent 70 34%
    • Total number of absent household members 141/1510 8.5%
    • Missed or separated by force 42/141 30%
    • Stayed to fight with KLA 45/141 32%
    • Households reporting at least 1 death by intentional violence
    • Period N°1 : 2/28/99-3/23/99 6 3%
    • Period N°2 : 3/24//99-4/25/99 2 1%
    • Total number of reported deaths for the 2 periods = 16
  3. Hosting conditions and basic needs coverage
    • Households paying a monthly rent 126 61%
    • Average monthly rent = 250 Dutch Marks
    • Average N° of household per housing = 1.8
    • Households not receiving food rations 41 20%
    • Households receiving a complete food ration only once 117 57%
    • Households receiving at least one blanket 46 24% Blankets per family among these 46 = 3.7
    • Households receiving soap 107 52%
    • Households wanting to stay in Kukes 147 72%

1. Introduction

Immediately after the beginning of the NATO offensive in the Federal Republic of Yugoslavia on March 24, 1999, Serb armed forces deliberately provoked a massive population exodus of ethnic Albanians living in Kosovo mainly toward the neighboring countries of Montenegro, Macedonia, and Albania.

One month after the beginning of the crisis the number of refugees having fled Kosovo is estimated at 500,000. Among those 3 destinations, Albania is the one hosting the larger number of refugees. According to the UNHCR, as of May 7, 1999, approximately 400,000 Kosovars have crossed the border into Albania. Currently the only open point on the border with Kosovo is Morina, a locality of Kukes district at northeast Albania. The capital of the district, Kukes town, has become a large crossing-point for refugees and serves as a base of a vast relief operation co-ordinated by the UNHCR. The responsibility for the distribution of food to the refugees is being shared by WFP and the International Federation of the Red Cross.

Doctors Without Borders/Médecins Sans Frontières (MSF) started to operate in the region beginning of April and, in Kukes town, is currently in charge of the management, water and sanitation, and primary health care in the so-called "UNHCR/MSF" camp and is providing primary and secondary medical assistance to the former "Italian 2" camp. Those two camps each contain a population of around 7,000 persons.

Currently several types of refugee settlement can be identified around Kukes town: scattered tractor encampments and tent camps which are the more exposed; collective centers (such as mosques and schools); and host families, which are less apparent but as significant in size. It was evident early on to relief agencies, that the basic needs and the coverage of those needs were going to be very different for each of those groups.

At the request of MSF, Epicentre carried out rapid need assessments in two of the settlements: the tractor camps and the host families. A report on the former group had been already distributed to concerned agencies. The results and recommendations presented in this report concern only the population hosted by Albanian families in the town of Kukes. The assessments were conducted in close collaboration with the French National Public Health Centre (Institut de Veille Sanitaire).

[Food and non-food items have been distributed since the survey was conducted; therefore data on distribution might by out of date. However, the information on human rights abuses, the makeup of the families, and the general problems they are facing as well as the overall recommendations are still relevant.]

2. Objectives

  • To provide information on demography, basic needs and basic needs coverage of Kosovar refugees hosted by Albanian Families in Kukes town
  • To estimate the proportion of Kosovars living in host families in Kukes town (excluding collective centers and camps)
  • To obtain detailed information on the causes and duration of the exodus from Kosovo

3. Methods

A sample of households living in Kukes was obtained using a two-stage random sampling method. A sample of 100 apartment buildings and houses were randomly selected from a map depicting all existing physical structures in the town. Each structure selected was visited and whenever a building with several apartments was found, a sample of two apartments was randomly selected.

The number of local Albanians presently living in the apartment, as well as the number of Kosovars, if any, was systematically recorded. A standard questionnaire containing information on date of arrival, family composition and host conditions was addressed to the head of the Kosovar household. A household was defined as the members of the family that lived under the same roof before the beginning of their exodus. A separate questionnaire was filled for each Kosovar household whenever there were more than two occupying the same apartment.

For each household we recorded the details of their exodus, the number of missing members, together with the numbers and causes of deaths that occurred during two periods of time: (i) between the February 28, 1998 (a Muslim holiday) and March 24, 1999 (the beginning of NATO offensive), and (ii) between the March 24, 1999, and the date of the survey.

