You can’t measure everything: Chaos Theory and Epidemiology

Have you ever had one of those moments where your mind is blown? It happened to me the other night as I was reading up on different theories and how they apply to epidemiology. Epidemiology — the study of that which comes upon the people — doesn’t have one unified theory of how it works. Instead, epidemiologists work with different theories to try and explain the phenomena of diseases and conditions and how they spread through populations. For example, one colleague of mine picked the Broken Windows Theory. That theory states that a building with broken windows is likely to have other windows broken because, hey, what’s one more broken window? When it comes to diseases, if all we see around us is unhealthy habits and unhealthy people, being healthy ourselves will not make a dent in the whole thing. So we “break one more window” and eat a twinkie or two.

I was browsing through different theories myself to try and see what other theories could be applied to epidemiology. One theory that I looked at was Game Theory. That theory is a little harder to explain. It has a lot to do with how rational, thinking actors behave within a system. Applied to epidemiology, game theory looks at how we choose to do things that benefit us, or harm others, and when we choose to do them. For example, a smoker who sees a “no smoking” sign may choose to smoke after they weigh all the possibilities. What are the chances of being caught? And, if caught, what are the chances of a severe punishment for not following the sign? So, in this example, the smoker plays the game of smoking and weighs the possible outcomes of breaking the rule.

What blew my mind was a paper by Dr. Pierre Philip from the University of Montreal. His paper talks about Chaos Theory and how it applies to epidemiology. To understand this better, you need a primer on what Chaos Theory is. Here’s a quick video:

The main gist of the theory is that complex systems — like our biological systems — behave differently based on very minute differences between them at the time that the system got started. For example, look at twins. They’re genetically the same, but two twins can be very different from each other in a lot of ways because of the tiny differences when they were born and all the different inputs along the way. Take away the inputs along the way, and the very fact that they were not born at the same time and don’t occupy the same space can explain the differences between them.

We can apply this in epidemiology when we are unable to accurately predict outbreaks of influenza (or any other infectious disease). We may have all the data available to us, but tiny differences in the initial conditions can produce very different outbreaks. Consider the 2009 H1N1 pandemic. That virus was created when a swine influenza strain and a human influenza strain found themselves in the same host (probably a pig in Mexico). Those two strains shared genes and a “novel” flu virus was created, one for which none of us had acquired immunity to. That virus was passed on to a person who passed it on to another person, and so on and so forth.

But what if we repeat the scenario again? Do we still get the H1N1 pandemic? What if the viruses share genes that are not conducive to antigenic “novelty”? Then that virus doesn’t spread to the index case and that index case doesn’t give it to the next person. The pandemic doesn’t even cause an epidemic. Or what if the pig doesn’t come into contact with the index case? What if? What if? What if?

There are two concepts in Chaos Theory that we need to be aware of in order to try and wrap our heads around it. The first is determinism. Determinism, in this context, means that the outcome of the system is going to happen no matter what. But it’s not like you cannot have different outcomes. As I wrote above, a change in the initial conditions of the system (e.g. the outbreak and ensuing pandemic) causes a change in the outcome. Even small changes at the molecular level of the virus could turn the whole thing on its head.

It’s pretty wild to think about that. It almost makes it an impossible task for us to predict when and where outbreaks will happen and how they’ll go. We can make educated guesses based on past experiences and probabilities of things happening, but we can’t measure everything. We cannot measure the initial condition of that virus, or the proximity of that pig to a susceptible human host. Not knowing that injects chaos into our understanding of epidemics and how they’ll go.

That’s not to say that we can’t make guesses as to how the epidemics will go, but we can’t be 100% certain. We epidemiologists were very surprised when the 2009 H1N1 pandemic turned out to be not-so-deadly. (Pleasantly surprised, mind you. Contrary to popular belief, I’m only joking when I say that people are supposed to die.) We had a virus for which none of us had immunity and which had previously caused a lot of disease and deaths. So it stood to reason that we might have a major pandemic coming on. We didn’t. A lot of people caught it, but, proportionately, it was a milder flu season than other non-pandemic ones. However, according to Chaos Theory (if I’m interpreting it correctly), one tiny change at the very beginning of the pandemic and everything could have gone really, really bad. Heck, one tiny change way back when the influenza virus first made it onto humans (probably thousands of years ago), and we could have been in a world of hurt.

Again, it’s those tiny little things that we don’t measure (or can’t measure) that make a big difference in the outcome of what we’re observing. That’s the main gist of this whole thing.

Speaking of hurt, thinking about all this really hurts my head. The concepts are difficult to understand, let alone fit them into how I see the world. So it is very possible that I just jumbled up Chaos Theory and didn’t really represent it well in this blog post. But, unlike some other people blogging out there in the big, bad world, I’m willing to correct my mistakes.

I'm a doctoral candidate in the Doctor of Public Health program at the Johns Hopkins University Bloomberg School of Public Health. All opinions posted here are my own, of course, and they do not necessarily reflect the opinions of my school, employers, friends, family, etc. Feel free to follow me on Twitter: @EpiRen