Wars today are characterised by increased violence against the civilian population. As a result, refugees and internally displaced persons (IDPs) often form unprotected groups which are difficult to access. Once the international aid does reach them, however, the collection of surveillance and survey data is essential to document past abuses, notably those perpetrated during the migration. Collecting testimonies is vital; the validity of which is reinforced by the methods used to attain them and by the quality of the information obtained.
After the civil war restarted in Congo-Brazzaville in December, 1998, a third of the population of Brazzaville fled into the forests of the neighbouring Pool region, where some 250 000 displaced people remained trapped for several months with limited access to international aid. From May, 1999, targeted surveys, and the collection of surveillance and screening data among returnees in Brazzaville enabled the documentation of the health consequences of war on this population.
A retrospective mortality survey registered a mortality rate during the migration of more than five times the alert threshold. Lack of food was a major problem for the displaced, as shown by the proportion of deaths due to malnutrition (50%), and by the prevalence of severe malnutrition among children younger than five years returning to Brazzaville (20%). Further, the 1600 cases of rape reported between May and December, 1999, from the hospitals of Brazzaville highlight the high prevalence of sexual violence directed against women and girls during migration.
In July, 1997, surveys were carried out in Ndjoundou camp in Congo-Brazzaville to document the 1500 km flight of Rwandan refugees through Zaïre during the previous nine months. Researchers found that 82.5% of the initial group disappeared or died during the migration, and that peaks of mortality matched the attacks of the AFDL forces along their journey. These findings support the argument that there was major violence directed against this civilian population.
Similar results were found in Rosaye, Montenegro, in 1999, where surveys have shown that a third of families who fled Kosovo to avoid the exactions reported being separated from at least one close family member - either "left behind" in Kosovo (28%), or "missing" (5%). The programme of attacks launched by Serbian forces was reconstructed through the documention of details from refugees on their villages of origin and the dates that they were forced to flee.
These surveys enabled detailed accounts of the events to which Kosovan refugees were subjected before they left their country, and will contribute to the process of recognition by the international community of the abuses directed against this group.
Most complex emergencies are chaotic. Intervention teams are overwhelmed with work and resources and qualified personnel are limited. Carrying out such surveys can put the refugee populations and the surveyors in danger. In such situations, collection of quality epidemiological data for monitoring or advocacy purposes becomes a challenge. Nevertheless, the work done so far indicates that these surveys are practical and worthwhile in order to quantify violence targeted at civilians. Collecting testimonies is a moral obligation; in some cases it should be a priority, whatever the consequences on the official authorisations, to provide further medical and humanitarian assistance.
Although documenting violence will not do much to help victims of violence in past wars, we are convinced that the documentation of these events will have an impact on the prevention of abuses against vulnerable populations, notably refugees, in the future.