The author team extracted weekly state-level counts from 474 Nigeria Centre for Disease Control situation reports covering 2017–2026. They then compared years at phase-matched points per a pre-specified analysis plan, and performed post hoc comparisons of weekly incident versus published cumulative totals. The aim was practical: determine whether the reported rise in case fatality ratio (CFR) for Lassa fever between 2025 and 2026 reflected worsening clinical outcomes or artefacts of surveillance and reporting. The data source was public NCDC situation reports and the analysis code and repository are available at the project GitHub and Zenodo links provided in the source.
Investigators decomposed the year-on-year CFR change into two components: within-state changes and between-state redistribution of cases. This allowed them to distinguish whether a national CFR shift came from local increases in fatality, or from changes in which states contributed most cases. They also calculated tipping-point ascertainment shifts—an assessment of how many missed survivors would be required to reconcile differing series. The study followed a pre-specified plan and reported confidence intervals for key comparisons.
The team predefined 11 candidate mechanisms that could explain changing reported CFRs, ranging from true changes in clinical care to surveillance artefacts such as late-added deaths, variable testing volumes, and increased post-mortem sampling. Of those 11, two mechanisms could be evaluated fully from the published situation reports and five others could be evaluated in part. The source did not report sufficient detail to fully assess the remaining mechanisms, and the authors explicitly noted those limits.
A central finding emerged when analysts compared weekly incident counts to cumulative published totals. Seventy deaths were added retrospectively to cumulative figures by week 36, and these additions were not present in the weekly situation reports. Those 70 late-added deaths increased the published cumulative CFR by approximately 5.4 to 6.2 percentage points. The investigators calculated that to reconcile the higher cumulative CFR with the weekly incident series, an additional 279 survivors would need to be present among suspected-but-unconfirmed cases—equivalent to 6% of suspected cases not confirmed. The source data and calculations supporting these numbers are included in the project repository referenced by the authors.
When analyses focused on weekly incident counts at the phase-matched midpoint, 2026 did not show a meaningful difference from 2025: the point estimate was a +0.3 percentage-point change with a 95% confidence interval spanning -5.0 to 7.6. The decomposition returned a within-state component of -1.1 percentage points (95% CI -4.7 to 3.0) and a between-state component of +0.5 percentage points (95% CI -2.0 to 2.9). Both component intervals crossed zero. Data indicated no clear signal of increased mortality attributable to deteriorating clinical care when weekly incident reports were used as the basis for CFR calculation.
The authors concluded that the observed rise in the 2026 CFR was driven by deaths added to cumulative totals outside the weekly counts, and that published surveillance data cannot on their own attribute CFR changes to failing clinical care. The source explicitly stated that endemic zoonotic CFRs cannot serve as reliable clinical outcome measures unless reports include accompanying metadata: testing volumes, the timing and provenance of late-added deaths, and post-mortem sampling rates. The article noted that several candidate mechanisms could not be fully resolved because the routine public reports did not contain the necessary granular detail.
Limitations and outstanding questions remain concrete and data-centred. The study depended on publicly available NCDC situation reports and therefore could not evaluate mechanisms that require patient-level clinical data or laboratory-confirmation timelines—those details were not reported. The analysis quantified the reporting artefact (retrospective death additions) but did not, and could not from the available sources, determine why those deaths were retroactively added or whether they reflect delayed reporting, reclassification, or other administrative processes. The authors also did not claim causation between surveillance processes and patient outcomes because the required clinical and laboratory metadata were absent from the published reports.
Funding sources for the work were disclosed in the source and included US NSF-NIH-NIFA EEID, UK BBSRC, and the Birgitta Sintring Fellowship at Uppsala University. The study was posted as a medRxiv preprint on October 15, 2026. The analysis code and extracted data are archived and linked in the source material for readers who wish to reproduce or extend the work.