A / A / A


Date de publication
Judith Soussan

Graduated in International relations (Institut d'Etudes Politiques de Paris), Logistics in humanitarian settings (Bioforce-Développement) and Anthropology (University Paris I), Judith Soussan joined MSF in 1999. After missions in Sri Lanka, Ethiopia, Sudan and the Palestinian territories, she worked in Paris, in particular on protection of civilians. Following a few years break from MSF working as a radio reporter, she has come back to Crash in 2015. She contributed to the book "Saving lives and staying alive. Humanitarian Security in the Age of Risk Management" (chapter "Security Issues and Practices in an MSF Mission in the Land of Jihad" - London: Hurst and Co, 2016).

PROLOGUE - 1996, BEFORE THE ADVANCE OF THE AFDL: MOUNTING CONCERN                                                                                                   


Assisting the refugees                                                                                                

Total loss of access and a call for armed intervention                                                    

The immediate consequences of the call for intervention; support for intervention runs out of steam                                                                   

Outline of the respective responsibilities of the parties involved during this episode where protection is concerned                                                      


First doubts about perverse effects, first large-scale operations – November-December 1996      

Large-scale operations in camps and the appalling

situation of the refugees – December 1996–January 1997                                             

Assistance as protection of refugees?                                                                           

Resumption of the flight of refugees and with MSF at their heels – February-March 1997       

  1. - THE RADICALISATION OF POSITIONS APRIL 1997                                  

The dilemma of whether to repatriate exhausted refugees or provide care on site – April 1997      

Communication strategy concerning the massacres – April 1997                                    

Issues disputed in the discussion of responsibility in the face of the massacres                  

Crystallisation of the sections’ positions: loss of access, massacres in present time,

disagreements over whether to speak out - late April                                                    

Through disagreements, the details of MSF-France’s position emerge more clearly        


Disagreement over a range of bad solutions                                                                   

Some aspects of content, or how to protect the refugees                                               


From the consolidation of information on the massacres to the grim outcome of repatriation…     

… A story that stops without coming to an end                                                            


To delimit MSF’s responsibility, in its own eyes, for providing protection at the time when the Rwandan refugees in Zaire were being hunted down and killed, we need to consider the various discourses relating to MSF’s role when confronted with violence – both the discourse of its public statements and that of internal debates which manifest a certain idea of our responsibility towards ‘endangered populations’, and towards the Rwandan refugees in particular.

We also need to consider the diverse decisions and actions taken by the different components of MSF throughout the many episodes of this exceptionally serious and large-scale crisis. It will be a matter, as we will see, of analysing the concrete details of delivering assistance – since the concern for protection arose not as a ‘side’ issue (i.e. in practices only marginally related to the relief effort), but in the very act of delivering care – in a context of such complexity that the relief effort became, by turns, an instrument for saving masses of people on the verge of death, a tool for tracking down refugees in order to eliminate them, and an insignificant, futile action.

In order to do justice to the complexity of the situation and the significance of the various positions adopted by MSF over time, the form selected for this account is a chronological sequence in which breaking points and prevailing rationales that define the various sections are identified. This case study is based almost exclusively on The Hunting and Killing of Rwandan Refugees in Zaire-Congo: 1996-97, by Laurence Binet.L. Binet, The Hunting and Killing of Rwandan Refugees in Zaire-Congo: 1996-97, MSF International/CRASH-FondationMSF-F, coll. “MSF speaking out”, 2004, 266 pages (internal document available on approval at bcescau@msf.org). The minutes of Board of Directors’ meetings and presidents’ annual reports were also consulted.Note: double inverted commas (“ ”) indicate a quotation; single inverted commas (‘ ’) are used by the authorAbbreviations and terms: AFDL: Alliance des forces démocratiques pour la libération du Congo (Kabila) (Alliance of Democratic Forces for the Liberation of Congo) – FAR: Forces armées rwandaises (prior to June 1994) (Rwandan Armed Forces) – Banyamulenge rebels, ‘Tutsi troops’: AFDL - IHL: international humanitarian law – HCR: United Nations High Commission for Refugees – WHO: World Health Organisation.  MSF jargon: sitrep: situation report – PR: press release – AR: President’s annual report – DirOp: Operational Director – DirCom: Communication Director – Ex.Dir: Executive Director – Field co: Field Coordinator – MSF-B: MSF-Belgium – MSF-F: MSF-France – MSF-H: MSF-Holland – MSF-US: MSF-United States – IC: International Council (of the presidents of MSF sections) – Back-up section: section responsible for coordinating the activity of the emergency team (ET) – task force: working group dedicated to a specific emergency operation - HAD: Humanitarian Affairs Department (MSF-Holland) Before the Alliance des Forces Démocratiques pour la Libération du Congo (AFDL) began its advance in October 1996, MSF-Holland was the only MSF section present in Kivu, the eastern region of Zaire.

In early 1996, when tensions in the region were rising between Hutus and Tutsis speaking Rwanda language (Kinyarwanda), the Tutsis being particularly in danger, the tendency at MSF- H was to get out of the camps. The new coordinator heard rumours of massacres in the Masisi area: “I said, ‘OK, let’s have a look at it and get some information’. Indeed, there did seem to be a lot happening”.

The team began to “move more around in the area”, collecting information that, though fragmentary, sufficed to show that violence was occurring and that more serious violence might be imminent (palpable tension, wounded patients, villages that at times were completely surrounded and whose inhabitants were terrified): worry over the fate of the people, which had subsided, was back in full force.

Owing to its growing perception of an unspecified menace hanging over a certain group, the team on the ground tried to alert other parties: this took the form of an “information and advocacy campaign” by the coordination teams in Goma and in Rwanda (embassies, aid agencies) as well as efforts to involve the desk, “but this looked rather like silent diplomacy” (MSF-H coordinator). As “the concern for these people” grew steadily amongst the MSF team, it called for preventive action to move the Tutsis out of danger, appealing to other stakeholders’ sense of responsibility: discussions with other agencies were organised “initiated by MSF and ICRC, to try to get something done to protect these people”; during one such discussion, MSF and the ICRC drew attention to the situation of the Tutsis of Mokoto and Kichanga, “stressing that in [their] view these people needed to be evacuated” (sitrep, MSF-H coordinator).

When the potential threat became a reality (the massacre of Tutsi Banyarwanda in a church on 12 May 1996), this concern was transformed into an absolute necessity: “we decided that since we now had so much evidence of things, we had to do something about this” (interview, MSF- H coordinator).

This responsibility to “do something” gave rise to a series of simultaneous or successive activities. In the field, it led to the deployment of medical assistance to the survivors: “MSF’s first response was the dispatch of a surgical team to care for the wounded”; the coordination team in Goma held discussions “to talk UNHCR into taking responsibility for the evacuation of the remaining 5,900 Tutsi” (MSF internal magazine). When it became apparent that the HCR would not evacuate the survivors, a public statement (press release of 21 May 1996) was issued: its purpose was to put pressure on the HCR’s headquarters, and its form was that of a warning: “the idea was to WITNESS not to DENOUNCE”, by “expressing concern about the Tutsi and calling on the relevant UN bodies to evacuate them”. At the same time, in an unusual move, MSF itself evacuated some of these displaced persons.

Here for the first time we see the link between, on the one hand, the clarity of the information available (i.e. not only information received but also, and most importantly, as we can see in this case, information collected, since concern for the Tutsis did not take shapeuntil the new coordinator took the pro-active decision to go out and collect information) on the violence (past and to come) and, on the other, the perception of an urgent need to act, and indeed, of MSF’s responsibility to take action.

This sense of responsibility is connected, in a confused manner, with the conviction that violence was both probable and imminent, and with the awareness that MSF possessed information indicating it. This explains why the specific content of this responsibility ranges in this case from giving the alert to advocacy to actual evacuation of people.



Events picked up speed dramatically in October 1996 with fighting between the Zairian   army and the Banyamulenge rebels of the AFDL in Kivu (Bukavu was taken on 29 October), and, in conjunction with this fighting, attacks by the rebels (with the help of Rwanda) on the camps of Rwandan refugees: first the camps in the Uvira area (13-20 October), followed by those in Kibumba and Katale on 26-27 October and that in Kahindo on 30 October. A tide of refugees began to converge on Goma. On 28 October, the press was already speaking of a “desperate humanitarian situation” where “500,000 refugees take to the roads”. The situation was extremely confused, with tens of thousands of Rwandan refugees fleeing the area, Zairians fleeing before the advancing ‘Tutsi troops’, the chaos of the fighting itself, etc.



In just a few days, the emergency situation arising from a massive displacement of the population was transformed into a context of war, characterised by violence, confusion, volatility and mobility.

The response of the international organisations was to sound the alert (a “cry of alarm” from UNICEF) and call for the establishment of “humanitarian corridors” (UNHCR, 27 October). The WHO sent an emergency team to assess “the epidemic risk and the measures to protect the hundreds of thousands of people left without assistance”. As for “the international community, [it] is currently considering only diplomatic solutions” (Le Monde, 1 November and 31 October).

MSF’s first action concerned the provision of relief: it helped to “offset up emergency clinics” (PR MSF-F, 29 October) in the Mugunga camp, near Goma, a gathering point for hundreds of thousands who had fled from other camps.

Faced with, on the one hand, the prospect of no longer being able to help the endangered population (risk of being forced to evacuate), and on the other, the announcement that the international community would take no action, MSF issued a public statement putting forward the notion of protection and invoking the responsibility of the other parties involved: MSF-F called for a protected assistance zone for the Rwandan refugees and Zairian civilians caught in the conflict and denounced the passivity of the international community: “the international community should focus on creating a safe space where civilians would have access to aid” (PR MSF-F, 31 October).

