Note: This article appeared in The Santa Fe New Mexican on September 8, 2014.
By Ben Althouse, Omidyar Postdoctoral Fellow, Santa Fe Institute
In the fourth grade I read Richard Preston’s contagion thriller The Hot Zone and became fascinated with the idea that an organism so small as to require a powerful microscope to view and sophisticated tests to discover could wreak such havoc on individuals and societies. This fascination led me to pre-med at the University of Washington. There, I fell in love with the beauty of mathematics and was delighted to discover that I could combine these passions, viruses and mathematics, to model the spread of infectious diseases in entire populations. Eventually that led me here, to the Santa Fe Institute, where I am fortunate enough to be able to pursue my interest in epidemiology in an environment where ideas combine continually into useful new ideas.
Epidemiology is the study of the patterns and causes of disease in populations of individuals. This population-level approach is what makes it different from medicine: while your doctor’s primary concern is the well-being of the patient sitting in her office or lying on her operating table, the epidemiologist’s concern is the general well-being of the population, on average and as a whole. Of course, the two professions work hand in hand (the physician’s recommendation to her patient to quit smoking is based on population studies of tobacco use), but each is focused on fundamentally different scales and processes. Just as the physician and the epidemiologist have complementary but different jobs, infectious disease epidemiology (what I study at the Santa Fe Institute) concerns itself with problems and processes distinct from that of classical, or chronic disease epidemiology.
A hundred years ago, Sir Ronald Ross, a British physician and army officer, first described the details of the malaria life cycle within the mosquito. (This work eventually earned him a Nobel prize. Until that point we weren’t sure malaria was spread by mosquitoes.) He also developed a groundbreaking mathematical theory of the spread of malaria epidemics. Anchoring his theory was the idea of dependent happenings: “Different kinds of happenings may be separated into two classes, namely (a) those in which the frequency of the happening is independent of the number of individuals already affected; and (b) those in which the frequency of the happening depends upon this quantity.” In other words, my risk of developing lung cancer after years of smoking is independent of the number of smokers in the population (a). But my risk of getting a cold in the dead of winter when there are lots of other sick people is much greater than in the summer when there are few sick people (b). That’s a dependent happening.
This is a subtle, but important distinction, because it implies that changes in my actions can indirectly affect those around me without them doing anything. This is why vaccines are so effective: not only is a vaccinated individual protected against disease, but if the overall level of immune individuals in a population gets to a high enough level, there are so few susceptible individuals left that the pathogen is unable to spread. Infectious disease epidemiologists, physicians, and public health practitioners eradicated the terrible disease smallpox from the globe 35 years ago using this principle. In 2014, an individual’s risk of being infected with smallpox is zero because it is dependent on there being other cases, of which there are none.
For many infectious diseases though, we do not have effective vaccines, and we have to rely on other measures to understand and control their transmission. Dependent happenings, this dependence on the infectious or immune status of the people I interact with, can complicate the way we observe patterns of disease spread in populations. My risk of becoming infected with the flu depends not just on what I do, but on many interacting factors: my immune system (whether I was vaccinated this season, whether I’m currently infected with another pathogen, the entire history of what other flu strains I have been infected with in the past, my genetics), my personal behavior (hand washing, diet, exercise), the current level of flu in the population, who I contact during my day (coworkers, family, friends), where I contact them and for how long (for a minute at my desk, for two hours at the dinner table), and their own immune systems. This can muddle the link of Person A infecting Person B, creating patterns in numbers of cases that are hard to explain directly.
Infectious disease epidemiology, with its dependent happenings and myriad influencing factors, is a highly complex system. To understand and hopefully control a pathogen’s spread in a population, we need some of the tools of mathematics and complex systems science.
With my Santa Fe Institute colleague Sam Scarpino, I have been studying the current resurgence of whooping cough worldwide. 2012 saw over 48,000 cases of whooping cough – the most since 1955 – and the death rate in infants was three times that of the rest of the population. There are several factors that could be contributing to the resurgence, but what we’re learning is that the presence of individuals able to transmit Bordetella pertussis (the whooping cough bug), but that do not show symptoms (so called asymptomatically infected individuals), could be the primary culprit in increasing the overall level of pertussis in the population – and thus increasing the number of whooping cough cases. Thus, the number of observed cases is dependent on the total number of observed and unobserved cases in the population. (Put another way, there are a bunch of us who have pertussis and are able to infect others, but don’t know it.)
Other research I have recently been involved with is related to the dengue fever and Chikungunya viruses, both widely circulating in the Caribbean and in parts of Florida, and both causing a lot of illness and death. What makes these viruses difficult to control is that they are not transmitted directly from human to human; rather, they are dependent on the mosquito for transmission. Developing theory and building mathematical models to understand and predict where dengue or Chikungunya will go next must consider the complexity posed by the role of the mosquito. Lucky for us, we not only have Sir Ronald Ross’ theory of dependent happenings, but also his mathematical models, upon which almost all modern models of mosquito-borne pathogens are based.
Keeping people healthy has kept scientists busy for thousands of years. It is a highly complex problem embedded in overlapping complex systems, and the more we learn, the more complex it becomes. We must draw from epidemiology, biology, sociology, mathematics, and economics to start to understand the dependent happenings that kill and maim us. Considering the full complexity of the problem offers opportunities to make the world a safer place. I’m glad to be a part of it.
This column is the latest in the "Science in a Complex World" series written by researchers at the Santa Fe Institute and published in The Santa Fe New Mexican.