Historically, pandemic preparedness has relied on outbreak lists that catalog viruses based on past epidemics and use that information to guide future response plans. While this approach has its merits, it is inherently reactive, designed to fight yesterday's pandemic rather than tomorrow's threat. In our new study, posted on the preprint server medRxiv, we propose a more proactive approach: one that categorizes pathogens based on shared epidemiological traits. By analyzing key viral characteristics, we've classified pandemic threats into distinct archetypes. This method could offer a more flexible and predictive framework for preparing against future outbreaks.
The retrospective nature of pandemic preparedness
Global health agencies such as the World Health Organization (WHO) maintain priority lists of dangerous pathogens. These lists help guide research funding, surveillance, and vaccine development, but they are inherently backward-looking. The focus is on viruses that have already demonstrated their potential for harm rather than those that may emerge in the future.
In our latest research, conducted in collaboration with John Edmunds and his team at the London School of Hygiene and Tropical Medicine, we explore a different approach. We analyzed 19 different viruses from the WHO priority list, along with multiple influenza strains and mpox. For those, we gathered 342 epidemiological estimates from 166 studies, focusing on traits that define how a virus behaves in an outbreak.
We examined various factors, for example: how easily a pathogen spreads, how long it takes for symptoms to develop, how quickly new cases arise, how deadly it is, and whether it can spread before symptoms appear. With this data, we applied machine learning algorithms to sort the pathogens into distinct clusters based on their epidemiological characteristics. What we have arrived at is a new way of thinking about pandemic threats, where viruses are grouped by behavior, rather than name or historical precedent.
The five pandemic archetypes
The analysis revealed that most viruses fall into one of the five broad epidemiological categories.
- Airborne pathogens with high transmission potential (e.g., SARS-CoV-2, influenza) – Fast-spreading and often transmissible before symptoms appear.
- Respiratory zoonoses with high fatality risk (e.g., MERS-CoV, H5N1) – Less contagious but highly deadly.
- Contact zoonoses with high fatality rates (e.g., Ebola, Marburg, Nipah) – Spreads via bodily fluids and animal-to-human spillover. Transmission is not rapid but is highly deadly.
- Contact zoonoses with presymptomatic transmission (e.g., Lassa fever, Mpox) – These pathogens spread before symptoms appear, with moderate transmission and fatality rates.
- Vector-borne pathogens capable of secondary human transmission (e.g., Zika) – These viruses jump between animals, insects, and humans.
More adaptable
Our new framework has the potential to change the way we think about pandemic preparedness. It is an adaptable tool that can be used to classify an emerging viral threat into an existing category and implement the appropriate strategies. Furthermore, as new epidemiological data emerges, these clustering methods can be refined and additional pathogens can be integrated.
By focusing on shared epidemiological traits rather than individual pathogens, public health preparedness can evolve from a static, reactive model to a more adaptable, forward-looking strategy for managing outbreaks
Read the preprint here