It should be noted here that the notion of a protected area has a very specific meaning: that of a demarcated geographical area. The direct justification given by MSF for the establishment of such an area was the need to deploy humanitarian assistance; the violence was mentioned only secondarily, along with the other dangers to the refugees: “in Goma, Bukavu and Uvira, where aid operations can no longer be conducted, more than 2 million people – refugees and Zairians – now face death, malnutrition and epidemics”. The priority was the complete vulnerability of these people in flight, and the response would necessarily combine security measures (and hence a protected area) and assistance.

On 1 November, the situation became still more difficult for the relief workers, who lost access to the area due to the fighting: evacuation of the Mugunga camp and of the projects in South Kivu and Gisenyi (Rwanda).

On this occasion, a field team once again took the initiative of physically evacuating people: “MSF staff in Gisenyi has been able to evacuate all children from their nutritional centre at the Umubano at the Goma border”, in lorries, to another camp “where the supplementary feeding programme can be continued” (PR MSF-H, 1 November). This action reflects a sense of responsibility to our patients (and all the more so to an “extremely vulnerable group of children”): a responsibility not to abandon them on leaving, to bring them to safety if possible. Apparently, this decision did not give rise to a debate.



The situation worsened further on 2 November with total loss of access. The NGOs were forced to evacuate Goma, which had fallen into the hands of the rebels. This loss of access seems to have caused the decision to call for “safe zones to be created to protect populations in danger in Zaire and to ensure that aid can be provided” (PR MSF-F, 4 November); failure to intervene would run the risk of “a repetition of the 1994 catastrophe when a delayed humanitarian action replaced effective protection of the population”. “The insecurity and chaos in Zaire is so bad that there is nothing doctors and bandages or any humanitarian assistance can do” (PR MSF-H, 4 November). This same press release mentioned the need for “an effective relaunch of the repatriation process”.

What was the process leading to this decision to call for armed intervention? The importance and rarity of such an initiative (taken for only the second time in the history of MSF) make it worth dwelling on. First, “everybody agreed that we had to say something” (S. Bolton, communication officer for the Great Lakes region) but did not necessarily agree on what should be said: “humanitarian corridor”, “intervention or no intervention?”, etc. It should be noted that the context was one of urgency and lack of information: “We were in Kigali and the pressure was incredible. We had no information about anything” (ibid.). “… The idea was that we couldn’t just stand by and do nothing. The international community had to provide protection. (…) Morally we couldn’t just stand by” (P. Biberson, President of MSF-F).

MSF-Holland was the section that argued the most strongly for intervention, based on the reports from the ground: “the last eyes and ears of the international community have left and given the (…) context this crisis will develop into a massive bloodbath” (J. de Milliano, President of MSF-H, quoting field personnel). As it was a matter of “predictable massacres”, the debate was couched in extreme terms: shall MSF merely “be the followers”, or shall we “choose to contribute to prevent the massacres of the civilian populations by raising our voice (…)”?, asked J. de Milliano. The President of MSF-H also emphasised that it was necessary “to measure and accept certain downsides” of such a decision: failure to achieve the desired result “is not a reason not to try, given the imminent catastrophe”.



The weeks that followed the call for the creation of protected areas were marked by the complexity of the tactics of the parties involved, while MSF continued to call for intervention:

  • The AFDL adopted conciliatory tactics to undermine support for an international intervention: on 3 November, just after the BBC’s broadcast of the interview with J. de Milliano, a cease-fire was announced, and Rwanda sent welcoming messages to the refugees; on 11 November, Kabila agreed to the creation of corridors and allowed NGOs access to the city of Goma (but not to the camps).
  • MSF remained sceptical of these ‘positive’ gestures on the part of the AFDL in early November, considering that it did not have enough information indicating an imminent return to normal conditions. MSF therefore maintained its call for intervention. The internal debate continued as to what specifically MSF was asking the international community to do. Following an inter-sections meeting held on 6 November, a position paper was issued that addressed the situation in extremely broad terms: the call for intervention was linked to repatriation, and hence to the conditions required for voluntary return to Rwanda, including “transit centres, reinforcement of human rights monitoring measures … safe environment for returnees, operational judicial system, [decent living conditions in] prisons, effective health care, solutions for the land issue”. The mandate of the intervention force was also discussed: “at this stage, MSF should consider formulating its position on this force, since it may be the case that the mandate of this force is considerably different from what MSF asked”; “a peace- enforcement mandate is a prerequisite for the chances of the force to succeed in their tasks”.
  • During this entire period, MSF’s attempts to reach the refugees continued day after day, in vain (to the point where its teams were blocked between the two borders in trying to reach Bukavu via Rwanda). The situation was completely obscure and concern was growing daily, along with the certainty that the refugees, if left without assistance, would necessarily be in a appalling state. Frustration over being powerless to help the refugees was running high. An apparent window of opportunity opened on 11 November, when the NGOs were allowed access to Goma; MSF immediately sent several tonnes of relief supplies. In reality, however, the NGOs remained confined to the town, with no control whatsoever over the supplies assembled there. The only action MSF was able to take was to establish “way stations” in mid-November along the route of refugees returning to Rwanda.
  • As a result, MSF spoke out repeatedly during this period of denied access and extremeconcern. Some of its public statements repeated the call for intervention to protect the refugees and restore access (letter from MSF-F to the President of France, participation in forums calling for intervention by the United States and by the United Kingdom, a declaration from MSF-US on 14 November on the mandate of the future military force). In addition, further warnings were issued concerning the precise (or presumed) situation of the refugees (PR MSF- F on 8 November stating that there had probably been 13,600 deaths since the beginning of the crisis, a figure obtained by extrapolation from mortality rates in similar situations; PR denouncing forced repatriation to Burundi). The latter public statements, when directed at a specific target, were in all cases aimed at the UN or Western countries.
  • Under continuing pressure from the NGOs (and the ICRC, which was also in favour of intervention), the ‘international community’ vacillated: although the UN Secretary-General declared on 8 November that he was in favour of an emergency action, and although the  Security Council agreed to “the principle” on 9 November, implementation of any such action was postponed. A resolution calling for intervention by a force with a humanitarian mandate (precisely what MSF had warned against) was adopted on 15 November, and this ‘achievement’ eased the pressure in a climate of general satisfaction.
  • The return of hundreds of thousands of refugees between 15 and 18 November brought a definitive drop in the pressure: on 15 November, the AFDL attacked the Mugunga camp, where fleeing refugees had gathered. On 18 November, it announced the establishment of a corridor for the return of refugees via Bukavu. In fact, between these two dates, a lightning repatriation took place: a tide of some 200,000 refugees crossed the border under brutal conditions, while others fled deeper into the forest. Rwanda thus succeeded in defusing the already wavering motivation of the Western countries and promoting the idea – which the developed countries were all too eager to endorse – that the problem was settled.
  • Disagreement concerning the facts was the underlying reason for the divergence of positions between MSF and the other international players: following this flood of returning refugees, controversy arose concerning the number of returnees and the number who were still scattered in Zaire. The United States, which had taken satellite photographs, took some time to reveal the findings, and the latter were problematic in several respects: the figures eventually provided to the NGOs (on 23 November) were different from those released to the press, and the satellites had not covered the entire area. Thus began a veritable ‘war of information’ in which the United States and Kigali asserted that “all the refugees had returned” (MSF-H communication officer), whereas MSF continued to ask what had happened to the hundreds of thousands of missing refugees. In turn, MSF began to be accused, by the United States and in the press, of having overstated the number of refugees and exaggerated the seriousness of their situation on the basis of extrapolated figures. This marked the beginning of the ‘figures crisis’: “after all this agitation, the story was over for the journalists. They were convinced that Africa had organised things perfectly well on its own (…) We were like little children being scolded for making arithmetic errors” (interview, communication officer for all MSF sections).




This episode brought into play a number of elements that reveal both implicitly and explicitly the various conceptions and questions concerning responsibility in the face of violence:

