Scientific Day 2026: Key takeaways

For nearly forty years, Epicentre has been supporting MSF’s tuberculosis (TB) control programmes by addressing the key practical issues that have shaped the evolution of this disease. TB-HIV co-infection has been one of the main drivers of Epicentre’s work on TB. Early research focused on the prevalence of TB amongst cohorts of patients living with HIV, as well as on the methods of self-administered treatment. At the same time, research has focused on optimising treatments, which were long characterised by their length and complexity. These research activities have helped to strengthen local research capacity, notably through the establishment of a laboratory in Mbarara, Uganda, which has become a key player in several research projects on TB and other diseases.

In collaboration with several partners, Epicentre has, in particular, taken part in studies assessing the benefits of increasing rifampicin doses in order to shorten the duration of treatment whilst maintaining its effectiveness, the STATIS study on the benefits of empirical initiation of anti-tuberculosis treatment in high-risk patients where a rapid diagnosis is difficult to establish, and the endTB study, which evaluated shorter, less toxic and more effective treatments against ‘multidrug-resistant TB’.

The issue of diagnosis has long been central to Epicentre’s work on TB. Numerous methods have been evaluated over the years, notably the TB-LAM, a point-of-care test that is useful for patients living with HIV (several TB-LAM studies), and the GeneXpert MTB/RIF, which has represented a major advance in improving diagnosis.

One of the major challenges remains the diagnosis of paediatric TB, which is still largely inadequate, with nearly 50 per cent of cases remaining undiagnosed. This issue was at the heart of the session. In the absence of suitable tests, the World Health Organisation (WHO) now recommends the use of treatment decision algorithms (TDAs), based on clinical signs and radiological abnormalities (where X-rays are available), which enables a score to be assigned and treatment to be initiated rapidly. Furthermore, the WHO recommends a short course of treatment for children with non-severe forms of TB, but this is very rarely used in practice.

As part of MSF’s TACTiC project, Epicentre is coordinating operational research to assess the diagnostic performance, feasibility and acceptability of these algorithms in five countries (the TB ALGO PED study), as well as the implementation of short-course treatments in two countries (the 4-months regimen study).

Diagnosis of TB in children with severe acute malnutrition (SAM)

Diagnosing TB is particularly challenging in these children. Weight loss is one of the signs indicative of TB, and symptoms associated with other common illnesses in malnourished children can resemble those of TB. The study shows that TB diagnostic tests have good sensitivity and acceptable specificity in these children. Furthermore, the study contributed to capacity building among local teams and to the integration of TDA into national guidelines.

The study also evaluated two TB-LAM urine tests in children with SAM. These tests, which are carried out using a urine sample, are simple to perform. The study showed that the Determine TB-LAM test, recommended by the WHO for children living with HIV, has low specificity in malnourished children without HIV, which limits its use. In contrast, the FujiLAM test, with its high specificity, could be of diagnostic value, but the test’s sensitivity remains low.

Artificial intelligence and chest X-rays in children

X-rays are essential for the diagnosis of TB, but interpreting them is often difficult for non-specialist clinicians. The study showed that computer-aided detection (CAD), approved by the WHO for the interpretation of chest X-rays to detect pulmonary TB in adults, remain underperforming in children due to radiological differences. The study shows that CAD is a promising tool, but its performance remains insufficient for paediatric clinical use. It is therefore necessary to develop specific models tailored to children.

Implementation of a short-course (4-month) TB treatment regimen for children

A mixed-methods study conducted in the Philippines and Uganda evaluated the implementation of a 4-month short-course treatment regimen for children with non-severe forms of TB. It showed that the 4-month treatment is feasible, with a high eligibility rate in primary care. It can be implemented even without chest X-rays. The large-scale adoption of short-course treatment could improve clinical outcomes whilst optimising the use of resources.

This session highlights concrete progress towards improving the diagnosis and management of paediatric TB, whilst emphasising the need for tools better suited to children.

