nabu.science
MethodologyAboutContact

 

  • Snapshot
  • Key Claims
  • Strengths & Limitations
  • Commentary
  • From the Authors
  • Author Note / Field Feedback
  • Dimension breakdown
  • Impact Profile
Free · About 5 minutes
Report #NBU-08B37D15·Framework v5.13·Pipeline v0.4·QA update 19 Nov 2026·Evaluated 19 Nov 2026·OECD: 3.3 Health sciences·Subfield: Public Health, Environmental and Occupational Health·Basis: Full text·Type: Cross-sectional study

The anthropogenic fingerprint on emerging infectious diseases

Rory Gibb et al.Rory Gibb, Sadie J. Ryan, David M. Pigott et al.·2026·Nature·10.1038/s41586-026-11058-6 ↗

Primary claim

We show that outbreak risks are typically highest in mosaic landscapes where people and livestock live alongside forests and fragmented ecosystems.

Tentative · reported-event associations against a log-population background do not independently identify true outbreak risk
In brief

A broad comparison of outbreak geographies reveals heterogeneous disease associations, but its spatial design cannot identify the causal risk and reporting effects claimed.

ReliabilityCaveats4 of 4 checks completeRead the entry ↓
Fragile — the claim's weakest requirement is that the effect exists at all. Experimental. Every point below is traceable to a quoted passage, but this forecast has not yet been validated against known replication outcomes.
Abstract

Abstract Emerging infectious diseases are a hallmark of the Anthropocene 1,2 . Human-driven ecosystem changes and human–wildlife–livestock contact can drive the emergence of zoonotic and vector-borne diseases 3–5 , but how much these processes shape landscapes of outbreak risk is poorly understood, beyond a few well-studied systems 6–8 . Here we consolidate 58,319 outbreak event records for 32 diseases and systematically test how 16 hypothesized social and environmental drivers impact outbreak geographies, while accounting for detection and reporting biases. We show that outbreak risks are typically highest in mosaic landscapes where people and livestock live alongside forests and fragmented ecosystems. These combined factors, along with long-term declines in precipitation, share strong impacts across several vector-borne diseases (for example, dengue, Lyme disease and zoonotic arboviruses). By contrast, directly transmitted zoonoses (for example, Ebola and mpox) share few common drivers, and the impacts of most other anthropogenic pressures (for example, deforestation, climate warming and agricultural intensification) vary widely between diseases. Most consistently, the observed geography of outbreaks is shaped by healthcare access: reporting declines by a median of 32% (range across diseases: 1.2–96.7%) for each additional hour’s travel time from a health facility. Our findings underscore that infectious disease spillover and emergence are multi-causal, and that no one-size-fits-all strategy can prevent epidemics and pandemics. Ecosystem-based public health interventions 9 should always follow system-specific evidence, and be paired with greater investment in health systems and One Health pathogen surveillance.

Key Claims

Primary finding · Tentative

We show that outbreak risks are typically highest in mosaic landscapes where people and livestock live alongside forests and fragmented ecosystems.

Fragile — the claim's weakest requirement is that the effect exists at all. Experimental. Every point below is traceable to a quoted passage, but this forecast has not yet been validated against known replication outcomes.

reported-event associations against a log-population background do not independently identify true outbreak risk

Abstract; Figs. 2c, 3, 4b,e
Quote verified
Secondary · Tentative

[Directly] transmitted zoonoses (for example, Ebola and mpox) share few common drivers, and the impacts of most other anthropogenic pressures (for example, deforestation, climate warming and agricultural intensification) vary widely between diseases.

sparse and uneven outbreak samples limit inference from few detectable associations to few shared drivers

Abstract; Figs. 3, 4c,f
Quote verified
Secondary · Tentative

Most consistently, the observed geography of outbreaks is shaped by healthcare access: reporting declines by a median of 32% (range across diseases: 1.2–96.7%) for each additional hour’s travel time from a health facility.

case-background odds combine infection and detection processes rather than separately identifying reporting probability

Abstract; Main Text p. 1306; Fig. 5a
Quote verified
Secondary · Tentative

We find evidence of a strong and consistent anthropogenic fingerprint on vector-borne diseases, suggesting that over the coming decades, risk is likely to increase unless there are substantial shifts in human activities.

temporally unaligned spatial associations do not identify future outbreak-risk trajectories

Conclusion
Quote verified
Secondary · Tentative

a consistent fingerprint of long-term precipitation drying on dengue outbreak risk across the Americas, Africa and Asia

spatial associations lack temporal identification and the regional dengue samples conflict with the full-database count

Outbreak drivers differ between disease systems

Derived from the full evaluation — not a separate score.

