What Outbreak Metrics Can and Cannot Tell Us About the Ebola DRC Response

The 17th Ebola outbreak in the Democratic Republic of the Congo (DRC) has surpassed 6271 confirmed cases and 3041 deaths. The case fatality ratio is nearly 1 in every 2 confirmed cases. A central question looms: Is the outbreak finally turning a corner, or is it getting worse?¹

Recent indicators reveal an uneven response to the outbreak: despite genuine progress over the last 4 months, persistent operational hurdles continue to stall containment.

Where the Numbers Align: Documenting Ground-Level Wins

Four measures of transmission have moved in the right direction, and they deserve to be stated plainly before anything qualifies them (Figure 1).

Cases are doubling far more slowly than in May; daily incidence remains well below its late July peak; weekly totals have dropped for 2 consecutive weeks; and the time-varying reproduction number (R_t) has followed suit.² Among these metrics, doubling time is the most reliable. It is straightforward math on what was directly observed, requiring no modeling or assumptions about the time between successive infections.

The response has moved too. The DRC rolled out 70,000 doses of Ervebo, licensed for Zaire ebolavirus, to frontline health workers, setting aside 20,000 of those doses for a phase 3 trial testing protection against Bundibugyo virus.² Concurrently, early-stage trials began for 2 Bundibugyo-specific candidates in the UK from July 24 and in Canada from August 3.^3,4

The operational challenges show why this is a more complex picture. On August 31, 43 of the 57 confirmed deaths recorded in a 24-hour window occurred in the community rather than in a treatment center.⁶ These community deaths represent 75.4% of the daily mortality.⁶ Over the last 14 days, 63% of the reported 719 deaths.⁶ A person who passes at home without safe burial practices, during peak viral shedding, that is the work still ahead.

So, what do the metrics really tell us?

What R₀, and R_t, Can and Cannot Tell Us About Containment

When we talk about viral reproduction, 2 different numbers answer 2 different questions.

The basic reproduction number, R₀, describes how a pathogen spreads before anyone intervenes. In June, a CDC branching process model estimated R₀ for this outbreak at approximately 2.51, inside the expected range for an Ebola virus.⁷ The time-varying reproduction number, R_t, describes what transmission is doing now, under whatever response exists around it.

The threshold is 1: Each infected person infects 1 other person. Above it, each infection produces more than 1 further infection, and the outbreak grows. Below it, transmission weakens, and the outbreak eventually burns out. The daily count has its own shape (Figure 2). Following late May's reporting backlog, daily cases stabilized near 35 through June, then surged to a peak of 127 in late July, easing back to roughly 76 in August.⁶ Mapping these case numbers against the estimated trajectory of R_t reveals just how sensitive transmission dynamics remained to shifts in intervention.

I estimated R_t using the method of Cori et al, applied to the daily increase in the DRC national count of laboratory-confirmed cases.⁸ That method requires an assumed serial interval, the typical time between 1 person's symptom onset and the next person's. I was unable to identify a published Bundibugyo-specific serial interval, so I used the estimate derived from Ebola virus disease caused by Zaire ebolavirus in West Africa in 2014.⁷

The trajectory follows the case curve (Figure 3). All values here use a 14-day estimation window.⁸ From mid-June through mid-July, R_t held between 1.25 and 1.32. Across the second half of July, it averaged 1.68 and peaked at 1.96. Over the final 2 weeks of August, it averaged 1.00 (ranging from 0.90 to 1.12), and over the final week, it averaged 0.92.

This range tells the real story. While there is room for cautious optimism that the response is working, it remains premature to conclude transmission has definitively turned.

Looking in the Rearview Mirror: Reporting Delays and the Limits of R_t

The WHO has attributed part of the reported increase to improved diagnostic capacity, expanded surveillance, and periodic backlog reconciliations, while cautioning that genuine transmission is still driving the uptick in reported cases.⁹ With both dynamics at play, R_t cannot separate real spread from better detection. That distinction matters; an analyst’s choice of estimation window can change the perceived trajectory. Either fewer people are transmitting the virus, and interventions are working, or more people are becoming infected, and the current interventions are not enough.