This session focused on two haemorrhagic fevers with distinct patterns of transmission: Lassa fever and Ebola virus disease.

Lassa fever remains one of the most neglected diseases in West Africa, despite causing thousands of deaths each year. In the absence of an approved vaccine, it is listed as a priority pathogen by the WHO. Epicentre is currently collaborating on a multicentre study being conducted in Liberia, Nigeria, Benin and Sierra Leone, aimed at better documenting the epidemiology of this disease: incidence, seroprevalence, identification of at-risk groups and assessment of its severity. These data are essential for guiding the design of future vaccine clinical trials.

Incidence of Lassa fever in at-risk communities in West Africa: some findings from the Enable programme

Among the key findings presented, it is clear that Lassa fever remains a relatively rare disease, but is likely to be under-diagnosed in areas of high prevalence. Children appear to be at a similar risk to adults. Preliminary results (not available in this video of the session) suggest that vaccine trials will require large sample sizes. However, careful selection of study communities and active surveillance could improve their feasibility. It also appears crucial to include children in vaccine development and to strengthen prevention measures at household level.

Heterogeneity and determinants of Ebola virus transmission: lessons learnt from the 10^th outbreak in the Democratic Republic of the Congo.

The second presentation focused on reconstructing transmission chains during the 10^th Ebola outbreak (2018–2020), caused by the Orthoebolavirus Zaire. This analysis provided a better understanding of the mechanisms by which the virus spreads. A key finding emerged: a minority of individuals are responsible for most transmissions, whilst around 80 per cent of those infected do not transmit the disease. However, the lack of a typical profile for the most contagious individuals complicates the implementation of effective targeted strategies.

Analyses show that several factors significantly increase the risk of transmission, including a failure to isolate patients promptly, death outside a specialist facility, failure to be identified as a contact case, and spending a prolonged period in the community without receiving care.

Although these findings are specific to the Orthoebolavirus Zaire species, this study conducted by Epicentre provides valuable insights into better understanding the dynamics of transmission during other epidemics, notably that caused by the Bundibugyo virus, which was declared on 15 May 2026 in the DRC.

In light of emerging health challenges (oncology, sickle cell disease, ABR), this session highlighted the commitment of MSF and Epicentre to addressing complex issues, despite the operational constraints they pose. It also highlighted the central role of operational research in generating high-quality data, providing patient care that closely aligns with their needs, and developing solutions tailored to the realities on the ground. This legacy is rooted in decades of innovation and learning around large-scale HIV care. Today, it continues to drive progress in the fight against HIV, particularly among the most vulnerable populations, such as sex workers in Malawi, while providing actionable approaches and tools to address new public health challenges.

Adoption and persistence of CAB-LA in sex work hotspots in Malawi

A study conducted among sex workers in 21 hotspots in Malawi explored the adoption of various HIV prevention strategies, including long-acting injectable cabotegravir (CAB-LA). Preliminary results show a marked preference for CAB-LA over daily oral PrEP, with participants highlighting its convenience and discretion. The study also observed persistant use among CAB-LA users, while identifying several levers to strengthen its long-term adoption, including awareness-raising, peer support, and the adaptation of care models.

AI and HPV genotyping to detect precancerous cervical lesions

Another presentation focused on the screening and management of precancerous cervical lesions. This cancer, for which nearly 80% of deaths occur in low- and middle-income countries, is caused by persistent infection with certain types of human papillomavirus (HPV), 13 of which are particularly high-risk for causing cancer. The PAVE study is evaluating an innovative approach that combines HPV genotyping using a self-collected sample with automated visual evaluation (AVE) powered by artificial intelligence. The goal is to identify women with the highest-risk precancerous lesions and offer them prompt treatment via thermocoagulation. Conducted among 50,000 women, the study shows that this strategy enables excellent risk stratification and has identified nearly 98% of high-grade precancerous lesions (CIN3+). The results suggest that combining HPV genotyping with AI-based screening could facilitate targeted and more accessible management of cervical cancer, including among women living with HIV.

