In an Ebola outbreak, the burial itself can become a transmission event. This study asks whether burials can be made both safer and more dignified—and which kinds of teams are most likely to succeed. Between July 2018 and April 2020, the Democratic Republic of the Congo experienced its tenth recognised Ebola Virus Disease epidemic, affecting North Kivu, Ituri and South Kivu and causing 3470 confirmed and probable recorded cases.

Of the 3470 recorded cases, 2287, or 66 percent, were fatal, showing both the outbreak’s scale and its high fatality burden. Despite improved knowledge, vaccines and multi-drug treatments compared with the start of the 2013–2015 West Africa epidemic, the epidemic proved difficult to control.

Response efforts were hampered by insecurity, political tensions, misinformation and public distrust, while the response was criticised as not inclusive and insufficiently responsive to community feedback and lived experience. Transmission of Ebola virus occurs via direct physical contact with an infected person, infected body fluids or contaminated fomites.

Corpses of Ebola cases are particularly infectious because of high viral loads and virus remaining viable for days after death. That makes post-mortem care and practices involving contact with the deceased high-risk, including washing, cleansing, dressing, touching during funerary rites, and placing the body in a coffin or grave.

In the Democratic Republic of the Congo, these practices are customarily performed by family members and the local community, so safe and dignified burials became a recognised pillar of Ebola epidemic responses. Evidence on the effect of safe and dignified burial on Ebola propagation was very limited, although it suggested that the intervention could reduce transmission considerably.

Earlier implementation in West Africa was marked by poor adherence to protocols and a biosafety-driven approach that excluded families, used burials or cremations without consent or knowledge, and sometimes used unmarked or mass graves. This approach contributed to low trust and uptake of the service.

The alternative proposed here was strict infection prevention and control while preserving dignity and enabling culturally acceptable family and community participation. During the eastern Democratic Republic of the Congo epidemic, the DRC Red Cross Society, supported by the International Federation of the Red Cross and Red Crescent Societies, implemented and coordinated the SDB service.

The package included community engagement, a burial site agreed with the family, modifications to traditional funerary customs, psychosocial support, Ebola testing, and decontamination of dwellings. Initially, mobile teams were staffed only by Red Cross volunteers. Civil Protection teams were then established on 23 September 2018, and they performed the largest share of SDBs.

From 12 December 2018, community-led emergency harm reduction burial teams also operated. Unlike other teams, they were static rather than mobile, working within their communities of residence and resupplied from neighbouring health facilities.

The analysis covered North Kivu, Ituri and South Kivu, with a population of approximately 6,300,000, from 6 August 2018 through 10 October 2019. That period comprised about 91 percent of all confirmed or probable Ebola cases during the epidemic, and all SDB alerts received during it were eligible for analysis.

Figure 1 represents the intended SDB process as a flowchart. Once a death was reported, the Alerts System was activated. For Ebola-positive deaths in an Ebola treatment centre, a dispatch alert went directly to the SDB team. For suspected cases in the community or other health facilities, the Ministry of Health and WHO Case Investigation Team first determined whether the decedent met the suspected-case definition.

If the decedent was determined to be a suspect case, an alert was sent to the closest SDB sub-coordination hub, which dispatched a team to the decedent’s location. Figure one traces alerts through the intended safe and dignified burial process, separating alerts from ETCs or transit centres from those originating in the community or other health facilities.

It shows how alerts were followed by testing or swab collection, body securing, environmental disinfection, and outcomes classified by EVD status, including non-case, confirmed case, and suspect. This matters because it makes visible where the SDB response succeeded, generated a failure alert, or had an unclear outcome.

Once on site, the SDB team was expected to collect an oral swab specimen for Ebola testing, secure the body with zero point five percent chlorine and a leak-proof body bag, and decontaminate the dwelling and the deceased’s belongings with zero point five percent chlorine.

The body was to be buried at a site agreed with relatives, in a grave at least two metres deep and covered by at least one and a half metres of earth to protect it from wild animals. For suspected cases, SDB was to happen before Ebola test results if the family consented; otherwise, the body was taken to a mortuary to await results.

Family representatives were invited to join the team in securing the body. Community engagement and accountability personnel trained volunteers, liaised with community and religious leaders and families, and assisted with feasible modifications to standard SDB.

