Information dated 20.10.1999 A scientific approach to "témoignage" y Alison Marschner MSF finds itself working more and more often in situations with high levels of violence. This has required the organisation to implement new tools and adapt existing ones in order to respond adequately to the needs of the populations with which we work. An interesting example can be seen in epidemiology, a branch of medical science that deals with the incidence and transmission of disease in populations, and one in which MSF has acquired much experience and developed a strong reputation. By adapting its methods of epidemiological survey and analysis, MSF is now able to obtain quantitative information about human rights' violations against populations in situations of violence. The first such survey was conducted in the vastly over-crowded prison of Gitarama, Rwanda, in 1995, and succeeded in drawing the world's attention to the extent of that drama. MSF has implemented the technique on several other occasions since then. More recently, MSF has used the services of Epicentre, an epidemiological research centre based in Paris with which it has developed a close collaboration. The unseen massacres in the forests of Congo Although the return at the end of 1996 of half of the million plus Rwandan refugees from the camps in Kivu (in the Democratic Republic of Congo) were the focus of world attention and figured nightly on our television screens, the other half were ignored or conveniently forgotten - but not by MSF. The results of an epidemiological survey carried out in July 1997 in Ndjoundou, in Bas Congo province, provided proof that they had been systematically massacred. The 3,121 refugees who had reached Ndjoundou (normal population around 500) were receiving food, medical and sanitary assistance from MSF. The stories they told of their 1,500-kilometre journey led to a decision to make a statistical documentation of their experience. A total of 266 families were selected by random sample out of a base of the "enlarged family" units registered for food distributions. The survey took place over a five-day period. Detailed questioning of one member of each family covered background information and the travel history of all members since leaving Kivu. The events documented included deaths, ³disappearances", repatriation and spontaneous return. The most significant fact revealed by the data was that only 17.5% of family members reached Ndjoundou. The others could be accounted for as follows: disappearances: 59.5%; killings: 19.7%; other: 3.3%. Between 1 October 1996 and 31 May 1997, the crude mortality rate was 15.5 per 10,000 persons per day; the comparative infant-juvenile rate was 18.2. The mortality rate, mainly the result of armed attacks, was probably under-estimated as deaths were only recorded if confirmed by the person interviewed. As a result of this survey, MSF was able to back up its témoignage with statistical proof of a kind that could justifiably be used to extrapolate evidence of the fate of tens of thousands of Rwandan refugees in the forest of Congo. This was especially important in view of the Congolese government's refusal to grant free access to refugees to a United Nations human rights investigation team. Mutilations and amputations in Sierra Leone Following a violent upsurge in the fighting in Sierra Leone's protracted civil war at the beginning of 1999, the MSF surgical team in Connaught Hospital in Freetown carried out a one-day systematic survey of patients on 14 February in order to document civilian casualties. A standard and anonymous questionnaire was drafted to identify the cause of each admission. Of the 284 patients interviewed, the study focused on 215 who had been wounded during fighting since the beginning of the year. Of these, 43% were female and 23% were under 15 years old. The results showed that 57.2% had received gun shot wounds, 30.2 had blade injuries. The remainder was suffering from the effects of shells, bombs, burns and other injuries. According to the stories told by patients, 79.1% had been intentionally wounded. These included 63 people who had suffered either full or partial amputations of limbs. As most of the violence could be shown to have occurred between 19 and 20 January, it is very likely that specific orders had been given for these attacks. As statements were obtained only from those lucky enough to survive, it is feared that the level of brutality described by the patients must have resulted in many more mutilations that are unlikely ever to be recorded. This survey offered irrefutable proof that civilians were being directly targeted. Forced Displacement of Kosovars Epidemiological surveys of Kosovo refugees were carried out at the end of April 1999. One was conducted among the 100,000 or so refugees hosted by, or renting accommodation from, local families in Kukes, Albania; another covered Rosaye, in Montenegro, where 3,000 (out of a total of over 10,000 refugees) were living in three disused factories. Apart from providing necessary information to determine the most urgent needs among these people, the intention was also to obtain detailed information on the causes and duration of their exodus from Kosovo and to evaluate the impact of the abuses, physical and psychological, with which they were confronted. A sample of households living in Kukes was obtained using a two-stage random technique (100 apartment buildings were selected and then two apartments from within each of these). Detailed information was obtained for 1,510 of the 1,572 people in the sample. The detailed information received from the 197 households interviewed (average size: 7.7) showed that 70% had left their homes as the result of threats by the Serbian armed forces. At least one family member was reported as absent from 34% of households; 60% of these were males aged between 15 and 54. Out of a total of 141 missing persons, 22% went missing during the exodus and 8% were forcibly separated from their families. Assuming that the approximately 90,000 other refugees in Kukes suffered in a similar way to this sample group, the number of deaths attributed to intentional violence could range from 200 to 600. In absolute numbers, there may be at least 2,000 missing persons among the Kukes refugees whose fate remains to be discovered. A similar number of families were chosen, again by random sample, from among refugees living in disused factories in Rosaye. An analysis of the replies produced comparable results. males aged between 15 and 54. Out of a total of 141 missing persons, 22% went missing during the exodus and 8% were forcibly separated from their families. Assuming that the approximately 90,000 other refugees in Kukes suffered in a similar way to this sample group, the number of deaths attributed to intentional violence could range from 200 to 600. In absolute numbers, there may be at least 2,000 missing persons among the Kukes refugees whose fate remains to be discovered. A similar number of families were chosen, again by random sample, from among refugees living in disused factories in Rosaye. An analysis of the replies received produced comparable results. In addition, almost half of the families in Rosaye had had their identity papers seized before being allowed to leave Kosovo. From these examples, it is clear that the use of proven epidemiological methods to obtain data enables MSF to offer solid statistical evidence to back up its témoignage on behalf of populations in situations of violence.