Feasibility and acceptability of hydroxyurea-based treatment for children with sickle cell disease 

A third presentation focused on sickle cell disease, one of the most common genetic disorders in Africa. In Niger, approximately one in twenty people carries the disease. Although hydroxyurea is known to reduce complications and improve patient prognosis, its use remains limited due to its cost and concerns about side effects. As part of a pilot project conducted among children cared for by MSF in the Diffa and Madarounfa regions, researchers assessed the feasibility, acceptability, and impact of this treatment on quality of life.

Preliminary results show a significant improvement in quality of life after six months of treatment, which was even more pronounced after eighteen months. The most significant benefits relate to pain reduction, the children’s emotional well-being, and a decrease in anxiety related to the disease. Interviews with patients, their families, and caregivers also highlight a decrease in the need for medical care, particularly blood transfusions, as well as improved socialization among the children and reduced anxiety among family members.

These encouraging results confirm the acceptability of hydroxyurea use in rural settings and support its wider implementation.

The goal now is to gradually expand access to hydroxyurea while maintaining good long-term clinical and laboratory monitoring and by strengthening screening, national patient referral pathways, and advocacy for affordable access to treatment. 

New surveillance and alert tool for Healthcare-associated infections

The final presentation focused on HAI-SAT, a tool currently under development that combines the KoboCollect and WHONET platforms to improve the surveillance of healthcare-associated infections (HAIs). HAIs affect approximately 7% of patients in high-income countries and up to 15% in low- and middle-income countries. The goal of HAI-SAT is to centralize and analyze surveillance data to accelerate the detection of nosocomial infections, the identification of clusters, and the implementation of appropriate operational responses. Ultimately, this tool could significantly strengthen the surveillance of healthcare-associated infections and help improve patient safety in MSF intervention settings.

Although cholera is an ancient disease that is easy to prevent and treat, it remains a major public health challenge, with nearly 600,000 cases and 8,000 deaths reported in 2025. Eradicated from Europe for over 150 years thanks to improved living conditions, access to safe drinking water, and sanitation, it continues today to affect the most vulnerable populations living in situations of poverty, exclusion, and conflict, which explains its persistent presence in the contexts where MSF and Epicentre operate.

Three to four years ago, the global resurgence of cholera necessitated a scaled-up response. But beyond the increase in the number of cases, the very nature of the disease has also changed: more countries are facing epidemics exceeding 10,000 cases, cross-border transmission is on the rise, and epidemic waves tend to last longer, even in areas historically less affected. The situation in the Democratic Republic of the Congo (DRC) illustrates this trend. The 2025 outbreak was the most severe observed in 25 years, with rapid spread even into non-endemic areas and case fatality rates exceeding 3%.

Yet the tools for controlling cholera are well known: early case detection, real-time surveillance, rapid adaptation of strategies, decentralization of care, community engagement, and improved access to water and sanitation. The question, therefore, is not so much what to do, but how to do it effectively in increasingly complex humanitarian contexts marked by insecurity, strain on health systems, population displacement, and misinformation.

It is precisely in these situations that operational research comes into its own. 

Cholera dynamics in the DRC and the impact of Oral Cholera Vaccination (OCV) in endemic areas of the Eastern part of the country

While reactive vaccination is used during epidemics alongside WASH (Water, Sanitation and Hygiene) interventions, the World Health Organisation (WHO) recommends two-dose mass preventive campaigns in endemic areas. 

The study examined two different contexts: Goma, an urban area marked by the massive influx of internally displaced persons, and Bukama, a rural area. In Goma, the analysis shows that the increase in cases coincided with the opening of camps for displaced persons and that the epidemic was primarily concentrated in health zones where camps are present. The impact of OCV vaccinations is therefore difficult to determine in a context where the arrival of displaced persons has obscured the signal; however, these results support prioritizing preventive campaigns among displaced populations, in addition to strengthened WASH measures.