Details about relief aid such as food and non-food items and the perceived priority needs of the households were also obtained. For the question on food items, only the delivery of a complete food ration was taken into account. We did not consider various items distributed by trucks on a first-come-first-served basis (fish/beef cans, cookies and sometimes fresh vegetables).

4. Results

4.1 General Information

The survey was conducted between the April 25 and 27. Altogether, 195 housing structures (apartments or houses) were visited. In 119 (61%) of these structures there were 213 Kosovar households, in 59 (30%) there were no Kosovars, 2 (1%) were rented to foreign journalists, and 13 (7%) were empty at the time of the visit. In two apartments (1%) the occupants refused the interview.

Of the 59 Albanian families not hosting Kosovars at the time of the survey, 30 (50%) reported having done so during the month preceding the survey.

The total number of Kosovars and Albanians in the sample was 1572 and 807 respectively (ratio 1.9:1). The average number of persons among the Kosovar household was 7.4.

4.2 Description of the sample of Kosovar refugees

Detailed information was obtained for 1510 of the 1572 persons of the sample (205/213 households). The distribution by age and sex is shown in Table 1. Females accounted for 55 % of the sample (sex ratio M:F = 1:1.24). However, the proportion of females was higher among the 15 to 54 year age group (62%, sex ratio =1:1.6).

Table 1. Age and Sex distribution of a sample of 1510 refugees. Hosting families survey, Kukes town, Albania. 27/04/99

Sex 0 - 4 (%) 5-14 (%) 15-54 (%) > 55 (%) Total
Male 98 (15) 230 (34) 284 (42) 62 (9) 674 (100)
Female 97 (12) 198 (24) 471 (56) 70 (8) 836 (100)
Total 195 (13) 428 (28) 755 (50.0) 132 (9) 1510 (100)

The large majority of the households interviewed (161/205; 78%) lived in villages before their exodus. Those villages were generally located around Prizren, Djacovtica, Dragash and Suarekha in South Kosovo.

4.3 Details of the exodus

Detailed explanations concerning the causes of the "departure" from Kosovo were obtained from 197 families. One-hundred-thirty-eight (70%) households said they were directly or indirectly threatened by Serb armed forces to leave their houses. Eleven families (6%) were compulsorily sent (in collective buses) or escorts to the Albanian border by armed forces, and 45 (23%) left because they feared Serb aggression.

For 186 (91%) of the families, the exodus to the Albanian border lasted less than 5 days (median = 2 days). Nine families (4%) reported having left their villages of origin 8 to 12 months before their arrival to Albania "because of the war" and had moved into relatives' houses in other towns and villages of southern Kosovo. The rest (5%) spent 1 to 4 weeks either hiding in the mountains or in relatives' houses in their way to Albania.

Information on the means used to reach the border was only obtained for 70 families. Of these, 62 (91%) reported having came by motor vehicle (tractor and/or car) and the other 8 (9%) had walked to the border.

All the 205 families except 2 (99%) arrived after March 24, 1999. The median number of days since the arrival in Kukes was 24. The number of families arrived during the 3 to 4 weeks preceding the survey was 133 (65%) while the rest rrived one to two weeks before (7% and 27% respectively).

4.4 Retrospective mortality

A total of 16 deaths were reported between February 28, 1998 and the date of the survey with seven of them (43%) occurring after the beginning of NATO offensive. Nine (56%) of the 16 deaths were explained by acts of violence conducted by Serb police or military forces (7 male adults and two children less than 10 years old), and one occurred during the exodus to Albania (elder).

The crude mortality rate per 10,000/day and per period are presented in Table 2. The mortality rate attributed to violence after the NATO offensive was 0.6/10,000/day [95%CI: 0,16 - 1,9]

Table 2. Retrospective crude mortality rate among a sample of 1510 Kosovar refugees. Host families survey, Kukes town, Albania February 28, 1998 - April 25, 1999

Period N° of deaths Crude mortality rate/ 10,000/day [95% CI]
2/28/98 - 3/23/99

(391 days)

9 0.15 [0.07 - 0.3]
3/24/99 - 4/25999 7 1.4 [0.6 - 3.0]
(33 days)

4.5 Absent family members

Seventy households (34%) reported at least one member absent (not in Kukes) at the time of the survey. The total number of absent persons was 141, 60% of whom were males between 15 and 54 years old. This figure represented 8.5% of the sampled population [141/(1510 + 141)]. According to the respondents, 31 persons out of the 141 (22%) became "lost" during the exodus and 45 (32%) stayed in Kosovo (or went back) to join the KLA forces (see Table 3).

