What are the odds of dying in the next pandemic?
Evidence quality 4.5/5
Eight-dimension review score against the quality rubric . Each dimension scored 1–5.
- D1 Source grounding
- 4/5
- D2 Source authority
- 5/5
- D3 Arithmetic
- 3/5
- D4 Uncertainty
- 4/5
- D5 Scope
- 5/5
- D6 Prose
- 5/5
- D7 Perception honesty
- 5/5
- D8 Caveat completeness
- 5/5
Lifetime probability · lifetime, global adult
1 in 208
0.5% lifetime chance
Most people underestimate this.
range 1 in 1,000 to 1 in 50
● your factors — click this risk ▾ to reveal
≈ As likely as
Perceived
Post-COVID perception of pandemic risk is shaped almost entirely by recency. Most adults in high-income countries now acknowledge that pandemics are "possible" or even "likely" within their lifetime, a sharp shift from pre-2020 polling where pandemic risk ranked well below terrorism, plane crashes, and violent crime. But the perceived probability has a peculiar structure: it is simultaneously higher than the actuarial estimate for any given year (because COVID is still vivid) and lower than the cumulative lifetime estimate (because most people assume "the next one" is decades away and medical science will handle it). The net effect is a perception that roughly tracks reality at the population level but fails badly on the tails — underweighting the possibility of a pathogen more lethal than SARS-CoV-2 and overweighting the possibility of an exact COVID repeat.
Rough estimate: 41.2% of US adults report being afraid or very afraid of a new pandemic or epidemic (Chapman Survey 2024)
Actual
~2% annual probability of a COVID-scale pandemic (Marani et al. PNAS 2021, pre-correction); ~0.5-1% annual probability post-correction; ~0.3-1.9% for a 1918-scale event
global, novel pandemic event probability
Show derivation
This is a forward-looking estimate, distinct from the retrospective covid-death- cumulative entry. The calculation proceeds in two steps. (1) Annual probability of a pandemic occurring: Marani et al. (PNAS 2021) estimated ~2% per year for a COVID- scale event, but a 2023 correction noted that a coding error inflated the probabilities. Post-correction estimates and independent analyses (CGDEV, Disease Control Priorities) converge on a range of roughly 0.5-1.5% per year for a pandemic causing >1 million deaths globally. Using a midpoint of ~0.75% per year. (2) Conditional mortality: COVID-19 killed roughly 1 in 400 global adults over its acute phase (see covid-death-cumulative). A future pandemic could be more or less lethal; the historical range spans 1918 influenza (~1 in 30 global population) to H1N1 2009 (~1 in 10,000). Using a conditional death probability of ~1 in 150 global adults per pandemic event (geometric mean of the historical range, reflecting both improved medical countermeasures and the possibility of a more transmissible or lethal pathogen). Combined: per-year probability of dying in a pandemic ≈ 0.0075 x (1/150) ≈ 5.0 x 10^-5. Compounded over 59 remaining adult years: 1 - (1 - 0.000050)^59 ≈ 0.00295. However, Marani et al. also found that pandemic frequency is increasing — roughly threefold in the next few decades due to zoonotic spillover acceleration. Adjusting upward by ~1.6x for the increasing trend gives ~0.0048, or roughly 1 in 210. The uncertainty band is wide because both the occurrence probability and the conditional mortality are deeply uncertain.
Caveats: This entry is forward-looking and therefore inherently more uncertain than retro…
This entry is forward-looking and therefore inherently more uncertain than retrospective entries. The central estimate of ~1 in 210 lifetime risk rests on two deeply uncertain inputs: the annual probability of a pandemic occurring (~0.75% post-correction, plausibly 0.5-3.3% depending on methodology and trend adjustment) and the conditional mortality per event (~1 in 150 global adults, with a historical range spanning 1 in 10,000 to 1 in 30). Small changes in either input produce large changes in the lifetime figure, which is why the uncertainty band runs from 0.001 (1 in 1,000) to 0.02 (1 in 50). The Marani et al. 2023 correction is a cautionary note about the fragility of these estimates: a single coding error halved the headline probability. The entry is distinct from covid-death-cumulative (which is retrospective) and from seasonal-influenza entries (which cover endemic rather than pandemic mortality). Candidate pathogens for the "next pandemic" include H5N1 avian influenza (which has shown sustained mammalian transmission in US dairy herds as of 2024-2025), novel coronaviruses, antimicrobial-resistant bacteria (the AMR pathway), and the catch-all "Disease X" of WHO priority pathogen planning. The entry makes no prediction about which pathogen or when — it simply converts the historical and statistical record into a lifetime probability envelope.