  • First, it should be recalled, at the risk of tautology, that the first actions taken by MSF are always concerned with providing assistance (and access is a necessary condition for such assistance, which explains why loss of access triggers such a strong response). Providing care is indeed perceived as the primary responsibility, the one that makes it legitimate for MSF to be present on the ground and to speak out.
  • The link between realism and demands addressed to the international community is unclear. Although the negative experiences of the recent past – the inaction of the Western states during the genocide of 1994 – were strongly present in everyone’s thoughts, this shadow cast on the present does not seem to have sewn serious doubt on the usefulness of calling for an intervention. Thus, this appeal, problematic though it may have been (and it was debated as such), was ultimately seen to be the only recourse available. The low probability that the international community would actually take action, and the even lower probability that such an action would be in accordance with the specific expectations of MSF, strengthened rather than weakened the demand for “protection”: the reference in the press release to the genocide, and more specifically to Operation Turquoise as a “delayed humanitarian action replac[ing] effective protection”, was intended to revive the spectre of culpable inaction (with humanitarian relief as the ‘alibi’) and to demand action corresponding to the imperatives of the situation (protecting people).
  • This public statement invoked a two-fold responsibility of the international community: “to protect populations in danger in Zaire and to ensure that aid can be provided”. It is not simply a matter of allowing access by NGOs (which bear the responsibility for providing assistance), but of ensuring “effective protection of the population”. In this respect, this appeal differed from the first (PR of 31 October), which called for the creation of protected areas only for the purpose of deploying relief.
  • In this case, since it was impossible to ‘do’ anything (if access had been possible, MSF would probably have leapt into the breach), it was by ‘saying’, i.e. by issuing a public statement, that MSF avoided the unthinkable course of “standing by”.
  • Loss of access (as a means of knowing what was happening and as a last defence against the outbreak of violence) combined with what was already known of the risk of massacres led to   the conviction that massacres were imminent – as if MSF were a witness ‘in advance’ to the coming violence. This sense of looming catastrophe gave rise, as noted above, to an imperative desire to ‘do’ or ‘say’ something to prevent it from happening. Thus, there is indeed the idea  that it is ‘MSF’s responsibility’ to try to help prevent or forestall foreseeable violence.
  • In making a connection between the departure of the “eyes and ears” and the prediction of a “bloodbath”, those on the ground – and those who quote them – were expressing a certain idea of the role played by MSF’s physical presence (as an international organisation): namely, that its presence attenuates the violence. Expressed in a more conventional (and more ambiguous) way, this is the idea that ‘presence protects’ and that a loss of presence leaves the door open to the worst abuses. By implication, there is obviously a perception that MSF should remain close to the ‘most vulnerable’ or ‘most threatened’ groups in order to lessen their vulnerability and the violence committed against them.
  • In addition to calling for immediate action, MSF took a position in favour of comprehensive, lasting solutions. The press release issued by MSF-H warned of the need to restart the repatriation process with guarantees of safety, as well as the problem that the Interahamwe were mingled with ‘real’ refugees. The question of “a return to normal via an organised return to Rwanda” was discussed from a number of aspects: legal, security and so forth. Thus consideration of and positioning on the political aspects of reaching a settlement seemed to be legitimately within the scope of MSF’s responsibility; they were discussed extensively within each MSF section.
  • The connection between one’s reading of the situation and the concern for protection is of central importance. In both our obsession with access and in the “war of information” over the question of missing refugees, the preponderant issue was one of seeing and knowing: access to information and description of situations are at the very root of responsibility to ‘populations in danger’, in contrast to the eagerness of other actors to declare the subject closed once and for all.
  • It may also be noted that the MSF-H document calling for armed intervention employed rhetorical arguments that are commonly used within MSF debates in general: refutation of objections based on the uncertainty of the outcomes and on perverse effects (the ‘futility’ and ‘perversity’ of reactionary discourse, according to A.O. Hirschman), and posing the problem in radical terms that force one to choose between looking like an upright person and looking like a scoundrel.

These various levels of discourse (internal and public) already display most of the horns of the successive dilemmas faced subsequently by MSF.



The period that begins after the rise and subsequent decline of pressure for intervention was characterised by constant tension between, on the one hand, delivery of relief supplies to fleeing refugees, whom MSF was desperately trying to locate and save, and on the other, the ambiguity surrounding these supplies, as the use to which they were put by the AFDL rebels could be glimpsed intermittently – all of this in the context of a serious health emergency, with access being partial, reversible and temporary, enormous logistical difficulties, extremely confused information and the glaring absence of the international community.




Authorisation to explore the area surrounding Bukavu (within a radius of 30 km) was given on 23 November. MSF set off at once and tried at first to go as far as possible; it exceeded the 30 km limit, and as a result, the NGOs were assigned AFDL “facilitators”. On 26 November, the team was informed that the AFDL had massacred some 500 refugees/displaced persons ten days earlier. After alerting the HCR and the ICRC, with no response, the MSF coordinator travelled to the site. He spoke with the survivors and wounded: they were refugees who had not fled the area, hoping to turn themselves in and return to Rwanda. They told him that the AFDL soldiers had pretended to register them and then “launched grenades at them and machine-gunned them”. The coordinator wrote a report and sent it to Amsterdam, where it was decided to forward it to Amnesty International so that the name of MSF would not be mentioned, “for reasons of (still) having difficulty in getting access”. Moreover, the team was increasingly worried about the role of the “facilitators”: “From the beginning, we suspected that these informers were going to report everything (…) We were going to places where we had heard there were refugees. The first day, we would see refugees and on the second, there would be security problems and we couldn’t get there anymore. That happened to us three or four times” (interview, MSF-H coordinator in Bukavu). The coordination team in Kigali was informed.

By 1 December, the picture had become clearer as some information was confirmed. On the ground, MSF was on the point of deciding to issue a denunciation: “along with UNHCR and ICRC, we reached the conclusion that we’d obviously been used as bait (…) The ICRC representative said, ‘it’s not ICRC’s tradition, but (…) we can’t treat refugees and later learn from villagers that they were shot after we’d left. (…) we have to do something’. (…) Just as we were making that decision [to leave the area and tell the world what was happening], the Alliance liaison officers opened the meeting room door and said: ‘You want refugees? There are 5,000 on the road, get to work!’ (…) They knew we were preparing to react and take a position against them. So they opened the floodgates and shoved 5,000 refugees in our direction as a bone to gnaw on”. A bone that shifted the focus of concern towards providing emergency care for this crowd of refugees: “that changed our strategy completely”. We asked ourselves a few months later “why did we stop paying attention?” (interview, Bukavu coordinator).

Thus, in the Bukavu area, the concern for denouncing the use of humanitarian organisations as bait gave way to ‘activism’ due to the necessity of caring for the refugees that come within sight of the teams. The latter tried to set up relief operations for these moving groups of refugees.

The decline in the attention paid to this issue occurred in conjunction with several developments on the Western scene: increased media coverage of past massacres (dating from mid-November), yet at the same time a drop in international mobilisation. As a result of Kabila’s conciliatory gestures, the planned international force now appeared to be “pointless” to the countries that were supposed to promote it, and it was dissolved on 16 December. The consciences of all parties seemed to ease, and the above-mentioned quarrel over the figures continued and grew; several articles appeared hinting that the humanitarian organisations had lied to increase their “business”: “why did the United Nations, Médecins Sans Frontières and Oxfam get it so badly wrong? (…) They grow only of they are good at raising funds (…) these messages [high-profile appeals] raise money” (BBC, 27 November).

The field staff had not completely stopped paying attention to the violence, however. The coordinator remembers that he alerted Kigali and Amsterdam again in late December: “I suggested three or four times that we leave and close the mission”. When he described the use of MSF as bait, he was subjected to questioning: “‘Did you hear the killers? Are you sure this was a massacre? You had stories from villagers saying they [the AFDL soldiers] arrived two days later and killed people but you don’t have any eyewitness statements’. After the problem with the figures, they were very, very cautious. Things were completely paralysed!” (interview, Bukavu coordinator).

What caused this misalignment between the field team and headquarters? The field team, which witnessed directly the harmful effects of its presence and its attempts to approach the refugees, saw the entire purpose of its activity disappear; the context in which an action can be useful seemed to have disappeared. The fact that the team was an ‘interested party’ in the process of hunting down and killing the refugees seems to have exceeded the limit of what it could tolerate. As a result, it wanted a strong response (denunciation, departure).

If we are to believe the version of the Bukavu coordinator, at the decision-making level in the capital – in this case, Amsterdam – the ‘figures crisis’ had resulted in a high degree of caution regarding supposedly obvious ‘facts’ and the veracity of information (whether quantitative or not, collected or generated): the fact that it was more difficult to speak out in this context strewn with pitfalls necessitated more rigorous procedures for collecting information. It is our hypothesis here that these two factors – the greater difficulty to speak out and the greater rigour required – combined to make MSF less responsive to incoming information, i.e. that its capacity for indignation was inhibited or weakenedIndignation is here taken to mean not the automatic and necessary reaction to an unacceptable situation, but rather the outcome of an intellectual stance that makes indignation possible (the ‘capacity’ or ‘predisposition’ to indignation).. In short, we find once again the link, mentioned above, between clarity of information and the perception that it is urgent to take action. Those whom the episode of the ‘figures crisis’ led to place increased reliance on ‘methodological doubt’ also had a less radical perception of the threat to the refugees and of the fact that MSF had served as bait: thus, in addition to caution about speaking out in public, these differences of perception emerged with regard to the question of whether MSF should withdraw, as proposed by the Bukavu coordinator. Whereas the latter thought that the situation had become unacceptable, with MSF being an accessory to the massacres, the head office in Amsterdam was questioning the accuracy of the facts reported.

Ultimately, the concern over the violence inflicted on the local peoples – and over MSF’s role in this process – was to be expressed at the operational level. The Bukavu team’s analysis of the instrumentalisation of the relief effort was confirmed by the Kigali coordination team on visiting the area: “When returning there after permission of the authorities, no refugees were found any more”, giving rise to the “fear that our presence is not improving the chances of the refugees to survive. Are we becoming a risk for the refugees? It is decided to change our intervention approach towards these groups of refugees. We will encourage people to come to the main roads (…) This way we are sure that our assistance is helping the refugees” (sitrep 11 January, MSF-H Kigali). Thus the quintessential humanitarian dilemma, in which aid is transformed into support for a criminal purpose, led not to a general withdrawal but to a series of small-scale withdrawals in the field. The practical modes of aid delivery were modified: we stopped trying to move closer to the refugees and to locate them in the forest, we rethought the question of where to deliver aid and with what messages, etc.

This adjustment by abstention obviously reflects MSF’s responsibility to ensure that its relief efforts do not place people in greater danger than they already are and do not contribute to the conduct of criminal policies.




Meanwhile, it gradually emerged as from early December that there were many more refugees further north, at various places along the Walikale-Kisangani road. For  example, 70,000 refugees wound up at Tingi-Tingi in mid-December (by the end of January the number had risen to 160,000). Here, in contrast to the Bukavu area, MSF rapidly undertook large-    scale operations in various locations: in Kisangani, establishment of an MSF  rear  base,  support to the hospital, a transit centre for displaced Zairians; in Tingi-Tingi,  which  had  become a refugee camp, water purification facilities, establishment of six dispensaries, a nutrition centre. The urgency of the work was clear and the logistical obstacles enormous. The moment for calling for intervention had passed: according to the President of the MSF International Council, given the change in the situation – the return of “600,000 refugees” to Rwanda, access in Zaire to new groups of refugees in a very worrisome state – “we must be capable of changing our public position”. We  had access to the refugees again, and they were    in dire need of assistance: the priority, for the moment, was clearly the operational side.

