On 05 May 2026, the World Health Organization (WHO) was alerted to a high-mortality outbreak of unknown illness in Mongbwalu Health Zone, Ituri Province, Democratic Republic of the Congo (DRC), including deaths among health workers.¹ The Institut national de recherche biomédicale confirmed Bundibugyo virus disease in 8 of 13 samples on 15 May 2026, and the country declared its 17th Ebola disease outbreak; the same day, Uganda confirmed an imported case in a Congolese traveller who died in Kampala.¹ On 17 May 2026, the WHO Director-General determined that the event constituted a Public Health Emergency of International Concern (PHEIC).² As of 16 May 2026, WHO reported eight laboratory-confirmed cases, 246 suspected cases and 80 suspected deaths across three Ituri health zones (Bunia, Rwampara and Mongbwalu), plus two imported laboratory-confirmed cases (one fatal) in Kampala.^1,2 As of 01 July 2026, the DRC outbreak had reached 1460 confirmed cases and 452 deaths, with cases reported across Ituri, North Kivu and South Kivu and continued cross-border risk involving Uganda.³

The 10-day interval between the initial alert (05 May) and laboratory confirmation (15 May) is an important operational anchor. During it, multiple signal streams were already active: Syndromic surge reports from Mongbwalu, clusters of health-worker deaths, cross-border mobility along the DRC–Uganda corridor and community mortality patterns.¹ Existing platforms captured many of these signals – the WHO Epidemic Intelligence from Open Sources (EIOS) initiative,⁴ the WHO African Region Integrated Disease Surveillance and Response (IDSR) framework, whose third edition formally integrates event-based surveillance,⁵ the Africa Centres for Disease Control and Prevention (Africa CDC) event-based surveillance framework,⁶ ProMED-mail,⁷ HealthMap⁸ and the WHO Early Warning, Alert and Response System (EWARS).⁹ The challenge was not signal capture but the interpretive integration of concurrent multi-source signals into a unified detection-decision workflow.

In this article, a signal denotes any data element from a heterogeneous source – laboratory, event-based, community, cross-border or open-source – potentially indicative of an emerging public health event prior to formal case classification. Signal integration failure denotes the systematic non-convergence of concurrent multi-source signals into a unified detection-decision workflow, resulting in delayed epidemic recognition despite pre-existing data. Both definitions are consistent with the WHO glossary for epidemic intelligence,⁴ the Africa CDC event-based surveillance framework⁶ and the companion methodological preprint.¹⁰

The Pan-African Real-time Epidemiological Intelligence System (PREIS) is a proposed methodological framework for reproducible epidemic intelligence workflows, designed to organise heterogeneous permitted data streams into standardised, auditable and reviewable intelligence products. Its full description – five-layer architecture, six design principles, a 16-field minimum data model, an eight-step reproducible pipeline, quality gates, the CORI operational cycle (Capture, Organise, Review, Inform), configurable risk-prioritisation and a prototype Ebola-specific module – is provided in a companion OSF preprint.¹⁰ The present communication summarises only what the case-application discussion requires and refers readers to the preprint for detail.

Pan-African Real-time Epidemiological Intelligence System consumes permitted inputs – WHO Disease Outbreak News and situation reports, Africa CDC bulletins, aggregate line lists and District Health Information Software 2 (DHIS2) exports where authorised and event-based signals where authorised – and produces provenance-tagged dashboards, prioritisation tables, risk maps and situation-summary templates. Current ingestion is from structured or semi-structured sources only; unstructured social-media crawling is on the roadmap and not yet operational.¹⁰ Any high or very high output requires qualified expert epidemiological review before being shared beyond the authorised team, and notification templates require explicit institutional authorisation prior to dissemination.¹⁰

Pan-African Real-time Epidemiological Intelligence System is not an autonomous open-source detection engine but an analytical intelligence layer that transforms already-captured, permitted signals into decision-oriented products. The distinction between detection capacity (identifying a signal from raw source data) and analytical capacity (transforming, contextualising and decision-orienting already-captured signals) is preserved throughout. Pan-African Real-time Epidemiological Intelligence System is designed to complement, not replace, established platforms including IDSR, EIOS, EWARS, ProMED, HealthMap and DHIS2, as summarised in Table 1; a fuller comparison appears under ‘Comparison with existing systems’ in the companion preprint.¹⁰

