The Quiet Indicators That Often Precede the Next Pandemic

Signal

Three measurable indicators that have historically correlated with increased zoonotic spillover risk are currently elevated above baseline:

First, deforestation rates in the primary tropical and subtropical hotspots for novel zoonotic disease emergence — the Congo Basin, the Amazon, Southeast Asian lowland forests — have been above the 2015-2020 average for the past three consecutive years. Habitat fragmentation places human populations in closer contact with wildlife species that serve as reservoir hosts for viruses with pandemic potential.

Second, global live animal trade volumes — legal and estimated illegal — have increased post-COVID to levels comparable to the pre-pandemic peak, despite regulatory commitments from multiple countries to reform wildlife trade practices. The legal wildlife trade generates approximately $23 billion annually; the illegal trade is estimated at $7-23 billion. Both bring diverse wildlife species into contact with domestic animals and humans in ways that provide opportunities for viral spillover and adaptation.

Third, wild bird influenza surveillance data — maintained by the FAO and WHO's EMPRES monitoring system — shows an unprecedented diversity of H5 influenza subtypes in wild bird populations globally, with ongoing evolution toward configurations that increase mammalian receptor binding affinity.

Interpretation

These three signals point toward increased background probability of novel zoonotic emergence, though not toward a specific pathogen or timeline.

The deforestation signal operates on a longer timescale: the relationship between habitat fragmentation and zoonotic emergence is well-documented in retrospective studies but has a variable lead time measured in years to decades. The wildlife trade signal operates on a shorter timescale: illegal trade in live animals provides direct transmission pathways in high-density markets that compress the spillover and adaptation timeline.

The H5 influenza signal is the most immediately concerning. H5N1 influenza — which has maintained a case fatality rate above 50 percent in the limited human cases documented since 1997 — has expanded its host range significantly since 2022, infecting dairy cattle in the United States (a documented first in 2024), marine mammals, and a widening array of terrestrial mammals. The expansion of mammalian host range is a key precursor to increased human transmission risk; the adaptive mutations associated with mammalian infection are in some cases the same mutations associated with increased human-to-human transmissibility.

Scenario

Scenario A — the limited outbreak scenario — involves a novel spillover that produces limited human-to-human transmission and is contained by existing surveillance and response capacity. This is the most historically common outcome of zoonotic spillover events; most novel spillovers do not produce pandemic-scale transmission.

Scenario B — the moderate pandemic scenario — involves a spillover with sustained but moderate human-to-human transmission, comparable to the 2009 H1N1 pandemic, which caused significant illness and economic disruption but was managed within existing health system capacity in most countries.

Scenario C — the severe pandemic scenario — involves a spillover with high transmissibility and high severity comparable to the 1918 influenza pandemic. This scenario requires either a pathogen that combines these characteristics inherently or a rapid adaptive process during an initial period of spread that increases severity before containment is achieved.

Current H5N1 is a candidate for Scenario B or C because of its existing high severity in humans and the ongoing evolution toward increased mammalian transmission. The gap between Scenario B and Scenario C is large; its determination would depend heavily on specific adaptive mutations that are not predictable in advance.

Probability

Metaculus forecasts a 27 percent probability that a novel zoonotic disease will cause a public health emergency of international concern (PHEIC) declaration before the end of 2028. The same community forecasts a 14 percent probability of a disease causing more than 1 million deaths globally before 2030.

Kalshi was trading a contract on whether H5N1 influenza will achieve documented sustained human-to-human transmission — more than two sequential generational links — before 2028 at 31 percent. This is the threshold at which pandemic preparedness protocols at national health authorities are expected to activate.

These probabilities are not predictions of disaster; they are honest estimates of non-trivial risks that the evidence supports. A 27 percent probability of a PHEIC over three years implies that the absence of such an event would be somewhat surprising, not assured.

Indicators to Watch

— H5N1 poultry and mammal cases in the US: the CDC and USDA's monitoring of dairy herd infections provides the closest real-time signal for US human exposure risk — Wastewater surveillance: syndromic surveillance for novel respiratory pathogens via municipal wastewater sequencing, which identified early COVID signals in 2020 — Illegal wildlife trade market surveillance: TRAFFIC's monitoring of wet markets in Southeast Asia and Central Africa for high-risk species combinations — WHO GOARN network alerts: the Global Outbreak Alert and Response Network activations, which signal that national health authorities have detected unusual clusters — Genomic surveillance data sharing: the rate at which countries share novel pathogen sequences to GISAID is a leading indicator of whether early warning is functioning

The signals are not alarming in the sense of predicting an imminent pandemic. They are alarming in the sense that they indicate elevated background risk that the public health and policy community is not adequately communicating to decision-makers.

---

Dr. Amara Singh is a staff writer at The Auguro covering medicine, science, and public health. She holds an MD from Johns Hopkins and a PhD in epidemiology from Harvard.