The teams working in these camps witnessed an “extremely rapid decline in the state of health” of the exhausted refugees: “the refugees were in an absolutely intolerable situation – a refugee camp, escapees in the forest, deep forest (…) it was monstrous, especially when you know what these people have lived through. You always think you’ve seen the worst, but that’s not true” (interview, MSF-F coordinator in Tingi-Tingi, on the French television channel TF1 on 24 January).

While the field teams were extremely busy, overwhelmed with work, the various MSF sections issued a series of public statements from 14 to 31 January (PR MSF-F/UK; press conference, MSF-F Operational Director; interview on TF1; PR MSF-F/US; PR MSF-F/B/US). All of these statements concerned the disastrous situation (as regards nutrition in particular) in the camps and emphasised the responsibility of international community, which was dithering over the deployment of relief for fear of strengthening the militias’ hold over the refugees. In the words of the press release, “the leaders and militiamen are not the ones penalised by delays in distributions and the lack of commitment on the part of the international community but civilian populations”. A solution was offered: “New procedures must be in place quickly so that aid passes directly to heads of households and the militias are prevented to control the camp. UNHCR protection is also needed urgently to guarantee the safety of the refugees who wish to return to Rwanda” (PR of 14 January).



The month of January thus saw a shift towards a logic of providing assistance while at the same time speaking out on aid issues.

  • For the people in the field, who were witnesses to the desperate situation of the refugees and the inadequacy of the relief effort, the need for speaking out was obvious: “I had no worries about speaking out. I was convinced that that’s what had to be done. Otherwise, all you’re doing is providing palliative care and in the end, everyone dies”. The aid workers found themselves in a position of helplessness, watching people die, and sounding the alarm seemed to them the only possible response. We see here an irreducible element of the need to speak out, a need which does not only originate from an expected impact or moral / ethical grounds.
  • At the level of the section headquarters, although the sections seemed to be in agreement on the need to alert the world to the catastrophic situation in the area, an exchange between the DirComs of the Swiss and French sections concerning these public statements reveals differing conceptions of MSF’s position with regard to other actors (a debate that runs throughout MSF’s history in connection with its ambiguous relations with the HCR). The DirCom of MSF-Switzerland questioned MSF’s practice of accusing the international community and “teaching UNHCR a lesson”: “in this situation, it is unwise to make accusations against our partners … Are we sure that those people [the HCR] aren’t doing anything?”. He took the position that we should be careful of what we said about ourselves (again, the use of quantitative data) and about others.

In reply, the DirCom of MSF-F reiterated a number of facts (“the WFP distributed 300 grams [of food] per person over three weeks … the repatriation of the refugees was accompanied by particularly high mortality”) and, in his way of designating the UNHCR, emphasized its responsibility to provide protection, which it had failed to do: “in the face of this pretty serious assessment, the reaction of the leadership of the UN agency responsible for protecting refugees is, at best, out of step given …In that context, it is only logical that there would be friction between us and UNHCR”. Here we see the idea that it is normal (i.e. legitimate and consistent with MSF’s role) for MSF, owing to its work of providing assistance and to what it witnesses, to have strained relations with other actors, and particularly UNHCR: reminding them of their responsibilities, as well as appealing to the more general responsibility of the international community, is viewed as being a part of MSF’s responsibility for protection.

In addition, the DirCom of MSF-F seems to regret retrospectively that the massacres were not addressed during the press conference: “Her [the Operational Director’s] comments were those of a doctor in the field. She said things like: ‘it’s going very poorly there’. At that time, we already had a fair amount of information on the abuses” (interview). The health emergency thus took precedence: “we had not really spoken out on the massacres”, in the words of the MSF-F communication officer, who arrived at Tingi-Tingi on 20 December with the initial objective of producing a report on the massacres: “We’d been thinking about that since mid-December. It was still a bit vague. It was a priority for some people at MSF, but not for others or for the field team”. In the end, she briefed reporters informally on the massacres but primarily emphasised the “food and logistical problems. That was the priority we focused on”. To cope with the emergency, it was necessary to set priorities for both communication and operations: the priority was given to dealing with the existing situation and helping those who had survived thus far to remain alive. In this specific context and at this time, documenting the violence inflicted on these refugees in the past seemed to be a second-order priority.

Thus, as from December, at both Bukavu and Tingi-Tingi, MSF did not give priority to denouncing the violence. It clearly emerges from these two episodes, however, that the issue of assistance remained constantly intertwined with those of preventing or avoiding violence, of protection in the legal sense of the term.

In Bukavu, as we have seen, the operational choices were dictated directly by the realisation   that attempts to reach the refugees and help them were in fact endangering them, which implicitly raised the spectre of Ethiopia. At Tingi-Tingi, MSF’s insistence on the fact that the victims of the lack of aid were “civilians” served as a reminder that the major problem in providing aid to the refugees since 1994 had been that members of criminal militias were mingling with and dominating the ‘real refugees’. This had the effect of depriving the category of ‘refugee’ of some of its validity (since some were ‘false’) – yet this category was of the utmost importance as a legal status authorising the application of the protection ‘mechanism’. In this case, the legal mechanism included assistance, and this is why the UNHCR’s “shilly-shallying” on delivering aid was rightly considered a failure to fulfil its responsibility  to  provide  protection for refugees.

The practical details of aid delivery were thus based on a constant concern that the relief effort should “improve the chance of the refugees to survive” (whether the threat is violence or malnutrition), in the words of the MSF Kigali coordination team quoted above.




The relative geographical ‘stabilisation’ (at least for MSF) of the situation in January broke down once again in February-March, in a scenario virtually identical to that of November- December 1996:

  • Flight of refugees before the advance of the AFDL troops and loss of access for MSF – when the fighting approached the sites where the refugees had gathered, thousands of them took to the roads in early February. To the west, 40,000 refugees left Shabunda; further north, 40,000 more fled from Amisi towards Tingi-Tingi. MSF activities in the latter camp were suspended: “… we knew and they knew, implicitly, that it couldn’t stop there … We tried to get them back on their feet so they could leave” (interview, MSF-F coordinator in Tingi-Tingi). The aim was not to protect the refugees physically, but to enable them to survive in their flight (both when they halted, as at Tingi-Tingi, and when they set out again; see below).
  • The resumption of pressure on the ‘international community’ concerning access to and protection of refugees – a briefing by MSF International before the Security Council recommended that the United Nations move the refugees to a safe area, force their way to civilian populations in South Kivu and assist the displaced in Rwanda, and requested the Security Council to work on a sustainable political solution. MSF-F issued a number of press releases on the same subject: 2 press releases on 13 February,  one concerning the briefing of    the Security Council, the other, issued jointly with other NGOs, on respect for international humanitarian law; one on 28 February, requesting protection for a specific group (specifically, requesting the preventive evacuation of vulnerable individuals from Tingi-Tingi, i.e. those who would be unable to flee if the camp were attacked); on 7 March, jointly with other NGOs,  calling for guarantees of protection for refugees wishing to return to Rwanda and those fleeing the fighting. Lastly, MSF made several statements in the French media, denouncing the “international indifference”. On 11 March, the UN Secretary-General’s proposal to reactivate the multinational force for eastern Zaire, supported by France, was rejected by the United States and the United Kingdom.
  • Virtually no access to the majority of the fleeing refugees – tens of thousands of refugees from Tingi-Tingi reached Ubundu on 10 March; the press went to the scene, but apparently the NGOs did not. From 12 March, MSF was allowed, bit by bit, to enter Amisi and Tingi-Tingi; in the latter camp, there remained only 2,000 of the 160,000 refugees who had occupied it. Obstructed, inadequate assistance was provided there until the end of March. “All they [the volunteers] could do was watch the refugees die”. After the fall of Kisangani on 15 March, the NGOs were permitted by the new AFDL authorities to explore in a radius of 20 km around the city (i.e. an area that the refugees had not yet reached). On 20 March, the refugees gathered in Ubundu were driven away by the resident population and took to the road again. MSF was forced to recognise its helplessness: it had still not gained access to the bulk of the fleeing refugees, and in those places where it was allowed to work, it was hampered and powerless, and its actions appeared futile.
  • Restricted access to individuals scattered far from the ‘mass’ of refugees, used by the AFDL as a tool to locate refugees – from 20 March, on the outskirts of the area where the majority of the refugees were probably located, a few slender opportunities appeared, and MSF seized them: permission from the AFDL to explore the area around Masisi (MSF-H, 20 and 23 March) and Shabunda, to the east of Bukavu (MSF-H again, joining an HCR mission, from 26 March to 3 April). Each of these exploratory missions helped to confirm that there was a policy of eliminating the refugees: villages were destroyed, the people terrorised and residents told the team, “they’re killing everybody”; the HCR representative possessed information on mass graves, and fresh traces were indeed found. Moreover, the idea of relief as bait became clearer day by day. The army announced that the humanitarian workers were coming and followed them after the refugees emerged from the forest. The MSF team learned that several of the people they had spoken to had subsequently been killed. Elsewhere, the AFDL soldiers learned that the MSF team had advised villagers willing to help the refugees to send the latter to the MSF health station by a “loop” “through the forest”: “Two days later there were AFDL soldiers that went there and there were no more refugees coming from Nzovu [the place on the loop indicated by MSF]” (interview, MSF-H exploratory team). Lastly, some AFDL members confirmed explicitly to MSF that they were using the latter to make the refugees come out of the forest.