The 2026 outbreak offers an operational case for examining how a PREIS-type analytical layer could contribute to signal integration in a live PHEIC context. It is used here as an illustrative case only: We do not report the causal effect of any intervention, present validated detection-performance estimates or reproduce restricted operational data. Any operational deployment would require formal institutional authorisation, written data-sharing agreements, expert-review protocols and communication safeguards consistent with international regulations.^10,11

Four features of the 2026 outbreak inform the signal-integration discussion. Firstly, the initial alert (05 May) preceded laboratory confirmation (15 May) by 10 days.¹ Secondly, the affected geography spanned three Ituri health zones (Bunia, Rwampara and Mongbwalu) with concurrent cross-border risk into Uganda.^1,2 Thirdly, initial testing targeted the Zaire ebolavirus strain; identifying the Bundibugyo strain required specific sequencing, and no licensed Bundibugyo vaccine or therapeutic exists² – an absence of post-detection countermeasures that amplifies the value of early multi-signal detection without itself constituting a detection-gap argument. Fourthly, the concurrent signals during the interval – syndromic surges, health-worker infections, cross-border mobility – were captured by existing platforms but not integrated into a single decision product at the required speed.¹

Table 2 lists the principal signal types active during the outbreak and their relevance to interpretive integration. The framing is descriptive; no counterfactual claim is made about what a PREIS-type layer would have produced during the actual outbreak.

The 2026 Bundibugyo virus disease PHEIC illustrates three lessons for African epidemic intelligence. Firstly, and most importantly, the principal challenge was not signal availability: Signals were present and existing platforms captured many of them; the gap lay in integrating concurrent multi-source signals into a unified decision product at the required speed. Secondly, established platforms – IDSR, EIOS, EWARS, ProMED, HealthMap, and DHIS2 – provide valuable, complementary capacities; their limitations are specific configuration and analytical-layer gaps amenable to complementary intervention rather than replacement. Thirdly, locally maintained analytical layers, developed by and for African public health practitioners under authorised data governance, may help strengthen interpretive-integration capacity – consistent with the Africa CDC blueprint for early warning surveillance¹² and the priorities of Africa’s continental public health architecture.¹³

A framework such as PREIS should be understood as an adaptable analytical layer that may complement – not replace – established platforms.¹⁴ Its potential contribution is a reproducible workflow model that helps practitioners organise heterogeneous permitted signals into transparent, reviewable products. Realising this requires co-design with authorised institutions, formal data-sharing arrangements, prospective external validation against verified epidemiological ground truth and disease-specific peer-reviewed studies – none of which is complete at the time of writing, as the preprint makes explicit.¹⁰

Pan-African Real-time Epidemiological Intelligence System is presented as a proposed methodological and prototype framework. It is not, at the time of writing, an official platform, policy, product or endorsed tool of Africa CDC, the African Union, the World Health Organization (WHO), any Ministry of Health, any Regional Coordinating Centre or any Member State, unless future formal endorsement is obtained through appropriate institutional processes.¹⁰ References to Regional Coordinating Centres, IDSR, event-based surveillance (EBS), EIOS, EWARS, DHIS2 and other systems are made solely for methodological alignment, interoperability and comparative context; they imply no endorsement, affiliation, operational deployment, data ownership or official outbreak notification.

No individual-level human participant data are reported, and no confidential institutional dataset is reproduced; the case, death and geographic figures are drawn from publicly available WHO Disease Outbreak News and PHEIC determination statements, each cited with retrieval date. Analytical quality assurance is treated as an ethical requirement: In operational deployments, risk scores, novelty flags and notification templates should be reviewed by qualified epidemiologists before external sharing, and the framework should communicate uncertainty and distinguish official confirmed data from early attention signals.

The 2026 Bundibugyo virus disease PHEIC illustrates that Africa’s principal detection challenge is often not signal availability but the timely interpretive integration of concurrent multi-source signals into unified decision products. Locally maintained, reproducible analytical layers that complement – rather than replace – established frameworks may help strengthen this capacity, subject to co-design with authorised institutions, formal data-governance arrangements and prospective validation.

Pan-African Real-time Epidemiological Intelligence System is proposed as one such analytical layer; its full methodology, prototype logic and explicit limitations are set out in the companion preprint.¹⁰ Future work should prioritise co-design with African public health institutions and competent authorities, formal data-governance arrangements, prospective external validation, implementation research and disease-specific peer-reviewed studies.