Risks at similar odds
Other risks with roughly the same likelihood — useful for calibration.
COVID-19
What are the odds of dying from COVID-19 over the course of the pandemic and endemic era?
Unsafe imported products
What are the odds of being harmed by an unsafe imported consumer product?
Childhood cancer diagnosis
What are the odds of a child being diagnosed with cancer before age 20?
Counterfeit medicine
What are the odds of being harmed or killed by a counterfeit or substandard medicine?
Recently viewed on this device
Stored locally — clear anytime.
Pick challenger
The statistical framework for pandemic recurrence is younger than most readers assume. Marani, Katul, Pan, and Parolari published the first rigorous extreme-value analysis of historical pandemic frequency in PNAS in August 2021, assembling a dataset of major epidemics from 1600 to the present and finding that a COVID-scale pandemic had an annual occurrence probability of roughly 2% — implying a ~38% chance within any given 20-year window. A 2023 correction identified a coding error that roughly halved the probabilities, bringing the post-correction estimate to approximately 0.75-1% per year. Independent analyses from the Center for Global Development (2.5-3.3% per year, using a broader methodology) and the Disease Control Priorities project (4.2% per year for a pandemic causing ~10 million deaths) bracket the corrected Marani figure from above. The convergence across methods is the important signal: regardless of framework, the annual probability of a major pandemic is somewhere in the 0.5-3% range, and the cumulative probability over a 25-year window is more likely than not.
The harder question is not whether a pandemic will occur but how lethal it will be. COVID-19 killed roughly 1 in 400 global adults over its acute phase (2020-2022), making it far less lethal than the 1918 influenza pandemic (~1 in 30 global population) but far more lethal than the 2009 H1N1 pandemic (~1 in 10,000). The candidate pathogen list for the next event — H5N1 avian influenza, which has demonstrated sustained mammalian transmission in US dairy cattle; novel coronaviruses from the bat-pangolin- civet reservoir; antimicrobial-resistant bacteria whose resistance profiles are outrunning the antibiotic pipeline — spans a similar lethality range. Converting the occurrence probability (~0.75% per year) and the conditional mortality (geometric mean of the historical range, ~1 in 150 per event) into a lifetime figure gives a forward- looking risk of roughly 1 in 210 for a global adult dying in a future pandemic. That is comparable to the retrospective COVID-19 cumulative figure and roughly five times higher than the lifetime odds of dying from seasonal influenza.
The trend line is the detail that separates this entry from a simple actuarial exercise. Marani et al. found that the rate of novel pathogen emergence has been increasing roughly threefold over the past 50 years, driven by deforestation, urbanisation, factory farming, bushmeat trade, and climate-driven range expansion of disease vectors. The 2023 correction did not affect this directional finding. If the trend continues — and no structural intervention in zoonotic spillover prevention materialises — the annual pandemic probability in 2050 will be meaningfully higher than in 2025, and the lifetime risk for a 20-year-old today will be higher than the headline number suggests. This is why the entry is tagged underrated: not because the public is unaware that pandemics happen, but because the forward-looking probability envelope is larger than most intuitions accommodate, and the trend is moving in the wrong direction.
Claim ledger
Every number below is what each source reported, with the verbatim quote we relied on and how we arrived at our figure. Click any link to verify directly.
-
[1] Proceedings of the National Academy of Sciences / Marani M, Katul GG, Pan WK, Parolari AJ — Intensity and frequency of extreme novel epidemics
Intensity and frequency of extreme novel epidemics- Statistic
Probability of a pandemic with similar impact to COVID-19: ~2% per year (pre-correction); probability of a 1918-scale pandemic: 0.3-1.9% per year; pandemic frequency increasing roughly threefold in coming decades due to accelerating zoonotic spillover- Excerpt
“"The probability of a pandemic with similar impact as COVID-19 is about 2% in any year, meaning that someone born in the year 2000 would have about a 38% chance of experiencing one by now. [...] The probability of novel disease outbreaks will likely grow three-fold in the next few decades." ”
- Source data from
- 2021-08-31
- Accessed
- 2026-04-18 · archived copy
- Calculation
- Marani et al. assembled a global dataset of epidemics from 1600 to present and used extreme-value statistics to estimate occurrence probabilities. The headline ~2%/year for a COVID-scale event was the most-cited result. A 2023 correction (PNAS 120(19):e2302169120) identified a code error that inflated probabilities by roughly 2x, reducing the COVID-scale annual probability to roughly 0.75-1%. The corrected estimates are used in this entry's calculation. The finding that pandemic frequency is increasing ~3x is based on the accelerating rate of novel pathogen spillover events over the past 50 years and is directionally supported by independent analyses (CEPI, WHO priority pathogen reviews).