Table 3. Reported reasons for the absence of household members. Host families survey, Kukes town Albania, April 27, 1999

Reason N (%)
"Lost" during exodus 31 (22%)
Separated by force 11 (8%)
Stayed voluntarily 29 (20%)
KLA fighter 45 (32%)
Others 25 (18%)
Total 141 (100%)

4.6 Hosting Conditions

When asked about their hosting conditions, 126 households (61.5%) said they were paying rent, 40 (20%) said they were hosted for free, and the rest (18.5%) were hosted by relatives. The mean rent paid by families was $137 (US) per month (ranging from $36 to $441).

The average number of households per housing unit was 1.8 (ranging from 1 to 6), and the mean number of Kosovar refugees per housing unit was 13 (ranging from 1 to 53).

4.7 Coverage of essential needs

None of the households interviewed had a registration card or any other document identifying them as refugees. The large majority reported having been visited at home and registered on a list. It was not clear though whether the visitors were administrative Albanian officers or "Mother Theresa" affiliated members.

Concerning food, 41 families (20%) reported not having received any food ration since their arrival in Kukes. Among the other 164 families, 117 had received a food ration only once (see Table 4). This food ration consisted generally of parcels distributed by the Albanian Red Cross. The number of food parcels received per family varied from 1 to 5 according to the family size. The reported mean duration of the food ration received was 4.5 days (oil and yeast aside). Bread was the only food commodity received systematically (usually daily) by the families interviewed.

The number of families in the sample that received at least one blanket was 46 (24%). For these 46 households, the average number of blankets received per family was 3.7. None of the families had received mattresses. About half of the households (107) had received some soap, and a small percentage (2.5%) said they had received occasionally non-food items such as clothes, pampers, and hygienic commodities.

Table 4. Coverage of basic needs among a sample of Kosovar households. Host family survey. Kukes town, Albania, April 27, 1999

Basic Needs Indicators N° of Households N=205 (%)
Not receiving food rations since arrival 41 (20%)
Receiving a complete food ration only once 117 (57%)
Receiving at least one blanket since arrival 46 (24%)
Receiving soap 107 (52%)

Perceived priority needs were similar for the majority of people interviewed: 135 out of 191 households (71%) mentioned food, blankets and clothes as their main concern, 39 (20%) only mentioned food, 11 (6%) only mentioned clothes and blankets, and the rest (3%) mentioned miscellaneous items such as money and hygienic items.

4.8 Immediate Plans

In response to the question about the household's immediate plans, the majority (72%) answered that they wished to stay in Kukes until the conflict in Kosovo was over (see Table 5). As many as 12% wished to be relocated in a third country (usually in Europe) where they have close relatives, while only a minority (1%) had relatives in Albania with whom they could move in the near future.

Table 5. Immediate plans of Kosovar households staying in Kukes (N=205). Host family survey, Kukes town, Albania; April 27, 1999

Plans N (%)
Stay in Kukes 147 (72%)
Leave Kukes for another location in Albania 11 (5%)
Join relatives in other locations in Albania 2 (1%)
Join relatives or be relocated in a third country 25 (13%)
Household head wants to go back to fight in Kosovo 13 (6%)
Don't know what do 6 (3%)

5. Discussion

The results of this survey are representative of the refugee population lodging in houses and apartments in Kukes town. Considering that the Albanian population of this town is estimated at 20,000 to 25,000 (figure given by local authorities), the number of Kosovars lodged in Kukes can be estimated at 40,000 to 50,000.

In general, refugees arriving in Kukes may go immediately farther south, or decide to stay in town. In Kukes there are four "categories" of refugees: those settled in collective centers, those staying in scattered tractor encampments, those "accepted" in tent camps, and those staying with host families. With hundreds, and up to thousands, of new arrivals to, and departures from, Kukes every day, the total number of refugees in the town at a given point in time can only be roughly estimated.

Table 6 shows rough estimates of the various groups of refugees in Kukes town as of April 25, 1999. According to these figures, refugees lodging in town represented more than 50% of the total refugee population. This and the fact that the group studied is also the most stable and will probably stay the longest in Kukes, are elements that must be taken into account when planning relief operations in the region.

Table 6. Approximate population figures of refugees settled in Kukes town as of the April 25, 1999.