Thus the overall picture gradually became clearer through the interactions with the villagers and decisions at the micro level based on the available information, all this happening in a context of great pressure and close surveillance by the AFDL informers, which the team tried to outwit using a variety of tactics. This picture made sense when the team observed that each of its initiatives to help the refugees met with failure or was turned against the refugees: collection of information (endangering of witnesses from local villages who communicated with MSF), relief operations and their geographical adjustments (the trap of the roundabout path recommended by MSF).

As early as mid-February, the ICRC had informed MSF and the HCR that it was calling a halt to its relief operations in favour of refugees in the forest, because these operations were instrumentalised to harm those they were supposed to help. Moreover, as mentioned above, as early as December it had become clearly apparent to the team working in the Bukavu area that the relief efforts were being used as bait. This information had not been passed on to the other sections. Were they transmitted to the exploratory mission team? If yes, how can we explain the fact that its members say they realised this only gradually, and on their own? Should this be attributed to the normal ‘latency’ period before a given piece of information takes on its full significance? Or was it rather that people had “stopped paying attention”, in the words of the coordinator who had argued in December for withdrawal?



The two related issues described above (serving as bait, and emergency relief for exhausted refugees) took a more radical turn for MSF in late March.

As we have just seen, it was by this time clear to the teams that conducted the exploratory missions that massacres were occurring and that the relief operations were being used by the AFDL troops to locate and kill the refugees, which raised the question ‘what to do’ (see below).

On 27 March, the mass of fleeing refugees reappeared; a train, on which the presence of the WFP, the HCR, MSF and journalists was allowed, went as far as km 82 of the railway from Kisangani. They found approximately 100,000 refugees halted at various points along the railway: “a skeleton march … they looked horrible”, many with gunshot wounds, and who “wanted to go back to Rwanda, because they were completely finished” (MSF-H coordinator for Eastern Zaire). “In the beginning, it was first things first there. It was a catastrophe. We set up feeding centres and tried to save as many people as possible” (MSF-B Executive Director).



APRIL 1997

As at Tingi-Tingi, large-scale relief operations were deployed very quickly to where the refugees were located – in this case, four sites along the railway where groups had halted at various stages of their flight and various levels of exhaustion.

In accordance with the wishes of the refugees, the HCR was determined to repatriate them as quickly as possible. To this point, MSF had not taken a public position on what solution would be best for the refugees, apart from general statements in favour of their return to Rwanda; now, however, the question would have to be faced squarely and practical decisions made. MSF formulated a critical position based on its certainty that a great many refugees would be unable to survive repatriation: once again, MSF’s contesting of the repatriation procedures envisaged by the HCR was based on a legitimacy derived from its relief work and its medical evaluation of the condition of the people.

On 12 April, the MSF spokesperson in Kisangani wrote: “the whole repatriation operation is likely to be a pretty distasteful affair, a showcase ‘solution’ to the refugee problem (the last thing on UNHCR’s mind is to protect refugees, and the relief aspect is likely to suffer while repatriation goes on)”. Other documents, both internal and for lobbying purposes, spelled out these concerns and requested that repatriation be voluntary and occur only after a medical assessment and stabilisation of the person’s health status. When security problems between the refugees and the local population began to increase, MSF recommended “to speed up repatriation for the valid families and to ensure protection to allow proper assistance to the camps” for the refugees needing care prior to repatriation.

The AFDL, for its part, after granting access to the refugees in late March, continued its on- again, off-again tactics concerning the relief operations (which were occasionally blocked) and repatriation (repeatedly postponed). On 21 April, repatriation was still under negotiation, and the NGOs lost access to the camps after pillaging and clashes occurred, probably provoked by the AFDL soldiers.


At the same time, the question of what strategy to adopt following the two exploratory missions in late March, during which the policy of elimination became obvious, was being discussed intensely within MSF.

On the ground at Bukavu, MSF personnel took different positions according to the activity in which they were engaged: the exploratory mission team wanted to make the information public; the team working in dispensaries in contact with the refugees thought that we should not jeopardise the aid being provided to 10,000 people by taking the risk of being expelled; another team engaged in development activities refused to leave.

The decision taken by headquarters (MSF-H) was to use the information reported by the exploratory mission team, but not publicly: “Apart from the weakness in the organisation, there was also an idea that we could not use the Shabunda report publicly because we prefer to continue our operations even if we have very limited access” (interview, MSF-H DirOp). “We thought that in such a case with such sensitive information, our operation running, and our people on the ground, we had to handle this in a very confidential way … the decision to keep it confidential – the field asked the desk or the OD at the time … the field was very concerned” (interview, member of the HAD). A few days latter, however, a member of the emergency desk who arrived on site on 14 April recalls that “the team as well said ‘we have to release it [the report]’ … I was happy and I was proud. The whole team thought we had to do this” (interview, MSF-H emergency desk).

Using other arguments, the DirCom of MSF-US urged that information relating to the massacres should be managed cautiously: MSF “is not Amnesty – we are not there to pass on what refugees say”; “MSF has no visual first-hand evidence of people being killed … Seeing ‘mass graves’ is proof of nothing … The only proof MSF can give with any credibility is MEDICAL especially in view of our number crisis and credibility deficit post-Kivu crisis, Sept-Dec 96” (memo, 9 April). She therefore recommended that MSF contact human rights NGOs that could handle the forensics, brief journalists on the mass graves so that they could investigate by themselves, and seek advice from the field teams before making any public statements whatsoever.

Around 10 April, MSF-Holland drew up a “protection plan” based on “confidential advocacy”. The overall objective was to “ensure Protection of people at risk and to improve operational access so as to reduce their mortality due to both disease and killings in Kivu”. This was to be accomplished through a number of more specific objectives: to “influence” the AFDL, Rwanda and Uganda in order to “control human rights abuses”; to contact Western governments concerning the problem of access and the massacres, so that they too would bring their influence to bear on the AFDL, Rwanda and Uganda to “ensure adequate protection” of both refugees and residents; to re-kindle the protective activities of the HCR; to ensure that these issues remained on the agenda (in the press), etc.

Implementation began with meetings at the highest level: the Security Council on 16 April, the Special Representative for the Great Lakes, briefings for Western governments and the European Union by the exploratory team, etc. As a preface to these meetings, MSF-H presented five areas of concern: lack of access, lack of protection, the use of aid as bait, the need to apply pressure to Kabila, and the need for a commission to investigate human rights abuses. Our watchwords continued to be insistence on sticking to the facts, the idea of ‘bearing witness’ based on presence on the ground, and caution concerning information that had not been confirmed objectively (as indicated by the restricted briefing provided by headquarters to the coordinator of the exploratory mission who was tasked with these high-level meetings).

At the same time, the ‘Shabunda report’ was finalised and sent to the other sections. MSF- Belgium, which had teams in the field, contested the form of the report, describing it as “weak” and “poorly written” (“just because the story has proven to be true, that doesn’t mean we can speak out every time we have this type of information”; interview, MSF-B Operational Director). It tried to prevent MSF-H from going to the meetings provided for in its advocacy plan. MSF-H suspended all its lobbying activities among journalists “due to objections of MSF- B (security for the team in Kigali might be at stake) … NO JOURNALISTS WILL BE BRIEFED, NOR GIVEN THE REPORT” (review of communication, MSF-H, 18 April).

MSF France, in contrast, was extremely shocked that the possession of such vital information did not lead MSF-H to alert the other sections and speak out: “this meant that more than a month after the events, no concrete, operational, or public decision had been made on the basis of that information ! I was appalled” (interview, MSF-F Legal Advisor). “What elevates this kind of issue above the level of rumour is when an organisation commits to saying that it believes it to be true and provides evidence backing up its convictions. None of the organisations involved – not UNHCR … MSF-Holland or ICRC – took a public position on such an important subject” (J.-H. Bradol, MSF-F DirCom).

From this point on, the various MSF sections quarrelled daily over strategy: MSF-H ‘vetoed’ a press release on the current situation (in order to leave the teams enough time to get organised); MSF-B vetoed the Shabunda report (the veto was rescinded once the report had been revised by MSF-F, pursuant to a joint decision, but MSF-H, the original author of the report, was left disgruntled); restrictions imposed by MSF-H (on the distribution list for journalists). Each of these ‘vetos’ was transgressed by MSF-France, deliberately for the most part. As noted above, MSF-F considered it urgent to take action: “to me, the report should have been made public as quickly as possible … information like that can circulate through e-mails for two months without prompting anyone to stop what they’re doing – that just hit me in the stomach” (MSF-F Legal Advisor). At MSF-B, however, the retrospective analysis was quite different: “The release of this report could easily have been postponed a few extra days. And in fact, what use was it? None at all! … we said what everybody already knew.… it was a fact, there was a war on, we were also being used in Rwanda”, and as for the ‘confidential advocacy’ strategy, MSF-B “trusted MSF-H, which “was the back-up section” (interview, MSF-B desk).

In the end, owing to a leak from MSF-Spain, the report was distributed sooner and much more widely than planned. The field teams were reduced and made ready for withdrawal.