Strengths · Limitations

Strengths

#contribution

Broad comparison beyond first emergence events

The study harmonizes outbreak records for 32 diseases and compares their wider geographies rather than only locations of first documented emergence. This supports a delineated comparison of shared and disease-specific associations.

↳ Introduction; Fig. 1; Extended Data Table 1

#methodological rigour

Sensitivity checks probe key analytical choices

Three hypothesis-selection thresholds test dependence on expert-selected covariates. Comparisons with incidence models for four US arboviruses examine information loss from reducing case counts to outbreak occurrences.

↳ Methods; Extended Data Figs. 5–7

#reporting

Figures explain the sequential model comparisons

The presentation follows the addition of spatial structure and detection proxies before moving to disease-specific and transmission-group comparisons. This makes changes in inferred associations visible.

↳ Figs. 2–5 and legends

Limitations

#methodological rigour

Background weighting changes the population reference

Background locations are sampled in proportion to log-transformed population rather than population itself. This does not provide the population-proportional reference invoked when interpreting the headline risk comparisons.

↳ Methods, Statistical modelling; Extended Data Fig. 2

#methodological rigour

Spatial associations exceed their identification limits

Temporally unaligned covariates do not establish exposure precedence, and travel-time associations do not separately identify reporting probability. Future-risk projections and immediate intervention effects consequently go beyond the spatial evidence.

↳ Abstract; Conclusion; Methods, Statistical modelling; Extended Data Table 2

#methodological rigour

Regional sample counts exceed database totals

Extended Data Fig. 8 gives 16,115 regional dengue records against 12,668 in the full-database table, and 3,522 yellow-fever records against 1,123. The correct samples underlying these regional comparisons remain unresolved.

↳ Extended Data Fig. 8; Extended Data Table 1

Does this match your read? Add your assessment ↓

Nabu Commentary

The cross-disease harmonization and comparison with earlier emergence-event research provide a meaningful contribution, while Figs. 2–5 make the analytical sequence accessible. Methodological Rigour is constrained by the log-population background reference, temporally unaligned covariates, and attribution of observed-event associations to causal risk or reporting effects. Hypothesis-threshold sensitivity analyses and the US arbovirus incidence comparison support descriptive use but do not resolve those identification problems. The discussion engages competing literature and many limitations, yet the categorical prevention conclusion exceeds what sparse directly transmitted zoonosis data establish. Regional sample-count conflicts remain unresolved reporting uncertainties rather than demonstrated reversals of the fitted results.

Evaluate another paper

Free · About 5 minutes · Research preview

Author Note / Field Feedback

Nabu’s assessment, alongside the field’s view.

Are you an author of this paper?

Dimension breakdown

Evidence Score

Out of 5
012345
3.5±0.2

Contribution
4.1

The harmonized cross-disease comparison extends earlier first-emergence analyses to broader outbreak geographies and examines shared versus disease-specific associations. Uneven disease coverage and uncertainty in separating reporting from infection constrain the advance without eliminating it.

Introduction; Figs. 1–5; Conclusion
Methodological Rigour
3.1

Spatial models, structured hypothesis selection, and incidence comparisons provide evidence for descriptive associations, but the log-population background, temporal misalignment, and causal attribution substantially constrain inference. Conflicting regional sample counts add unresolved reporting uncertainty without a separate reliability cap.

Methods, Statistical modelling and Limitations of data and methodology; Extended Data Figs. 7–8; Extended Data Tables 1–2
Reporting
3.7

The progression from pooled models to disease-specific comparisons is followable, with figures explaining the analytical sequence. Interpreting event odds as risk or reporting still requires care, although the underlying identification defect is priced in Methodological Rigour.

Figs. 2–5 and legends; Abstract
Positioning
3.5

The discussion engages competing ecological explanations and traces numerous measurement limitations to particular disease systems. It does not sufficiently carry the low power for directly transmitted zoonoses into the categorical prevention conclusion.

Introduction; Shared drivers differ by pathogen transmission mode; Methods, Limitations of data and methodology; Abstract

Reliability

Caveats4 of 4 checks


Results checkNoted

Extended Data Table 1’s per-disease counts sum to 57,418, 901 fewer than the 58,319 reported in Methods; the gap is 1.5% of records and does not affect the fitted results.