In Bukama, the incidence of cholera showed strong seasonal variation; preventive vaccination proved effective in preventing future outbreaks during the following rainy season, but only under optimal conditions (two doses with high coverage). 

Overall, the results confirm the value of preventive cholera vaccination in endemic areas under certain circumstances, while emphasizing the importance of achieving high vaccination coverage and prioritizing the most at-risk populations, particularly displaced persons and communities living in lake regions. 

Randomized immunogenicity trial evaluating the non-inferiority of a delayed second dose of the oral cholera vaccine (OCV), Euvichol-Plus®

Oral cholera vaccination with Euvichol-Plus® follows a two-dose schedule administered two weeks apart. However, this schedule is difficult to implement during large-scale campaigns, particularly in emergency settings, where logistical constraints and a global vaccine shortage led to the temporary adoption of a single-dose strategy in 2022.

While a single dose allows for rapid expansion of vaccination coverage, it provides shorter-lasting protection. The second dose therefore remains essential to ensure optimal and long-lasting protection. Furthermore, several studies suggested that a two-week interval might be too short to fully stimulate immune memory.

To address this question, a randomized trial conducted in Conakry, Guinea, involving participants aged 1 to 39 years compared three vaccination schedules: a second dose administered after 2 weeks (standard schedule), 6 months, or 12 months. The objective was to demonstrate the immunological non-inferiority of extended intervals for the Inaba and Ogawa serotypes of Vibrio cholerae O1.

The results show that regimens with a second dose administered at 6 or 12 months are non-inferior to the standard regimen in terms of the immune response against V. cholerae O1, which is responsible for nearly all current epidemics. These results provide strong evidence in favor of more flexible vaccination strategies, allowing the second dose to be delayed without compromising long-term immunity.

Outbreaks of meningitis and measles remain major threats in humanitarian crisis situations. For nearly 40 years, Epicentre, alongside MSF and its partners, has been conducting research aimed at improving patient care and adapting prevention strategies to the realities on the ground in the most vulnerable settings.

In the case of meningitis, Epicentre’s work has helped to gradually simplify treatments to make them accessible in settings with severe operational constraints. Trials conducted in Niger and Mali first validated the use of oily chloramphenicol, and subsequently that of a single dose of ceftriaxone, thereby considerably facilitating management during outbreaks. The research has also helped to improve the World Health Organisation’s early warning systems and response thresholds.

Assessment of the effectiveness of the Nm5CV vaccine in Niger

The introduction of the MenAfriVac vaccine in 2010 led to the elimination of major epidemics of meningococcal A meningitis in the African belt. More recently, the pentavalent NmCV-5 vaccine, which targets serogroups A, C, W, X and Y, offers a new prevention strategy. The case-control study presented during this session, which was conducted in Niger, demonstrated a vaccine effectiveness in the field of over 90 per cent. However, certain serogroups, notably C and W, continue to circulate, highlighting the need for further research, particularly into the vaccine’s impact on bacterial carriage.

Developing measles vaccination strategies tailored to the local context in the Democratic Republic of the Congo 

The second part of the session focused on measles. Despite the availability of an effective vaccine, nearly 100,000 cases were still recorded worldwide in 2024. A major turning point came in 2005 when Epicentre’s work in Niger demonstrated the benefits of vaccinating even in the midst of an epidemic. Since then, strategies have been refined through the analysis of epidemiological data, enabling better targeting of the groups and areas most at risk.

Today, the aim is to anticipate rather than react. The work presented in the Democratic Republic of the Congo illustrates this shift. Through the UrgEpi project, Epicentre and MSF combine vaccination coverage surveys, mixed-methods studies and qualitative research to identify barriers to vaccination, guide campaigns and tailor messages to local realities. This approach has led to an increase in vaccination coverage of up to 27 percentage points in some areas.