These differences of positioning between MSF sections, between the field and headquarters and between people in the field, reflect different readings of the situation, different perceptions of our responsibility in the face of the massacres and of the fact that our presence was endangering the refugees:

  • Discussion of the consequences of possible actions: public declarations were likened to a ‘death sentence’ for current operations by all those opposed to speaking out concerning the violence and the instrumentalisation of the relief effort (during the discussions between field staff and headquarters at MSF-H, and subsequently at MSF-B). Rhetorical argument, gut feeling or carefully considered view of the reality? At any rate, this link was taken to be obvious, automatic and not dependent on the context, and as a result the terms of the debate shifted: the question was no longer whether it was necessary or useful to speak out in view of the seriousness of the information collected, but well and truly whether MSF was willing to sacrifice operations and endanger the field teams; apparently the pros and cons were not weighed up. All those who participated in the discussions in early April, except for the team that conducted the exploratory mission, preferred the option of passing information on confidentially. MSF-B went further, opposing even confidential distribution of the Shabunda report.
  • References to the identity or ‘mandate’ of MSF, when used, meant that content be classified according to whether it came within ‘our field’ for speaking out. The objection that “we are not a human rights organisation” (MSF-US DirCom) implicitly pointed to MSF’s medical identity, meaning that the legitimacy of its public declarations must be based on duly verified medical data – which, it should be noted, has become the predominant position today. The frame of reference of MSF-F, in contrast, was a ‘humanitarian mandate’ that in this case was placed under intolerable stress justifying an immediate reaction: “not only is humanitarian aid no longer effective, but it’s contributing to getting people killed” (MSF-F Legal Advisor).
  • The status of veracity versus that of conviction: how was the choice made between insisting that very serious information be validated and coming down in favour of radical action? On the one hand, the emphasis on what constitutes evidence (MSF-US DirCom) and criticism of the Shabunda report as “very poorly written” (MSF-B Operational Director) refer to the two-fold requirement of legitimacy and credibility, which were understood as depending on MSF’s ability to produce figures and objective, confirmed information. The paradoxical co- existence of the conviction that the facts described were true and reservations about communicating these facts may be explained in one of two ways: either the fear of releasing information that might be contested (owing once again to the disagreeable experience of the ‘quarrel over figures’, mentioned by the DirCom of MSF-US); or the judgement that the situation itself did not merit radical action, or in any event did not merit action at any price (MSF-B did not want to endanger its operations in Kigali). On the other hand, at MSF-France,the events described in the report were regarded as extremely serious (see below), and it was considered evident that MSF should take radical action based on the report (alert the other sections, halt all operations with the AFDL, etc.), as indicated by the indignation felt when this was not done. As a result, the fact that this information was not quantifiable was not regarded as a reason for proceeding with caution; rather, it demanded that MSF put its responsibility (credibility) on the line while making the information public.
  • Qualification of the seriousness of the events described: the central importance of the issue of complicity in the disappearance of refugees for MSF-F. Taking a position completely opposed to those of MSF-H and MSF-B, MSF-France focused on the content of the report, taken ‘as is’, as the most important factor: “this information was at the heart of the fight that all sections had been leading for two months to make people understand that not all the refugees had returned to Rwanda and that they were in danger in the forest” (MSF-F Legal Advisor). Here, protection of the refugees is described as the attempt to make known the persistence of the danger to a population group that had hidden from view, ‘evaporated’ into the forest, and whose disappearance the international community had connived at by evading the question of the ‘missing refugees’. In addition to direct complicity in a policy of massacres by serving as bait, the concern was over connivance (by failing to react) in rendering the refugees invisible, as a pre-requisite for eliminating them. The events described were not weighed against other factors, whence the implicit predisposition to accept the risks associated with the ensuing decisions on operational matters and public positioning.
  • A possible link with the hierarchical structure of the different sections. In this episode, decisions at MSF-H “were really taken by the people in the field” (coordinator, Shabunda exploration team), i.e. those in the field, the majority of whom recommended discreet communication, whereas at MSF-France decisions taken at headquarters took precedence (possibly because MSF-F has few operational activities compared to the other sections).
  • Relations with other actors and the abdication of responsibility by the international community: The necessity perceived by MSF-F of ‘not being an accomplice’ seems also to have been related to what, ever since Bosnia and Rwanda, has been the ‘intolerable’ topic of the abdication of responsibility by the Western countries in the face of criminal policies.It was precisely in response to such abdication of responsibility that the notion of a “responsibility to protect” took form in the late 1990s. On the notion that such abdication is intolerable, our reference is an analysis put forward by Michel Feher on changes in the discursive polarities that frame humanitarian action. On the notion of the intolerable, as mentioned in the main document (footnote no. 12), we rely on D. Fassin and P. Bourdelais (eds.), Les constructions de l’intolérable, Paris, La Découverte, 2005, 228 pages.  In this sense, the above-mentioned “fight” consisting of contesting the international community’s self-satisfied version of history and denouncing its “indifference” (see above) is an attempt to prevent such abandonment from recurring. This explains the great importance attributed to the information collected and to making it public in order to shake these countries out of their indifference. MSF-H, in contrast, viewed the position of the international community less radically in terms of abdication of responsibility, probably because of its concerns relating to security and operationality. Thus, it can be seen that the confidential advocacy plan reflects a concern with awareness-raising and influence on actors taken one at a time, instead of criticising them collectively for their inaction.



In late April, the polemic over whether to speak out against the massacres had begun. Access to the Biaro and Kasese camps had been cut off since 21 April, and international pressure for access was growing (United States, UN). At this time, the two processes – that of massacres and the use of the humanitarian effort as bait, and that of access and assistance to exhausted refugees – came together for MSF: rumours of massacres in the camps where MSF had been working were gaining substance; the massacres were now happening in present time and were perceptible even though they were taking place at a distance from any possible witnesses.

On 24 April, the news came that these camps were now empty: “where are these people, most of whom were unable to walk?” There was indeed a “strategy of eliminating refugees”, and refusal of access was one of its facets (minutes, Zaire task force meeting, 24-25 April).

On the ground, on the fifth day without access, the tone was one of failure: “we were reminded today, once again, that humanitarian agencies, in general, and UNHCR, in particular, are impotent … neither the number or tenor of the condemnations can make up for this third failure: each time, their number is reduced and so is the feasibility of repatriation”. This led to the “conclusion [that] operations are impossible” and to uncertainty as to what action to take: leaving exposed MSF to “the unbearable prospect of not being here just when we would have access to the refugees again”; staying exposed it to meaningless activity, “without even a guarantee of being able to save individuals” (sitrep, MSF-B coordinator in Kisangani, 25 April).

Marlène, an MSF-B volunteer just back from Biaro, testified the same day before the Board of Directors of MSF-France: “the eight tents overflowed with people who were very close to death

… For about a week, we thought we could do something… Then the rebels blocked access to the camps … On Saturday/Sunday, when we came back, we counted the bodies. Medical activity had become impossible. We had to do something else; stop our work and condemn what was going on”.

This was not the majority view at MSF-B, neither in the field nor at the higher levels: “There were 32 expatriates in all and not everyone shared this view. Most stayed on site with a ‘passive advocacy’ mindset”; the coordination team in Kisangani (which as we have seen was thinking about a withdrawal) eventually decided to stay so as not to abandon the refugees. The volunteer protested: “the MSF officials decided not to speak out but to stay on site … For me, such advocacy rings completely hollow … Watching people being killed and counting the bodies: that does not constitute genuine refugee ‘protection’ … there is in any case a deliberate strategy of bringing about the death of as many as possible”.

At MSF-F, this testimony, expressing a sense of utter powerlessness owing to the inability to give any meaning even to medical care, reinforced a position that was already in divergence with those of the other sections. It definitively confirmed the conviction that a systematic policy of elimination was at work, against which it was imperative to speak out (i.e. speaking out should take precedence over operations if a choice must be made between the two): “if you are convinced that the refugees are being exterminated by every means possible, how can you propose a ‘silent advocacy’ strategy in response? This is part of MSF collective responsibility and is beyond the pale of what we are willing to understand, even while trying to be constructive. This position led to the killing of several thousand people” (MSF-F DirCom). Another member of the Board declared that “silence has killed tens of thousands of people”. The tension between maintaining a presence and speaking out was called into question: “We cannot give in to the terrorism that says speaking out is no longer possible because there are teams in the field” (MSF-F Legal Advisor); “We have to stop pitting advocacy against our presence in the field” (MSF-F Executive Director).

The Board declared that “the primary purpose of MSF action, which is focused on defending populations in danger … must guide MSF’s action” and take precedence over institutional arguments. It voted to eliminate the veto principle, and declared that it was against repatriation to Rwanda.




The details of this debate at the MSF-F Board of Directors meeting give us information that extends the above analysis on different readings of the situation and different levels of responsibility. The French section’s harsh criticism of the policy of MSF-H and MSF-B serves to indicate how MSF-France conceives of its own role.


Without repeating what has already been said, we will merely highlight the following:

  • The conceptual link between an elimination policy and the necessity of speaking out in public appears explicitly in MSF-F’s lack of understanding of the silent advocacy position taken by the other sections. The opposition between bearing ‘passive witness’ (being present, witnessing events, being ‘with’ the people in danger) and bearing public witness (with a more explicit aim of providing protection and sounding the alarm to change the course of events) was drawn.
  • The argument concerning the consequences of our action was used, in a negative, accusatory manner, by several people in describing the disastrous consequences of silence. Although this interpretation had no objective value, it highlighted the idea of incurring direct responsibility for people’s lives through inappropriate action; the other MSF sections were explicitly accused of causing, by their silence, the deaths of thousands. We come back to the idea (once an MSF slogan) that speaking may not always save people, but it is certain that “silence kills”.
  • The automatic link between speaking out and calling MSF’s presence into question was definitively criticised and contested by key members of MSF-F at this Board meetingIn his annual report delivered at the May 1997 General Assembly, the President, P. Biberson clarified this position (see the excerpt of the annual report in the main document, Part II, section “confronted with the ‘logic of extermination’”) (in the preceding documents this question was more or less ignored, the focus of the discussion being on the question of whether to speak out). The MSF-B volunteer who testified, however, repeated the idea that speaking out and presence are mutually exclusive, though she reached the opposite conclusion to that of her superiors: that MSF should “stop working” and “condemn”.
  • In short, a core element of the debate was a concern for protection: the rejection of passive testimony, the accusation that MSF was responsible for the deaths of thousands, and the emphasis on “defending populations in danger” as the ultimate purpose of MSF’s activities all point directly to this, although the word “protection” was not much used.
  • It can nevertheless be seen that there were different contents underlying this common concern. For Marlène, the volunteer, it was the impossibility of “doing anything” (conducting a meaningful medical action) about the massacres that made her switch towards “doing something else”, “stop our work” (i.e. speak out). The two positions were not concomitant; rather, there is a shift from one field (providing care) to another (protection, to which she refers obliquely when she says that silence does not constitute “genuine ‘protection’”). For the Board members who took the floor – and the president in particular – however, reaffirming that the principle of defending populations in danger should “guide our actions” meant that this principle is at the very heart of any action: any activities undertaken, whether they come under the heading of health care or public statements, are meaningful only insofar as they contribute to this “primary purpose”. Thus this view contests the dichotomy between ‘care’ and ‘protection’, and the opposition between presence and speaking out. In so doing, it provides grounds for MSF-F’s willingness to put operationality at risk, or even to sacrifice it altogether, when the situation warrants (as it did in the eyes of the Board members on 25 April): it is no longer a question of switching to a fundamentally different position, but of a different balance between various facets of a single position, namely defending populations in danger. It might be objected that in the end, the conclusions reached are the same as those in Marlène’s reasoning and that all these distinctions are mere quibbling; in our view, however, this conceptual divergence underlies many of the debates concerning protection.


