Computer technique could help gauge when pandemic is ‘over’
At the start of 2022, nearly two years after Covid was declared a pandemic by the World Health Organization, experts are pondering a big question: when is a pandemic “over”?
So what’s the answer? What criteria should be used to determine the “end” of the Covid pandemic phase? These are deceptively simple questions and there are no easy answers.
I am a computer scientist who studies the development of ontologies. In computer science, ontologies are a way to formally structure the knowledge of a domain, with its entities, relationships and constraints, so that a computer can process them in various applications and help humans to be more precise.
Ontologies can uncover insights that have been overlooked so far: in one case, an ontology identified two additional functional domains in phosphatases (a group of enzymes) and a new domain architecture of part of the enzyme. Ontologies also underpin Google’s Knowledge Graph which sits behind these knowledge panels on the right side of a search result.
Applying ontologies to the questions I posed initially is helpful. This approach helps to clarify why it is difficult to specify an endpoint at which a pandemic can be declared “over”. The process involves collecting definitions and characterizations from experts in the field, such as epidemiologists and infectious disease scientists, consulting relevant research and other ontologies, and investigating the nature of what is the “X” entity.
“X”, here, would be the pandemic itself – not just an abbreviated definition, but an examination of the properties of this entity. Such precise characterization of the “X” will also reveal when an entity is “not an X”. For example, if X = house, a property of houses is that they must all have a roof; if an object has no roof, it is definitely not a house.
With these characteristics in hand, a precise and formal specification can be formulated, aided by additional methods and tools. From there, the what or when of “X” – the pandemic is over or not – would logically follow. If not, it will at least be possible to explain why things are not so simple.
This kind of precision complements the efforts of health experts, helping humans be more precise and communicate more precisely. This forces us to make implicit assumptions explicit and clarifies where disagreements may lie.
Definitions and diagrams
I conducted an ontological analysis of the “pandemic”. First, I needed to come up with definitions of a pandemic.
Informally, an epidemic is an event in which there are multiple instances of an infectious disease in organisms, for a limited time, which affects a community of said organisms living in a region. A pandemic, at a minimum, expands the region where infections take place.
Read more: When will the COVID-19 pandemic end? 4 essential readings on past pandemics and what the future may bring
Then, I was inspired by existing fundamental ontologies. This contains generic categories such as “object”, “process” and “quality”. I have also used domain ontologies, which contain domain-specific entities, such as infectious diseases. Among other resources, I consulted the Infectious Diseases Ontology and the Descriptive Ontology for Linguistic and Cognitive Engineering.
First, I aligned the “pandemic” to a fundamental ontology, using a decision tree to simplify the process. This helped determine what kind of thing and what generic category is “pandemic”:
(1) East [pandemic] something happening or happening? Yes (enduring, i.e. something that unfolds in time, rather than being entirely present).
(2) Are you able to attend or participate in [a pandemic]? Yes (event).
(3) East [a pandemic] atomic, that is, has no subdivisions and has a definite endpoint? No (accomplishment).
The word “accomplishment” may sound strange here. But, in this context, it is clear that a pandemic is a temporal entity with a limited lifespan and which will evolve – that is, it will cease to be a pandemic and become epidemic again, as stated in this diagram.
Next, I looked at the characteristics of a pandemic described in the literature. A complete list is described in an article by American infectious disease specialists published in 2009 during the global outbreak of the H1N1 influenza virus. They put together eight characteristics of a pandemic.
Read more: New COVID data: South Africa has reached pandemic recovery stage
I have listed and evaluated them from an ontological point of view:
Wide geographic extension. This is an imprecise characteristic – whether fuzzy in the mathematical sense or estimated by other means: there is no sharp threshold where “broad” begins or ends.
Disease movement: there is transmission from place to place and it can be traced. A yes/no characteristic, but it could be made categorical or with ranges of slowness or speed of movement.
High attack rate and explosiveness, or: many people are affected in a short time. Many, short, fast – all indicate inaccuracy.
Minimum population immunity: Immunity is relative. You have it to some extent for some or all variants of the infectious agent, and likewise for the population. This is an inherently fuzzy feature.
New: A yes/no function, but we could add “partial”.
Infectiousness: it must be infectious (excluding non-infectious things, like obesity), so a clear yes/no.
Contagion: it can be from person to person or by other means. This property includes human-to-human, human-animal intermediate (eg, fleas, rats), and human-environment (including: water, as in the case of cholera), and their associated aspects .
Severity: Historically, the term “pandemic” has been applied more often to serious illnesses or those with high mortality rates (eg HIV/AIDS) than to milder ones. This has a certain subjectivity and therefore can be unclear.
Imprecise boundary properties annoy epidemiologists because they can lead to results that differ from their prediction models. But from my ontologist’s point of view, we are getting somewhere with these properties. On the computational side, automated reasoning with fuzzy features is possible.
COVID, at least in early 2020, easily ticked all eight boxes. A sufficiently automated reasoner would have classified this situation as a pandemic. But now, at the beginning of 2022? Severity (point 8) has largely decreased and immunity (point 4) has increased. Point 5 – are there any worse variants of concern to come – is the million dollar question. A more ontological analysis is needed.
Highlight the difficulties
Ontologically speaking, a pandemic is therefore an event (“fulfillment”) that unfolds in time. To be classified as a pandemic, there are a number of features that are not all clear and for which the imprecise boundaries have not all been defined. Conversely, this implies that classifying the event as “non-pandemic” is equally imprecise.
It’s not a complete answer as to what a pandemic is ontologically, but it does highlight the difficulties of calling it “over” – and illustrates that there will be disagreements at this subject.