Type of Settlement Approximate numbers Capacity
Collective centers 20,000 1 25,000
Tent camps 16,000 2 30,000
Scattered tractor encampments
(Potato factory and Zabel)
5,000 3 ?
Refugees lodging in town 45,000 3 ?
Total 86,000  

1 Source ADRA 2 Source UNHCR 3 Source MSF

Causes and Consequences of the Exodus and Human Rights Violations

The results of this survey as well as refugee accounts of the exodus to Albania fully support previous reports on serious human right violations committed against ethnic Albanians by Serb forces in Kosovo. Close to 70% of the families were either physically and/or verbally threatened by armed militias or military forces, or terrorized by systematic acts of vandalism including burning and shelling of their villages, and finally compelled to depart immediately from their homes. (Families report that they had to leave within 5 to 15 minutes or else risk their lives.)

The excess of crude mortality observed since the March 24 and the fact that more than half of the deaths reported were caused by intentional violence underlines the seriousness of the situation. Assuming that the rest of the refugee population in Kukes (~ 90,000) suffered the same type of aggression, the number of deaths attributed to intentional violence among this group could range from 200 to 600.

Also alarming is the fact that close to one-third of families reported at least one member of the household absent. In absolute numbers our results suggest that, among those refugees staying in Kukes, there may be at least 2,000 missing persons, and that another 2,000 to 3,000 were left behind. More than 60% of the absent family members are male adults. This implies that women and the elderly head a significant number of households. This is clearly supported by the strong under-representation of men 15 to 54 years old in our sampling.

Basic Needs Coverage of Aid-Relief Operations

According to the results of this study, relief aid has not appropriately reached the refugees lodged in Kukes town. In the absence of a proper system for registration, refugees can not be identified as such. As a consequence, it is impossible to ensure that each family receives the minimum goods necessary to cover its basic needs, such as complete food rations, blankets, clothes and soap (and plastic sheeting and water containers for refugees sleeping outdoors). As a result, families are spending their lean savings purchasing food and other basic items for daily subsistence. This is aggravated by the fact that at least 60% of them are also paying high rents.

The above problems are, unfortunately, not limited to this group of refugees. A rapid assessment conducted in two tractor camps led to similar conclusions.

The permanent arrival of refugees, their mobility, and the multiple and diverse number of settlements existing in the town are factors that make difficult the delivery of aid. However, there is no such a thing as scarcity of aid in Kukes or major logistic constraints hampering it. Trucks delivering all kind of items are seen daily in the streets of the town and the warehouses of the agencies in charge of distribution of commodities are full.

That such a situation exists in the midst of a massive and without-precedent mobilization of the international community is unacceptable. Since a large sector of the refugee population in Kukes is not benefiting from relief aid, it should be acknowledged that the current aid mechanisms are not efficient enough. The bottleneck begins with the lack of a system for clear identification of the target population as well as a poor understanding of the basic needs that should be covered as a priority for each category of refugees.

Current Situation and Perspectives

The overall health status of the refugees is acceptable for the time being (see Annex 1 for the Ministry of Health surveillance epidemiological report). No energy-protein malnutrition cases were observed, and aside from some sporadic cases of measles no other cases of potentially epidemic diseases have been reported in the town.

However, with the price of food, basic commodities, and rents rapidly rising, the refugees will soon run out of money. Increasing sectors of the population will start relying exclusively on external aid, and the health status of the most vulnerable groups, i.e., children and the elderly, may progressively deteriorate. Should the coverage of basic needs not be rapidly improved, this grace period would not last much longer.

The significant proportion of elderly among this population (8%) and the prevalence of chronic diseases such as diabetes and cardiovascular disorders should be carefully considered by health agencies since the list of essential drugs do not always allow adequate case management of those conditions.

The reason that the majority of refugees don't want to move from Kukes in the near future may be explained by the striking lack of information the refugees have received on vital issues such as the locations and size of settlements in the south of the country (or how to reach them), assistance available for finding missing family members, and existing procedures for those wishing to join relatives in a third country. There is a strong feeling of uncertainty among the people interviewed that is not being adequately addressed.