After the turning point represented by the Board of Directors meeting of 25 April, where the members of MSF-F decided that henceforth nothing could justify hindering them from speaking out in public, intense disagreements continued to arise concerning the conditions in which they should speak out: strong positions were taken concerning adherence to procedure, owing to the presumed impact on the safety of field teams (crisis between the Belgian and Dutch sections and MSF-F over the latter’s press release of 26 April, which spoke of a “policy of extermination of Rwandan refugees” and was released without warning the other sections). In this respect, it is worth noting that the facts (as observed by MSF-H itself) did not bear out these fears about public declarations: “there were no threats against MSF teams in the region following this very strong advocacy message [the Shabunda report, following which the teams were on the point of being withdrawn]” (news flash MSF- H, 28 April). However, MSF-F continued to be accused of endangering the teams and suspected of being irresponsible because there were no MSF-F staff in the field. Surprisingly, however, the MSF International Council, at a closed meeting on 1 May, sided with the position defended by Paris: “in the current situation, speaking out should take priority over direct assistance”. This stance adopted by the IC did not, however, put an end to the conflicts between the sections.




Subsequent developments involved the same difficulties, the same decisions and the same debates as those explored above; the gulf between sections over the issue of speaking out and recommended solutions grew rapidly wider and deeper.

We will not examine further the dissent between sections over reports concerning the massacres, apart from noting that, after the Shabunda report, it was the dissemination of the report Forced flights, finalised on 15 May, that gave rise to further clashes. After the report was leaked to the French press (on 20 May, the front page of the French daily Libération bore the headline “MSF accuses”), media attention to the elimination of the Rwandan refugees reached its highest point in the entire refugee crisis. This also marked the culminating point of the discord between MSF sections over presence versus speaking out.

We will focus rather on the repatriation issue, which became particularly acute in May when practical implementation of the process began, in disastrous conditions. Every day, the situation of the surviving refugees worsened:

  • Still playing on ambiguity, the AFDL stepped up its pressure on the refugees: after the obstacles to access and the massacres of April in the death camps south of Kisangani, the AFDL seems to have decided to put an end to the refugee problem once and for all, along with the criticisms of the humanitarian organisations. It authorised repatriation of the refugees to  Rwanda, immediately gave the HCR an ultimatum to carry it out within 60 days (27 April) and then undertook to implement the repatriation itself (30 April), once again catching the humanitarian agencies unawares, although they had been calling for repatriation for weeks. At the same time, the AFDL intensified its efforts to intimidate those capable of causing it problems, namely journalists and over-talkative NGOs (i.e. MSF): AFDL  representatives  visited the team in Kisangani, demanding that it withdraw its public declarations on the massacres. This led to MSF-H’s decision to keep a low profile where communication was concerned (massacres, the responsibility of Rwanda, impunity) and increased the above- mentioned disagreements with MSF-F concerning safety. In the end, on 11 May, the AFDL declared that it was giving the NGOs 10 days to evacuate Biaro. Further along the path of the fleeing refugees, not far from Congo-Brazzaville, the hunt was still on: on 13 May, when Mbandaka was taken, hundreds of the 40,000 refugees who had arrived there were massacred.

  • The HCR took the course of repatriation at any price: on 28 April, the High Commissioner, Mrs. Sadako Ogata, told the Security Council that when the protection of the country of asylum (in this case, Zaire) was not assured owing to conflict or lack of security, it could become necessary to repatriate refugees without strict guarantees as to the voluntary nature of repatriation nor the safety of the refugees.In the summer of 1994, Mrs. Ogata had declared that she was in favour of immediate and rapid repatriation; subsequently, under pressure from the HCR’s “protection” department, she moderated her approach, stating that the wishes of the refugees were the key to repatriation. For a detailed description of the HCR’s positions and the issues surrounding refugees in the Great Lakes region, see Arnaud Royer, De l’exil au pouvoir, PhD thesis in sociology, University of Paris-I, March 2006, esp. pp. 20-54 and 324-370, where the author describes the “primacy of security and financial interests over protection of populations”. The author shows how, in 1993-94, HCR came to see repatriation as “the” solution, as a result of the financial difficulties encountered (it became imperative to cut short the time spent by refugees in host countries), in a context where the refugees had gradually become a security and billeting problem. Within this broader trend, he describes concretely how the HCR incorporated forced return in its policy regarding action in Zaire in 1996-97. She denounced the AFDL’s efforts to hamper repatriation as well as the killing of refugees. On 6 May, a trainload of refugees organised by the AFDL left for Kisangani in very poor conditions: 91 of the refugees were found to be dead when the train arrived. The HCR denounced the lack of access to the Biaro camp and its lack of control over the evacuation. Nevertheless, repatriation was under way, with 2,000 to 2,500 refugees evacuated daily to Rwanda. Between May and June 1997, 50,000 refugees were repatriated by air.




The MSF sections, each in turn, held internal meetings and reached conflicting positions on the repatriation issue, while in the background, the disagreements over communication concerning the massacres were sharper than ever.

On the ground, operations were undertaken in the various places where repatriation was taking place. There were no grounds for disagreement in this case. It was obviously necessary to assist the ill, exhausted refugees wherever they were: in the area south of Kisangani (the Lula transit camp, near Biaro, which had a mortality rate of 70/10,000/day, the generally accepted level for defining an emergency situation being 1/10,000/day); in the sites around Mbandaka, some of which were very difficult to reach; and in Rwanda, where the activities of MSF-B were stepped up.


The differences deepened, however, regarding the positioning of the various sections:

  • MSF-France considered that repatriation was not a solution, as Rwanda was directly involved in the policy to eliminate the refugees (cf. the stance against repatriation taken at the above- mentioned Board meeting of 25 April). Its position was that MSF should say “loud and clear that the refugees must not return to Rwanda” (MSF-F Legal Advisor) and should consider the fundamentals of the situation in order to find a solution. In the view of MSF-F’s Legal Advisor, Mrs Ogata’s statement of 28 April indicated that the HCR had abandoned its founding principles: in this “exceptionally serious” shift in position, the HCR had institutionalised, through “a new doctrine”, the fact that it could sacrifice the pre-requisites of voluntary return and security to the necessity of repatriation (message from MSF-F Legal Advisor, 6 May 1997). MSF-F made its position public: on 3 May, it denounced the “media operation” of the Zairian rebels, who were engaged in “forced repatriation” of the refugees in Biaro to “convince the international community that the problem is being solved” (quoted by AFP); it then called for “an immediate halt to the repatriations” (PR MSF-F, 7 May).
  • MSF-Belgium (which meanwhile had become the back-up section, replacing MSF-H) requested from the outset that the other sections not follow the course of MSF-France. Its position was that “MSF is not against repatriation since it is the only feasible way left to try and get the people out of the shitholes of Kisangani and Biaro and save them”; as a basis for this position, it cited the fact that the refugees themselves had expressed a desired to die in Rwanda, since they were going to die anyway. Several days earlier (on 5 May, following the death of the 91 refugees), MSF-B had issued a press release describing the desperate situation in Biaro and requesting the AFDL to bring about acceptable conditions for repatriation. In Rwanda, MSF-B was acting as the “official partner of the authorities for medical screening at the camps through which all refugees transit”. The idea was to try to “find the right balance  between critical observations, communication of humanitarian info and advocacy to try and improve the conditions in which repatriation is taking place” (MSF-B update, 6 May).

On 11 May, however, the positions of MSF-F and MSF-B came together in their common refusal of the AFDL’s ultimatum to the NGOs to evacuate Biaro and in their request for “unrestricted access” (PR MSF-F, MSF-B, MSF-USA). On 26 May, in a joint press release, MSF- F and MSF-B requested funding for transferring refugees located in swampy areas, where they were very difficult to assist, to more easily accessible camps.

Differences of opinion resurfaced on 28 May, however: the ex-president of MSF-F denounced the HCR which “instead of proclaiming the right to asylum and guarantees of protection for the refugees in Rwanda itself, is bowing to international pressure and rushing this repatriation”, sending the refugees back to their “oppressors” who “are destroying the refugees by the thousands” (Le Monde, 28 May). The MSF-B team in Kinshasa was outraged by this statement and predicted an immediate decline in its safety.