The International Federation of the Red Cross and Red Crescent Societies maintained an individual, standardised Excel database of every SDB alert, meaning a request to deploy a team to assist with burial. The database recorded demographic details, alert and completion dates and times, whether key activities were completed, and reasons for unsuccessful or incomplete burial deployments.

Because the dataset did not distinguish mobile from static CEHRB SDBs supported by Civil Protection, the analysis considered three actors: Red Cross mobile teams, Red Cross-supported CEHRB, and Civil Protection mobile or CEHRB teams. After range and consistency checks, the SDB database was analysed to compute key performance indicators of intervention fidelity and performance.

The targets were set by the Red Cross team as aspirational benchmarks of success, not as quantitative epidemiological control requirements. Table one defines the authors’ key performance indicators for safe and dignified burial, separating process measures from output measures. Process targets include one hundred percent for responding to dispatch alerts, performing swabs, securing corpses, and disinfecting dwellings, while community or family nonacceptance has no defined target.

Output targets include eighty percent for conducting safe burials, successful responses, and completion within twenty-four hours, alongside one hundred percent within seventy-two hours; these were aspirational benchmarks rather than epidemiological requirements. The analysis modelled the binomial probability of unsuccessful SDB with a logit link function, using the explanatory variables defined for the study.

A generalised linear mixed model estimated the log odds of SDB failure and specified the sub-coordination hub as a random effect because observations within each hub were assumed to be correlated. The analysis first computed univariate associations and screened out variables with a p-value of zero point twenty or higher, then retained variables using significance, coefficient influence or Akaike Information Criterion goodness-of-fit rules.

In terms of timing, 89 percent of SDB responses were completed within 24 hours, exceeding the 80 percent target, while 99.9 percent were completed within 72 hours. From December 2018 onwards, the proportion of timely SDBs increased in both Ituri and North Kivu, but both provinces experienced large monthly fluctuations in successful SDBs throughout 2019.

Overall, 66 percent of all alerts were successfully responded to, falling short of the 80 percent target. Figure three links SDB performance with insecurity over time in Ituri and North Kivu. Panel A shows weekly percentages of successful responses in orange and responses completed within twenty-four hours in blue, alongside the eighty-percent target; the paper reports that eighty-nine percent were timely, while sixty-six percent were successful overall.

Panels B and C show weekly attacks against the EVD response and deaths from other recorded insecurity incidents, highlighting how the period of especially low success coincided with intensified violence. Reduced health facility coverage per population was associated with higher odds of SDB failure: the odds ratio was 1.57 when comparing the lowest- and highest-coverage health zones.

Suspensions of Ebola response activities and insecurity were associated with 1.61 and 1.73 times higher odds of failure, respectively. A local Ebola treatment facility, high mobile-telephony coverage and high radio coverage were also associated with higher odds of failure.

SDBs carried out by local community members under the static CEHRB programme had zero point two five times the odds of failure, while those mainly supported by Civil Protection teams had zero point six one times the odds of failure, compared with other SDBs. Community or family nonacceptance was strongly associated with SDB failure.

Table four reports univariate and multivariate odds ratios for SDB failure, using reference categories marked as one. In the adjusted model, children aged one to four years had an odds ratio of zero point seven three, while zones with fewer than twenty-five health facilities per one hundred thousand people had an odds ratio of one point five seven.

Epidemic and post-epidemic stages were associated with odds ratios of two point one four and three point six zero, respectively, and confirmed EVD cases in the previous six weeks had an odds ratio of zero point four nine. SDB teams almost ubiquitously responded to alerts and met timeliness targets, but the proportion of successful responses was below target.

Low performance was mainly concentrated from February to May 2019, a period with high-profile killings of Ebola responders, attacks against treatment centres and increased insecurity in health zones reporting Ebola transmission, particularly in Ituri. Performance was also lower when teams responded to suspected rather than confirmed Ebola deaths.

Despite these reductions, a separate paper in preparation found that SDB was associated with substantial reductions in Ebola transmission. The factor models used multiple adjustment and were subject to the well-described Table 2 fallacy, so they should be interpreted with caution.

Here, the model results should be interpreted with caution and used to identify factors for further evaluation and careful consideration in future interventions. A large-scale SDB service was feasible and usually timely, but success fell during insecurity and nonacceptance.

Locally managed community and Civil Protection teams had lower odds of failure, pointing toward more locally rooted response models.