Concerned about the proximity of Kukes to the border of Kosovo (30 km) and the risk of an extension of the conflict into this region, the UNHCR and the Albanian authorities have expressed their intention to move all of the refugees from the town of Kukes to the camps in the south, leaving the town as a mere transit site. However, the continuous influx of refugees into northern Albania, the inevitable saturation of other sites throughout the country, and the fact that many refugees are already "installed" and are not willing to move outside Kukes town, mean that it will be difficult to move these refugees. So far there is no solution for the families running out of money. In any case, all groups of refugees require more information about what sort of accommodation they would be moved to, where, when, how, and so on.

Moreover, an episode in the tractor camp has shown that the re-location of refugees may be difficult to handle and could create the grounds for violations of the refugees' basic rights. On April 21, more than 3,000 people staying in the tractor camp around the Mosque were compelled by the Albanian police to empty the camp and to leave immediately to the south in buses made available by the Albanian authorities. Since no reasonable solutions were proposed to the owners of the tractors, men decided to stay behind. The methods used caused the splitting of families and strong psychological disarray among the population. The UNHCR, the agency in charge of refugee protection, should closely monitor the process of re-location toward the south to prevent such deplorable incidents from occurring again.

While searching for a durable and acceptable solution for those who have been or will be in Kukes for long, the concerned agencies and donors/governments should make all possible efforts urgently and appropriately to address the basic needs of these populations, including protection. To fail to do so in the midst of a plethora of aid and multiple aid-relief agencies, will be regrettable.

6. Recommendations

  1. The following measures should be undertaken urgently to ensure that the basic needs of refugees, old as well as new arrivals to Kukes, are appropriately covered:
    • Implement as soon as possible a simple registration system in order to ensure proper access to relief aid. This essentially means giving a registration card to each family already arrived or arriving in Kukes.
    • Ensure systematic weekly distributions of food parcels containing recommended energy and protein daily requirements in cold weather (2300 kcal/person/day).
    • Distribute a minimum of one blanket per person.
    • When possible, distribute to each family 2 mattresses as well as clothes for adults and children.
    • Insure the regular distribution of hygiene commodities such as soap, toothpaste and toothbrushes, sanitary napkins, and diapers.
    • Improve the refugees' access to information through leaflets and continuous messages on radio and television. The identification of a site or an office where people could be advised and oriented according to their needs may substantially improve access to key information and allow the refugees to participate more actively in the search for solutions to their problems.
    • Broadcast publicly a well-defined strategy for the re-location of refugees currently settled in Kukes farther to the south of Albania. Avoid the use of force, coercion, or intimidation before or during the process of re-location.
    • Document and immediately report any measures that violate refugees' basic human rights. Events like the forced displacement of tractor camp residents from Kukes should not be allowed.
    • Encourage a dialogue with the refugees, such as the one initiated by UNHCR through the "Blue River Strategy". Those initiatives should be multiplied and continued until consensual solutions are agreed upon.
    • Improve the visibility and strengthen the existing program for the search of missing family members.

    The above recommendations also apply to other categories of refugees and particularly those who remain in the tractor camps.

  2. A definitive solution to the current crisis may be several months away. As time passes, the needs of the refugees will evolve and the agencies in charge will have to adapt their activities to respond to new priorities. The following are some elements to be considered in the short and medium term:
    • Continue the implementation and strengthening of the epidemiological surveillance system proposed by the Ministry of Health/WHO/French National Public Health Center. This should be a priority for all agencies implementing health activities, as the potential risk for the appearance and rapid spread of infectious-disease outbreaks in Kukes is very high.
    • Carefully consider the prevalence of chronic health disorders, particularly diabetes, when planning an essential drug supply.
    • Implement a mortality surveillance system based on grave counts and hospital deaths. An initiative attempting to count graves was launched by UNHCR several weeks ago. This initiative needs to be maintained and a systematic data collection system promptly implemented.
    • Implement vaccination against measles and polio for all new arrivals, and continue the EPI strengthening at local health structures.
    • Extend the mental health program initiated in some of the tent camps to cover the population lodging in the town.


Beyond formalities, four persons deserve all the credit for the work done: Gadgick, Zabit, Filoreta and Hakim, all of them refugees themselves. The exercise they had to go through was a permanent mirror of their own distress and I can do nothing but to express my deepest respect for their understanding and implication on the study. I hope their work will contribute to the improvement of the situation of their and all Kosovar families. Thanks also to Marta's invaluable help during the whole survey and to Kristine and Ian who allowed the assessment to be finished in such a short time frame. The map and photos are the result of the restless and ingenious mind of Denis.