FIFA World Cup 2026 — International Infectious Disease Importation Risk

The 2026 FIFA World Cup brings 48 national teams to 11 US host cities over 38 days. Beyond the routine baseline of international arrivals that the US public health system already manages year-round, the tournament is expected to add on the order of 0.75 to 1.5 million additional inbound trips into the country.

The relevant question is not whether the tournament creates risk in absolute terms, but whether it creates excess risk relative to the existing baseline of international travel. That excess flow corresponds to roughly 5 to 11 percent above the concurrent June–July baseline overall, with larger relative increases on routes from the 48 participating nations.

This dashboard breaks that increment down by pathogen, source country, and US destination, drawing on the GLEAM mobility framework and OAG airline data.

Net-new visitors

0.75–1.5M

additional inbound trips that would not have occurred without the tournament

Qualified teams

national teams from across all six FIFA confederations

Excess share of baseline

5–11%

net-new arrivals as share of the 38-day baseline flow

US host cities

across nine states, 78 matches Jun 12 – Jul 19

The 5–11% increment averages across all source countries. The relative increase is larger from the 48 participating nations, which are estimated to account for roughly three-quarters of net-new visitors. The compositional shift in the arrival mix is what shapes pathogen-specific importation risk.

Explore the data

By pathogen →

Per-pathogen tables of expected imports by source country, with baseline and three WC-concentrated travel scenarios (+10%, +20%, +35%). Country-level maps and host-city routing.

Arboviruses Dengue · Chikungunya · Yellow Fever

Vaccine-preventable Measles · Pertussis · Mumps · Rubella

High consequence Mpox Clade I · Ebola · Marburg

Enteric / waterborne Cholera · Typhoid

12 pathogens · 40 source countries · 4 scenarios

By host city →

Select a US host city to see expected pathogen importations from each source country, ranked by local expected imports, plus the match calendar for the venue.

11 host cities · per-pathogen breakdown · match calendar

The World Cup is a bounded perturbation to a continuous system

Available projections estimate the tournament will generate on the order of 0.75 to 1.5 million net-new inbound trips into the US over the 38-day window. The range reflects different assumptions about visa friction, ticketing, and global demand, none of which are settled at this stage. These are trips that would not have occurred without the tournament, after netting out visits whose timing happens to coincide with the event.

The US currently sees about 11 million international arrivals per month, or roughly 14 million across a 38-day window. The World Cup's net-new flow is therefore on the order of 5 to 11% above the concurrent baseline overall. The aggregate masks substantial compositional change: roughly three-quarters of the net-new arrivals are expected from the 48 participating nations, which means routes from those countries carry larger relative increases than the overall figure suggests. Once domestic stadium attendance is included, total World Cup footfall is broadly comparable to the concurrent US concert and festival audience over the same dates.

Germany 2006, the most relevant precedent. Enhanced surveillance across 12 host cities found no significant effect on infectious disease epidemiology attributable to the tournament. Elevated counts of measles, norovirus, and Campylobacter were observed but explained by pre-existing trends. The empirical lesson is that the relevant risk increment is geographically specific and pathogen-specific, not aggregate.

Methodology

About this dashboard

1.Purpose and Scope

This framework estimates the risk of disease importation into the United States during the 2026 FIFA World Cup (WC2026), hosted across 11 US cities.

For each participating country, the framework computes the probability that infectious travelers reach the US under baseline travel conditions and under scenarios of elevated travel demand driven by WC2026 fan mobility.

The analysis covers twelve pathogens of public health concern grouped into four categories: arboviruses (Dengue, Chikungunya, Yellow Fever), vaccine-preventable diseases (Measles, Pertussis, Mumps, Rubella), high-consequence pathogens (Mpox Clade I, Ebola, Marburg), and enteric or waterborne diseases (Cholera, Typhoid). Results are produced at two geographic levels: the United States as a whole (country-level) and individual US host cities (city-level).

2.Data Sources and Active Case Computation

Active case estimates are derived from publicly available surveillance data accessed on May 8, 2026.

Surveillance-derived pathogens

For pathogens with active surveillance signals, active cases per month are estimated as the average monthly count over the last 6 months of available data with reported activity. If fewer than 6 months of historical data are available, the most recent month with reported activity is used as a fallback. Data sources:

Dengue — WHO Dengue Global Dashboard .
Chikungunya — ECDC Chikungunya Monthly Situation Report .
Yellow Fever — WHO Immunization Data Portal — Yellow Fever . A 0.001 reporting threshold is applied to per-country expected imports; 24 source countries below the threshold are excluded from the table, with combined cumulative contribution below 0.01 expected imports per month.
Measles — WHO Immunization Data Portal .
Pertussis — WHO Immunization Data Portal — Pertussis .
Mumps — WHO Immunization Data Portal — Mumps .
Rubella — WHO Immunization Data Portal — Rubella .
Mpox Clade I — WHO Mpox Global Dashboard . Active cases are estimated for countries with confirmed Community Transmission only.
Cholera — ECDC Cholera Monthly Situation Report .
Typhoid — WHO Immunization Data Portal — Typhoid .

Ebola and Marburg

Marburg is modeled as a fully hypothetical scenario — no confirmed active outbreak as of the data access date — with case counts set manually to represent a plausible outbreak size for risk planning.