It seems, however,  that the teams on the ground (Kisangani and Mbandaka, near the border   with Congo-Brazzaville, covering respectively 5,000 and 2,000 refugees) recognised the  harmful effects and limitations of the repatriation process as conducted by the HCR and the AFDL: they resisted the zeal of the HCR officials who appeared in MSF facilities looking for patients to evacuate: “that busybody from the HCR quite annoyed the MSF medical staff and  had to be resisted daily”. Some patients did indeed disappear, probably evacuated by the HCR. More generally, “the basis of this transportation (a more appropriate term  [than “repatriation”]) is sinister”: direct and indirect threats to the refugees, no alternative offered by the HCR (the refugees were thus agreeing out of resignation, weary of being hunted and repeatedly attacked), movement of the refugees towards another source of danger (they were worriedly asking MSF staff whether the humanitarian workers would be present in the towns    of Rwanda – where, according to the ICRC, there would be “300-400 arrests  every  day”)(report, information officer, 31 May 1997).

The MSF-B desk continued to think that instead of criticising the HCR, MSF should rather consider both institutions as “partners in the refugee crisis”, since all were in the same predicament with respect to this situation (document, MSF-B desk, 5 June 1997).



Many aspects of this disagreement over repatriation – from procedures on making public statements to the criteria for deciding between medical action and lobbying – reflected the rationales described above. In all cases, they revealed that the same concern – saving as many people as possible – had a different content from one section to the other. How can we explain such sharp disagreement between the sections, which faced the same situation and whose concerns were identical?

We will extend here the remarks made above concerning the variations on the theme of a desire for actions that actually protected people, trying to uncover the actors’ reasons and ways of acting. As we have noted repeatedly in recounting this story, the divergences between MSF- F and MSF-B reflect the opposition between, on one hand, a radical, denunciatory stance and, on the other, pragmatism and caution (a tension that likely runs throughout the history of MSF, dividing but also running transversally through the individuals and institutions concerned). This tension arises from both different readings of the situation and different ways of doing things; these two differences are probably linked in circular causality, with practice shaping points of view and vice versa. Here, after the sections’ positions had had several months to crystallise, they are particularly noticeable:

  • At MSF-France, the situation was described in general terms as an elimination policy whose perpetrators had been identified. The result was a perception that it was absolutely necessary to ‘do something about it’: to denounce those responsible “to try to put an end to the slaughter” (President’s annual report, May 1997), to dissociate MSF from the policies pursued by the perpetrators or policies that served their purposes (at the time, repatriation). It was thus hoped that speaking out in denunciation would be an effective form of action to “defend populations in danger”, to the point where at times it was the only way of contributing to this objective: “on several occasions, denunciation or speaking out has been the last rampart against the loss of all meaning, and perhaps the last way of doing anything useful” (ibid.). It was thus urgent to speak out, for both external (trying to influence the situation) and internal (trying to ensure that MSF’s actions were meaningful) reasons. It should be noted, however, that this did not lead to a radical decision to leave the country; the position described seems not to preclude remaining in the field and making every effort to save people (providing care on site), even though MSF personnel had few illusions about their ability to do so, which surely had an impact on the degree of radicalness adopted.
  • MSF-Belgium, though probably in agreement with MSF-France’s overall reading of the situation, saw concrete situations first and was confronted with specific problems to resolve (probably as a result of its presence in the field). Its aim was pragmatically to ‘cope with’ the situation as it was, to exert influence on the perpetrators to use less violence, to allow access to the most vulnerable people, to set less stringent conditions for repatriation. It also sought to cooperate with the authorities so that the repatriation would proceed under the least bad conditions possible, at both the point of departure and the destination. MSF-Belgium wished to establish a “dialogue” with the AFDL and considered the HCR and the Rwandan authorities as “partners”. It can be thus seen that it had a different assessment of the roles and capabilities of the different players: from this standpoint, its perception was that the HCR could hardly do any more, that the AFDL could be influenced in the right direction, that Rwanda was rational and powerful, and hence to be accommodated. Accordingly, the thinking at MSF-B was first and foremost operational; the objective was not to save the refugees but to save individual refugees, as patients, and MSF-B’s public declarations were concerned not with denouncing the violence but rather with sounding the alarm about specific situations, on the basis of descriptions backed by medical evidence. In keeping with the idea of circular causality between analysis and practice, we may suppose again that it was partly the wish to remain operational and on site that fostered the perception that the actors could be influenced and would be willing to cooperate – an assumption that was surely questionable where the AFDL and Rwanda were concerned.

A few words on the opposition between denunciation and caution are in order. As shown, in our view, by the description of the differing positions of MSF-F and MSF-B, it seems to us that speaking out is generally viewed as more ‘noble’ than caution, which is often associated with concessions, compromises and timidity. Many have argued against this view, viewing prudence as a philosophy of action based on acceptance of what is real, on an ethics of the possible, as opposed to the black-or-white moral stance, with little concern for what happens in the real world, that underpins indignation and denunciation.Cf. the philosophical concept of phronesis (i.e. prudence); see Gilles Achache (ed.), La Prudence, Editions Autrement, Paris, 1996, 186 pages. And in our view, it is precisely towards such practices characterised by caution – action based on what is actually possible, public statements based on medical facts, neutral description of situations and less use of political epithets – that MSF-France seems to be leaning today, ten years after the hunt for Rwandan refugees in Zaire and the quarrels it provoked in terms of policy decisions related to protection.

Providing medical care before repatriation, checking the conditions of reception on arrival, trying to reach people wherever they may be, preventing a return towards a dangerous area, refusing to transport people who would not survive the trip, trying to put an end to the whole process, etc. – we can see clearly that the question of repatriation, like the dilemmas discussed previously, gave rise, based on differing views of ‘how to protect people’, to positions and practices that were partly in harmony and partly irreconcilable, and that cannot be reduced to a simple list of what would be ‘good’ and ‘bad’ solutions.





In June-July, information concerning the massacres was consolidated through collection of eyewitness accounts, with a view to tracing the experience of the refugees throughout their flight across Zaire. This collection gave rise to a report entitled Atrocities Against and Massacres of Rwandan Refugees in the Territory of Democratic Republic of Congo: Refugee Testimonies (18 June 1997). Disagreements continued between those who questioned the added value of such a document and those who argued for a “responsibility to inform the world of what these people lived through”. In the end, the report was not made public, but released confidentially to human rights organisations.

In late June, MSF-B reviewed the results of the process of repatriation to Rwanda: it was a sombre assessment that clearly recognised the lack of protection for those repatriated, knowingly caused by the Rwandan authorities – lack of security, lack of food and health assistance in Rwandan towns for the repatriated, despite their deplorable condition, attacks on the towns where most ended up, massive arrests, the impossibility for the NGOs of monitoring the situation, lack of access to the entire eastern part of the country. There was no longer any question of the Rwandan authorities as “partners”, and the situation was deemed to be “humanly unacceptable”. MSF-B called for the protection and registration of refugees, for voluntary return only, for separating refugees from militia members, for an investigative commission, for an end to impunity, etc. (report, MSF-B desk, 28 June; PR and Board minutes,  MSF-B, 11 July 1997). In early July, the HCR also noted how little information was available on the fate of the refugees repatriated to Rwanda; nevertheless, it declared itself in favour of continued repatriation, admitting that it was unable to protect the refugees in Congo-Kinshasa itself. In conjunction with this, both the HCR and MSF decided that they would “no longer go looking for [the refugees still hiding in the forest]… because [the forest] is their last protection”. Thus, once again the humanitarian organisations were forced to recognise the impossibility of protecting the refugees, whether by presence or by action, and whether in the country of asylum or the country of origin.

In Congo-Brazzaville, lastly, MSF-F noted that there was no way out of the impasse until the ‘real’ refugees were separated from former FAR soldiers; as lives were not in danger, it planned to withdraw in February 1998.



In July 1997, Epicentre initiated an epidemiological survey at the request of MSF-F. The aim was to document a posteriori the violence committed through a retrospective survey on  mortality conducted among refugees in Congo-Brazzaville. The survey – one of the first of its kind – established that “the breakdown of the events that occurred during the flight of the refugees was as follows: missing: 59.5%; killed: 19.7”. “Only 17.5% of the members of families initially present in Kivu had reached the Congo”. This ‘epidemiology of violence’ (as it has  been called since then) made it possible to identify peaks of violence or mortality, which corresponded precisely to those moments when MSF was denied access to the refugees. Although in this survey there was no longer any question of protecting people, there is the idea of a ‘responsibility to know and to inform’ that extends beyond the present. This  notion probably arises from a deep-seated reflex (bearing witness as a ‘duty’, as a ‘moral imperative’,  to ‘prevent it from happening again’), but also from the fact that, throughout the crisis, establishing the true facts and figures had been a central issue – the ‘classification battle’ between the MSF version, in which there were fleeing refugees in danger for months, and the official version, in which there were no missing refugees but only killers who had been routed. The report was finalised in early September, but not made public until November, as some sections had deemed it inaccurate.

During the same period, the forced repatriation of several hundred refugees still in Congo- Kinshasa came to an end, with no indication that the facts would be established, that responsibilities would be pointed (through the end of impunity, as requested by MSF), or that there were any positive prospects for the survivors.

The fall of 1997 saw the publication of several reports on the violence that tried to draw attention back to this crisis (HRW, Amnesty); some of them designated certain actors, particularly Rwanda and the United States, as bearing responsibility for the events. MSF- Epicentre published its report in The Lancet and spoke at a hearing in the US House of Representatives. Attention to the crisis gradually eased.

Lastly, the UN investigative commission which had arrived on site in June 1997 (following repeated requests by several agencies, including MSF) produced a report that was damning for Kabila in July. It thereafter saw its work constantly and deliberately obstructed. The commission eventually, in April 1998, declared that it was abandoning its task.

The story of the hunt for refugees, which for MSF was a long ordeal of impotence and attempts to do something against the violence, was about to be declared over, even as its ramifications were already provoking further crises, giving rise to other emergency actions by MSF in each of the countries of the sub-region.