- Independence
- Marani et al. is the primary statistical analysis of historical pandemic frequency. It is methodologically independent of the WHO and NCBI/Disease Control Priorities sources, which use different frameworks (expert elicitation, epidemiological modelling) rather than extreme-value statistics.
-
[2] Proceedings of the National Academy of Sciences — Correction for Marani et al., Intensity and frequency of extreme novel epidemics
Correction for Marani et al., Intensity and frequency of extreme novel epidemics- Statistic
Code erroneously computed number of epidemics per year by summing events in two subsequent years, resulting in inflated probabilities and lower mean recurrence intervals- Excerpt
“"The code erroneously computed the number of epidemics in one year by summing the number of events in two subsequent years. This resulted in an inflated number of events/year and, as a consequence, in larger probabilities of occurrence and lower mean recurrence intervals." ”
- Source data from
- 2023-05-02
- Accessed
- 2026-04-18 · archived copy
- Calculation
- The 2023 correction is critical context for interpreting the Marani et al. 2021 headline figures. The coding error roughly doubled the estimated annual probability, meaning the widely cited "2% per year" should be read as closer to 0.75-1% per year post-correction. This entry uses the corrected range. The directional finding — that pandemic frequency is increasing — was not affected by the correction.
- Independence
- This is a correction to the Marani et al. 2021 paper, not an independent source. Included because the correction materially changes the headline probability used in the normalised calculation.
-
[3] Disease Control Priorities, Third Edition / Jamison DT, Gelband H, et al. — Pandemics: Risks, Impacts, and Mitigation
Pandemics: Risks, Impacts, and Mitigation- Statistic
In any given year, approximately 4.2% probability of a respiratory pandemic causing ~10 million deaths; 35% probability in a given decade; 66% probability over 25 years- Excerpt
“"In any given year there is a roughly 4.2 percent probability of a respiratory pandemic causing approximately 10 million deaths, amounting to a 35 percent probability in a given decade, and a 66 percent probability over a 25-year period." ”
- Source data from
- 2017-11-27
- Accessed
- 2026-04-18 · archived copy
- Calculation
- The Disease Control Priorities estimate of ~4.2%/year for a pandemic causing ~10 million deaths is higher than the Marani post-correction estimate because it uses a different methodology (expert elicitation and historical analogy rather than extreme-value statistics) and a different threshold (~10 million deaths rather than COVID-scale ~18 million excess). This source anchors the upper end of the uncertainty band and validates the order of magnitude: both frameworks agree that a major pandemic within a 25-year window is more likely than not.
- Independence
- Methodologically independent of Marani et al. — uses expert elicitation and epidemiological modelling rather than extreme-value statistics on historical data. Published before COVID-19, so it is also independent of any pandemic-recency bias.
-
[4] Center for Global Development — How Big Is the Risk of Epidemics, Really?
How Big Is the Risk of Epidemics, Really?- Statistic
Annual probability of a zoonotic spillover pandemic of COVID-19 magnitude or larger: 2.5-3.3%; 22-28% chance within 10 years; 47-57% chance within 25 years- Excerpt
“"The probability of a future zoonotic spillover event resulting in a pandemic of COVID-19 magnitude or larger is estimated between 2.5-3.3% annually." ”
- Source data from
- 2021-09-01
- Accessed
- 2026-04-18 · archived copy
- Calculation
- CGDEV's analysis synthesises multiple pandemic-frequency estimates and arrives at 2.5-3.3% per year for a COVID-scale event. This is higher than the Marani post- correction estimate (0.75-1%) because CGDEV incorporates the accelerating trend in zoonotic spillover and uses a broader definition of "COVID-scale." The 47-57% probability within 25 years is the figure most useful for the lifetime calculation. This entry uses the lower end of the range (closer to the corrected Marani estimate) for the central calculation and the CGDEV range for the upper uncertainty bound.
- Independence
- Independent of Marani — CGDEV synthesises a broader literature and uses different modelling assumptions. Partially dependent on the Disease Control Priorities framework as one input.