Ebola is anchored to reported Bundibugyo ebolavirus activity in DR Congo. The scenario stipulates 1,000 active cases as a hypothetical upper bound on the current situation; it does not incorporate ground-level outbreak detail and is intended for risk planning rather than as a forecast.

3.Travel Model and Baseline Mobility

The underlying mobility model is GLEAM-EPIRisk (Balcan et al., 2010 ; Davis, Chinazzi et al., 2021 ), a metapopulation framework that combines international flight data and commuting networks to estimate the probability that an infectious individual in a source country reaches a destination. Origin-destination flight data is sourced from OAG (Official Airline Guide) scheduled seat data aggregated to the basin level (approximately 3,200 geographic units worldwide). Commuting data is derived from a gravity model calibrated to population and distance.

The baseline scenario uses OAG summer 2026 scheduled flight data with no WC2026 perturbation. This reflects typical June travel patterns between each participating country and the United States.

4.Excess Travel Scenarios

WC2026 is expected to generate substantial increases in international travel to the US. The framework models four scenarios:

Baseline: no excess travel above OAG scheduled flows.
+10%: total travel from the source country to the US increases by 10%.
+20%: total travel increases by 20%.
+35%: total travel increases by 35%.

How excess traffic is distributed

The excess volume (in passengers) is defined as a fraction of the total baseline flow from the source country to the US. For example, a +10% scenario for a country with 50,000 annual travelers to the US adds 5,000 extra travelers.

At the country level, the full excess volume is added to the source country's US-bound link in the flight network, increasing the overall importation probability proportionally.

At the city level, the excess is routed to the US host cities where the country's team plays group-stage matches. Cities where the team plays more matches receive a proportionally larger share of the excess traffic. The departure distribution across source basins (sub-national units) follows each basin's existing share of US-bound traffic. For cities with no existing direct air connection, the excess is distributed proportionally to each basin's total US-bound flow, creating new effective OD pathways.

This approach reflects the assumption that WC2026 fans travel to the specific cities where their national team plays, and that their departure patterns mirror existing connectivity from the source country.

Important note: only group-stage match cities are used for excess traffic redistribution. Knockout-stage venues are not included because the schedule is not deterministic at the time of scenario construction.

5.Outputs

Country-level risk block

For each pathogen and scenario, the framework computes the full probability distribution of imported cases P(K = k) into the US, from which the following user-facing quantities are derived:

Imports per month (technical: E[K]) — the expected number of imported infectious cases per month, on average. Most months will see fewer; occasional months will see more.
Probability of zero (technical: P(K = 0)) — the chance that no cases reach the US from the source country in a given month.
Probability of ≥1 (technical: P(K ≥ 1)) — the chance that at least one case reaches the US in a given month.
Excess imports per month at +5/+10/+20% — the difference in expected imports under a WC-concentrated scenario minus the baseline. The dimension of risk added by the event itself.

These quantities are computed using the GLEAM mobility framework, which integrates the number of active cases in the source country, the per-individual daily probability of flying to the US (derived from OAG origin-destination flight data), and the length of the traveling window of exposed carriers, alongside global flight and commuting networks.

City-level relative risk

For each pathogen and scenario, the framework computes a relative risk score for every indexed US destination basin (~400 cities and metropolitan areas). The relative risk answers the question: given that at least one case reaches the US, what is the probability it arrives in each city?

Relative risk values are normalized to sum to 1 across all US destinations and are reported for all indexed cities, not just WC2026 host cities. Non-host cities can still emerge as relevant importation destinations through airline routing and connectivity.

Why relative risk and not absolute city-level probabilities? Absolute importation probabilities at the city level are affected by factors not captured in the travel model, including match attendance patterns, fan mobility within the US, and on-ground transmission dynamics. Reporting relative risk avoids overstating the precision of the model and focuses attention on the comparative ranking of cities.

6.Interpretation Guide

Results should be interpreted in the context of the following assumptions and limitations:

The model captures flight-based mobility, including indirect routing through non-US hubs, since it works with origin-destination passenger flows. It does not account for domestic overland travel (e.g. by car) within source countries or after arrival in the US.
Active case counts reflect the best available surveillance data as of May 8, 2026, and may underestimate true burden due to underreporting.
The incubation period is used as a proxy for the traveling window of exposed carriers. Cases that become symptomatic before departure are assumed not to travel.
Excess traffic scenarios are approximations. Actual WC2026 travel patterns will depend on team performance, ticket availability, and booking behavior.
Knockout-stage cities are excluded from excess traffic redistribution because the schedule is not deterministic at scenario-building time.

7.References

Balcan D, Gonçalves B, Hu H, Ramasco JJ, Colizza V, Vespignani A. Modeling the spatial spread of infectious diseases: the GLobal Epidemic and Mobility computational model. Journal of Computational Science, 1(3):132–145 (2010). link
Davis JT, Chinazzi M, Perra N, Mu K, Pastore y Piontti A, Ajelli M, Dean NE, Gioannini C, Litvinova M, Merler S, Rossi L, Sun K, Xiong X, Longini IM, Halloran ME, Viboud C, Vespignani A. Cryptic transmission of SARS-CoV-2 and the first COVID-19 wave. Nature, 600:127–132 (2021). link