The results of this work confirm the importance of a data-driven approach to better target interventions, strengthen community buy-in and identify pockets of under-vaccination. They illustrate the shift in strategies for combating vaccine-preventable diseases: moving from a response to outbreaks to a proactive approach focused on anticipation and prevention.

This session began by highlighting the importance of mortality surveys in humanitarian action. The experience gained by MSF in refugee camps has gradually led to the identification of 10 essential public health priorities - which have since been expanded - in complex humanitarian emergencies. These include rapid assessments of population health status, such as mortality surveys, nutritional surveys, and evaluations of vaccination coverage, shelter conditions, and access to water and sanitation. These tools have become indispensable for guiding interventions and tailoring responses to the actual needs of affected populations.

Since its founding, Epicentre has conducted nearly 170 mortality surveys, often in contexts of conflict, mass population displacement, or extremely limited access. Several landmark surveys have shaped the organization’s history. One of the first surveys focused on internally displaced persons arriving in El Meiram from Abyei and Northern Bahr el-Ghazal in Sudan in 1988. It revealed an exceptionally high mortality rate -40 per 10,000 people per day - with nearly a quarter of the population having died within four months. 

In 1992 in Merca, Somalia, in the midst of a civil war, a mortality survey based on systematic random sampling revealed extremely high mortality rates, estimated at between 91 and 121 deaths per 1,000 people per year. Beyond its findings, the study sparked significant debate within MSF due to the conditions under which it was conducted: teams had to travel under military protection to access certain areas.

Another landmark survey was conducted in Goma in the Republic of Zaire (now the Democratic Republic of the Congo (DRC)) in 1994, at the time of the massive influx of Rwandan refugees following the genocide. Within a matter of days, nearly 800,000 people crossed the border into the eastern part of the DRC, creating an unprecedented humanitarian crisis. Despite initial reluctance from certain international actors, including the United Nations High Commissioner for Refugee (UNHCR) and the Center for Disease Control and Prevention (CDC), a mortality survey was conducted among approximately 4,000 people to document the scale of the crisis. The results revealed extremely high mortality rates, estimated at between 20 and 25 deaths per 10,000 people per day. The study showed that approximately 10% of the population had died within a month, primarily due to the cholera epidemic that followed the massive influx of refugees, and then to an outbreak of shigellosis that occurred in the weeks that followed.

Over the decades, numerous other investigations have marked the history of MSF and Epicentre. Conducted in contexts of conflict, epidemics, or mass population displacement, they illustrate the teams’ ability to produce reliable data despite extremely difficult access conditions. 

Why measure mortality and violence?

These mortality surveys play a central role in MSF’s work. They help estimate the scale of a crisis, guide resource allocation, and ensure that the right interventions are implemented in the right place at the right time. Beyond their operational value, they also serve as a powerful tool for reporting and advocacy. By independently and rigorously documenting deaths among affected populations, they help alert the international community to crises that are sometimes overlooked and enable decision-making based on reliable data.

MSF’s independence lends strong credibility to the findings produced by Epicentre. Even when not conducted with this specific objective in mind, these surveys can also help establish the facts in contexts marked by controversy or allegations of violations of international law. The data collected has thus been used in international legal proceedings, particularly regarding the Rohingya, where it served as a major source of information to document the impact of violence on civilian populations.

Overview of methods for estimating mortality and violence

The session then presented the various methods available for measuring mortality and violence in conflict zones, each of which addresses specific constraints. In these environments marked by insecurity, population displacement, and difficulties in gaining access, epidemiologists must constantly strike a balance between speed of implementation, operational feasibility, and accuracy of results.

Rapid estimates, based on simplified surveys, are relatively inexpensive and quick to implement. They generally require few resources and can be conducted with the consent of local authorities or communities. Although less accurate than more comprehensive surveys, they provide rapid insights into the severity of a crisis and can have a significant impact on the direction of the humanitarian response or on communication efforts.

