An estimated 800 000 people have been affected by the war in Congo-Brazzaville. Thousands of civilians have been killed.

In December, 1998, after a year of temporary respite, fighting returned to the Congolese capital, Brazzaville, and caused the displacement of nearly a third of the city's inhabitants.

Most of the displaced sought shelter in the forests in the region of Pool, where their numbers grew as fighting spread across the region.

Here they were finally trapped, without food or medical assistance. (see footnote 1) The first displaced returned to Brazzaville in May, 1999, with more following up to February, 2000. In total, 250 000 people arrived at the Brazzaville Centre Sportif, which served as a transit centre at which the displaced were registered. A fifth of these people originated from Pool and had fled their homes in search of aid in the Congolese capital.

Médecins Sans Frontières (MSF) began providing assistance to the populations in February, 1999, first in the northern districts of Brazzaville, and then, when the displaced began to return in May, in the southern districts that were secured by government

forces. Four therapeutic feeding centres were set up and medical screening and medical referral was carried out by MSF in the Centre Sportif. In October, 1999, 10 months after the onset of violence, aid agencies were able to access the south of the country.

MSF began working in Pool--in Kinkala in November, 1999, then in Mindouli in February, 2000. We have collated the findings of qualitative and quantitative assessments to assess the health consequences of the war on the displaced population and the residents of Pool.

A survey of mothers of malnourished children, carried out in October, 1999, in the MSF feeding centres in Brazzaville, enabled us to retrace the movements of 191 displaced families, including 1033 people.

Among these families, 150 (79%) had been displaced from Brazzaville since December, 1998, when the fighting restarted. 41 (21%) originated from Pool and had sought shelter in Brazzaville because of the poor living conditions in Pool.

The median duration of flight was 8 months. 89 families (47%) returned to Brazzaville because of health problems (malnutrition and illness) and 60 (31%) after hearing that the situation in the capital had improved. 109 (57%) of the 191 families reported having been held in the region of Pool against their will, and 124 (65%) said that they had been robbed. In total, 92 (48%) families reported the death of at least one family member since December, 1998.

139 people had died (13.5 % of initial sample) of which 48 (35%) were children aged under 5 years. The main cause of death was malnutrition. A retrospective mortality survey was done in May, 2000, in the town of Mindouli among the town's 10021 inhabitants.

Between November, 1999, and April, 2000, 736 people were reported to have died, comprising 6Ã?·8% of the initial population. 205 were under 5 years of age. The overall daily mortality rate was more than five per 10 000 people between November, 1999, and January, 2000. After this date, it decreased and normalised in April, 2000, to less than one per 10 000. The daily mortality rate in children under 5 years exceeded ten deaths per 10 000 between December, 1999, and January, 2000.

More than half of all deaths were from malnutrition. Between May and December, 1999, 1600 women and girls who were admitted to the hospitals of Makelekele and TalangaÃ?¯ in Brazzaville said they had been raped. Most rapes occurred on the route linking Kinkala and Brazzaville, named the "corridor of death" by the survivors. The signing of the peace agreement in November, 1999, did not stop these crimes from occurring.

In March, 2000, 22 rape victims were treated at Makelekele hospital, aged between 3 and 40 years old (nine of the victims were under 15). 21 of the rapes were committed by armed men and 13 by at least two people.

In one case, the number of rapists was eight. After several months of negotiations with the Congolese ministry of health, MSF was given permission to open a medical programme for victims of sexual violence, including the provision of antiretrovirals. This programme continues today.

Between August and September, 1999, the prevalence of severe acute malnutrition among under fives screened at the Centre Sportif was between 30% and 40%. The prevalence of global acute malnutrition was greater than 20%, of which 75% were cases of kwashiorkor.

8061 children were admitted to the four therapeutic feeding centres in Brazzaville between May, 1999, and February, 2000, of which 286 died. We have been unable to calculate with precision the number of people who disappeared or died between December, 1998, and January, 2000.

Government sources have estimated that the conflict has caused the death of at least 10 000 to 15 000 people.The numbers presented here cannot be extrapolated to the total displaced population, but suggest that the actual number of deaths is much higher.

The conflict in Congo-Brazzaville, like those of Kosovo, Timor, or Sierra Leone has been characterised by repeated violence directed at the civilian population.

The displaced, forced to flee their homes, found themselves trapped and used at times as human shields. The stories brought back of murders were numerous and rapes were committed on a massive scale and at times systematically.

Malnutrition was the principal cause of death among the displaced. A third of children seen by doctors at the Centre Sportif in Brazzaville had global acute malnutrition. In total, more than 10 000 cases of acute malnutrition were treated in MSF's feeding centres.

This figure does not take into account the medical activities of other aid organisations present in Congo in 1999. The violence that ravaged Congo-Brazzaville between December, 1998, and January, 2000, unfolded behind closed doors and was met by general indifference from the international community.

The systematic documentation of violence targeted at civilians in war is essential to ensure that such abuses do not go unnoticed and are not forgotten, an important step in advocating for increased humanitarian protection and access for aid agencies.