The deeper issue at play is temporal. The date a case is a confirmed positive is not the date the symptoms began. Between exposure and reporting lies an obstacle course of delays: an incubation period of 2 to 21 days, the decision to seek care, specimen transport, and laboratory processing.⁹ By the time a case is counted, the transmission event that produced it is days to weeks old. Consequently, an R_t calculated from reporting dates is artificially delayed and smoothed relative to the true transmission chain.^5,6

Despite its inherent flaws, anR_t describes the epidemic that surveillance can see. It is a metric, a data point that, in tandem with others, provides insight into how the virus is spreading. For this outbreak, as with every outbreak, all eyes are on the metrics that could have the greatest impact.

Outbreak Metrics vs. Ground Reality: Evaluating Core Indicators

On paper, the epidemic appears to be settling into a manageable rhythm. Official summaries point toward stabilizing metrics, but beneath those aggregated numbers lies an outbreak that is continually slipping ahead of containment. The fundamental issue is that public health interventions rely on complete visibility, yet ground operations are currently tracking only a fraction of actual transmission.

• Fractured Contact Tracing

According to the Africa CDC, successful contact tracing and follow-up is the metric that would have the most impact in containing this outbreak.⁶ On the surface, frontline efforts are successfully following up with known contacts of patients with confirmed Ebola (Table 1).

However, this one metric tells only a fraction of the story. Containment requires identifying at least 95% of new infections among people already placed under surveillance, yet only 15% to 20% of new confirmed cases were previously identified as known contacts.⁴

Responders are logging an average of just 10.6 contacts per confirmed case, roughly half the minimum of 20 contacts per case recommended by the Africa CDC.⁶ To give context to the impact of this, over a recent 21-day period, 1627 confirmed cases produced an expected pool of 97,600 contacts. But tracing efforts captured only 19.2% of them (18,784 individuals), leaving an estimated 78,816 contacts unaccounted for.⁶

If 4 of 5 new cases never make it onto a contact list, the surveillance system is capturing only a minority of infections, undermining the reliability of every downstream metric.

• Diagnostic Gaps and Untested Alerts

In outbreak surveillance, the goal is to keep laboratory test positivity low; Africa CDC sets the target at 0%.⁶ At first, this may seem counterintuitive, as a high positive rate looks like the lab is finding things, so it reads as success. It is the opposite: it means the net is being cast where cases are already known. With roughly 1 in 4 people testing confirmed positive, testing practices are primarily reaching people already likely to be infected rather than the wider population where transmission is happening. The concern is that testing is reaching individuals who are already critically ill, leaving milder, earlier transmission chains unrecorded.

Laboratory diagnostics reinforce this blind spot. Half of all positive results reported on August 31 came from postmortem samples, and in some areas that number is as high as 66%.^2,6 To address this gap, the DRC adopted a community-centered strategy to boost local reporting, driving a reported 25% increase in suspected case alerts from villages.⁶ However, samples were collected and tested for only 72% of the suspected cases identified, leaving a substantial volume of suspected cases unassessed.²

Field teams are casting a wider diagnostic net, yet they are still detecting most confirmed cases only after death.

• Community Mortality and Funeral Transmission

When health officials cannot track where the virus is spreading, people end up dying at home instead of in a treatment center.⁶ The August 31 pattern described earlier fits a 3-week trend in which nearly 60% of people who died did so at home.² In an outbreak response, the target for community deaths is zero, because an unmonitored death outside a treatment center represents uncontained transmission and entirely unknown contacts.

The human and biological cost of these home deaths is more than you might expect. Viral load peaks at death. Every bodily fluid is heavily contaminated and infectious. When patients die at home, family members care for them without access to personal protective equipment. They wash the patient, clean up fluids, and comfort them in their final hours.

Currently, fewer than half of the affected health zones have an active safe and dignified burial team, falling far short of the required 100% coverage. When a family loses a loved one at home and no safe burial team is available, they conduct traditional burials on their own. These unmonitored funerals create additional opportunities for infections among mourners.