  • "(Latin America, n = 3,782; Africa, n = 2,271; Asia & Pacific, n = 10,062)" (Extended Data Fig. 8) vs "Dengue fever | | Dengue virus (Flaviviridae) | Virus | Literature | 12,668 | 1985-2015" (Extended Data Table 1, dengue fever row). The regional dengue counts sum to 16,115, exceeding the stated full-database count of 12,668 by 3,447. The effect on the result is unclear.
Reporting and transparencyClean

Funding sources with a funder-role statement, competing interests (including CEPI grants and government advisory roles), ethics status (no human-subjects data), and data and code availability (Zenodo archive, with restricted datasets available on request) are all declared.

Conflict of interest: declared · Ethics approval: declared · Preregistration: not stated · Data availability: declared · Code availability: declared

Reference IntegrityClean

31 of 32 checkable references verified

33 references in manuscript 1 are books, websites or datasets — counted, but not index-checkable

References not verified automatically are resolved manually within 24 hours.

Post-Publication RecordClean

No retraction or expression of concern found in Crossref (includes Retraction Watch).

Last checked: 19 Nov 2026

Sources: Crossref (incl. Retraction Watch) ✓

Impact Profile

Impact areaHealth · Infection · Other unspecified infectious diseases
Type of advanceEmpirical
Route to useNamed
AdvanceExtends
BeneficiariesCatch-all cause
CitationsToo early
Policy · patentNot yet traced
ReachedNot reachedNot assessable or unverified
ReadinessValidated
Foundational
Controlled
Piloted
Validated
Implementable
FoundationalImplementable

The primary claim is supported by 58,319 real-world outbreak events across 169 countries from multiple independent surveillance systems, with sensitivity analyses, though evidence remains observational.

From the paper
“Ecosystem-based public health interventions should always follow system-specific evidence, and be paired with greater investment in health systems and One Health pathogen surveillance.”

The paper explicitly names public health and health systems as the actors who should apply its findings to tailor interventions and surveillance investment.

Pathway

Public health agencies and One Health programmes would use the disease-specific driver findings to target ecosystem-based interventions and healthcare access investment in high-risk mosaic landscapes.

Barrier

The associations are observational with residual detection bias, and drivers vary by disease, so translating them into effective tailored interventions requires system-specific causal evidence.

Evaluated 19 Nov 2026 · GBD 2023 · glm_v1_3_3 · impact_profile_v1_3_3

Report an issue →

AI-generated, human-governed. Something look off? Contact us to request a review.

Keep reading

Related evaluations

Other papers Nabu has evaluated blind, starting with the same field.

  1. Evaluation of energy density and levels of fat, sugar, and salt in processed and ultra-processed food with existing and novel nutrient profiling models to identify unhealthy food and drink: cross sectional analysis of a nationally representative database of food and drink items

    Dicken, van Tulleken, Brown, et al. · 2026 · Health sciences

    Primary claim

    Food and drink products that, according to UK government dietary guidance, contain excess amounts of harmful nutrients are not identified by current systems for regulating UK food and drink (FSA NPM and front-of-pack traffic lights labelling system).

    Tentative
    Evidence Score
    3.1±0.2
    Reliability
    Caveats
  2. The Association of Physical Activity and Sedentary Behavior with Maternal and Cord Blood Anti-Oxidative Capacity and HDL Functionality: Findings of DALI Study

    Zafaranieh, Stadler, Pammer, et al. · 2023 · Health sciences

    Primary claim

    MVPA at <20 weeks and 24-28 weeks of gestation were positively associated with anti-oxidative capacity, as well as PON-1 activity of HDL in cord blood.

    Tentative
    Evidence Score
    2.7±0.2
    Reliability
    Caveats
  3. Physical activity and sedentary time across pregnancy and associations with neonatal weight, adiposity and cord blood parameters: a secondary analysis of the DALI study

    Dieberger, Obermayer-Pietsch, Harreiter, et al. · 2023 · Health sciences

    Primary claim

    Both higher maternal mean MVPA and increasing MVPA were associated with lower offspring FM% in males (-0.520%; 95% CI: -1.011%, -0.031% and -4.649%; -7.876%, -1.432% respectively).

    Tentative
    Evidence Score
    3.1±0.2
    Reliability
    Caveats

How the rubric works →