Surveys of displaced populations, on the other hand, provide more robust data on mortality and violence. However, they require more preparation, particularly obtaining ethical and regulatory approvals, which can slow down their implementation. In contexts of successive displacements, they also have certain limitations: respondents must sometimes recall events dating back several months, which introduces memory bias, and the results apply only to households that have actually reached host sites.

Two recent surveys of people who fled the Zamzam camp in Sudan illustrate the value of this approach. Conducted in Tawila, Sudan and in refugee camps in Chad, respectively, they aimed to document living conditions and the violence experienced before, during, and after the flight. In Tawila, the crude mortality rate was estimated at 1.24 deaths per 10,000 people per day, with a significantly higher mortality rate among men. More than half of the recorded deaths were attributed to acts of violence, primarily gunfire and explosions, that occurred in El Fasher or during the flight. In Chad, the crude mortality rate reached 1.41 deaths per 10,000 people per day, and the reported deaths were also predominantly linked to violence experienced in Zamzam, El Fasher, or during the initial displacement. These surveys provide important information on the violence experienced by displaced populations, while highlighting their limitations, particularly in documenting sexual violence and the fact that they cover only those who reached the site or where the survey took place. 

When direct access to populations is impossible, other strategies can be employed. Cohort surveys of “proxy” populations, for example, involve interviewing groups whose characteristics closely resemble those of the target population. This approach was notably used in Gaza with the households of MSF staff and their cousins. It allows for the collection of information in contexts where access is extremely limited, but the results cannot be considered representative of the entire population.

Phone or SMS surveys are another alternative when communication networks remain functional. Their main advantage is that they allow for rapid remote data collection. However, they require reliable and accessible phone number databases and often suffer from low response rates. A study conducted in Cameroon showed that mortality estimates obtained by telephone were comparable to those collected through door-to-door surveys, but less reliable for measuring population movements. SMS surveys, on the other hand, encountered particularly low response rates.

More recently, Epicentre has been exploring approaches based on family networks, which involve collecting information on members of the same family scattered across different geographic areas. This method holds great potential for reaching inaccessible populations, and a pilot project is planned in Sudan to evaluate its performance and limitations.

All of these retrospective approaches face several methodological challenges, including memory biases associated with long recall periods and difficulties in reconstructing household structures in contexts marked by family separations and repeated displacement. Other tools can also be used, such as surveys of key informants, satellite image analysis, or combined approaches based on capture-recapture methods. Prospective monitoring conducted by community health workers can also be considered when security conditions permit.

Finally, the panelists emphasized that surveys are only one part of the crisis documentation framework. Data from MSF’s medical programs also provide essential information on mortality and violence, including sexual violence and torture as well as their consequences for nutrition, mental health, and public health. Combining quantitative and qualitative data is now considered indispensable for understanding crises, adapting humanitarian operations, and strengthening advocacy on behalf of affected populations.

Trauma and burns in gaza: documenting the long-term consequences of war

The final presentation focused on a retrospective study conducted at the Nasser Medical Complex in Gaza, one of the few hospitals that was able to maintain trauma and burn care services despite the conflict. The results illustrate the scale of the human toll caused by the war: more than 30,000 patients admitted between May 2024 and June 2025, primarily for trauma related to violence, particularly from explosives and gunfire.

The study shows that trauma primarily affects adult men, while burns disproportionately affect children and often occur in the home. Variations in the number of trauma admissions over time also reflect changes in the security situation and difficulties in accessing care.

Despite extreme constraints, the teams reported low in-hospital mortality, a low rate of recorded complications, and a high proportion of patients discharged home. However, the significant number of amputations, severe burns, and complex fractures highlights massive and long-term needs for reconstructive surgery, rehabilitation, and disability care.

Beyond mortality directly linked to conflict, the data presented underscore the importance of documenting morbidity, disabilities, and long-term care needs. Protecting health facilities and maintaining surgical and rehabilitation capacities are essential to limiting the long-term consequences of humanitarian crises on populations.