• Treatment Center Bed Capacity and Regional Strain

The national-level treatment center bed capacity is promising, with open beds available. However, when this metric is broken down at the regional or health zone level, bed occupancy tells a more complex story. Treatment centers in North Kivu, an active hot spot, are at roughly 129% capacity, with some facilities at 140%.^2,6 This strain tracks with transmission: North Kivu recently produced more than 2.5 times its usual share of daily infections.² Driven by active urban outbreaks, the province's case fatality ratio sits near 68%, far above the national average of about 49%.²

Local-level metrics show where capacity is short, so support and resources can be deployed where they matter most.²

Ground Won: Measuring Real Victories in Inactive Health Zones

Despite uneven metrics, there is genuine progress. Nationwide, 91 of the DRC's 151 health zones under outbreak surveillance currently have zero confirmed cases.⁶ Another 9 zones that once fought active transmission have now gone at least 21 days without a single new case, completing a full incubation cycle. These are real victories. The immediate task ahead is to concentrate resources on the 51 health zones where the virus is still actively spreading.⁶

Reaching a 21-day benchmark in those 9 zones shows that the response model works when baseline containment holds. The challenge is that containment is inherently fragile. An inactive health zone is only as safe as its borders, and travel along trade corridors connecting North Kivu and Ituri means that silent reintroductions remain an ongoing threat. Maintaining zero cases requires sustaining alert networks and burial teams even after active clinical caseloads disappear.

References

1. Alert and response. World Health Organization. Accessed September 3, 2026. https://www.who.int/emergencies/alert-and-response
2. Institut National de Santé Publique, Ministère de la Santé Publique, Hygiène et Prévoyance Sociale, République Démocratique du Congo. Rapport de situation de la 17ᵉ épidémie de la maladie à virus Ebola/RDC [Situation report on the 17th Ebola virus disease epidemic/DRC]. SitRep No. 109/MVEBDB/31/08/2026. Reporting date August 31, 2026. Accessed September 2, 2026. https://insp.cd/sitrep-n109-mvebdb-31-08-2026/
3. World Health Organization; Africa Centres for Disease Control and Prevention. WHO and Africa CDC welcome the allocation of Ebola vaccines to the Democratic Republic of the Congo. News release. August 20, 2026. Accessed September 2, 2026. https://www.who.int/news/item/20-08-2026-who-and-africa-cdc-welcome-the-allocation-of-ebola-vaccines-to-the-democratic-republic-of-the-congo
4. Kabasele D, Meyer E, Kabore I, et al. Notes from the field: characteristics and monitoring of the 2026 outbreak of Ebola disease caused by Bundibugyo virus: Democratic Republic of the Congo, August 2026. MMWR Morb Mortal Wkly Rep. Published online September 1, 2026. doi:10.15585/mmwr.mm7535e1 
5. Special briefing on Ebola outbreak response. Africa CDC YouTube page. August 27, 2026. Accessed August 27, 2026. https://www.youtube.com/live/Lz2bF2VeiSA
6. Special briefing on Ebola outbreak response. Africa CDC YouTube page. September 3, 2026. Accessed September 3, 2026. https://www.youtube.com/live/u_gvmNapxYk
7. Mooring EQ, Koval WT, Routledge I, et al. Modeled scenario projections for the Ebola disease outbreak caused by Bundibugyo virus, 2026. MMWR Morb Mortal Wkly Rep. 2026;75(22):285-289. doi:10.15585/mmwr.mm7522e1
8. Cori A, Ferguson NM, Fraser C, Cauchemez S. A new framework and software to estimate time-varying reproduction numbers during epidemics. Am J Epidemiol. 2013;178(9):1505-1512. doi:10.1093/aje/kwt133
9. World Health Organization. Ebola disease caused by Bundibugyo virus: Democratic Republic of the Congo. Disease Outbreak News. August 28, 2026. Accessed September 2, 2026. https://www.who.int/emergencies/disease-outbreak-news/item/2026-DON616