“Diseases Desperate Grown”

Adam Kucharski began writing The Rules of Contagion: Why Things Spread – and Why They Stop before the Covid-19 outbreak in the West but was able to incorporate information about it prior to publication. Kucharski is an epidemiologist at the London School of Hygiene and Tropical Medicine where he analyzes infectious disease outbreaks. He didn’t start his career in the field. He first interned as an investment banker. The two fields have more in common than one might think, but more on that in a moment. 

Kucharski explains in easily comprehensible terms the mathematical bases for modeling contagions and making predictions about their course. Primary factors are the infectiousness of the pathogen (the likelihood of passing it on to a susceptible person), and what portions of the population subject to exposure fall into the categories Susceptible, Infectious, and Recovered. A bacterium or virus with a reproduction number greater than 1 (i.e. each infected person can be expected to infect more than 1 other person) will spread while less than 1 (even 0.9) will self-limit and burn out. The reproduction number changes as the percentage of Recovered in the population rises – assuming, as is usually the case, the Recovered have immunity. A population thereby can achieve herd immunity with a smallish percentage of Recovered in the case of a disease with an initially modest positive reproduction number (e.g. 1.1) since the Susceptible pool needn’t drop much to get the rate under 1. Similarly, one needn’t kill all malaria-carrying mosquitoes to eliminate malaria from a given region – just enough of them to get the reproduction number from individual to individual via the mosquito under 1. Naturally, getting something close to the correct numbers for these factors requires extensive investigation since they are different for each pathogen. He quotes the old line, “When you’ve seen one pandemic, you’ve seen one pandemic.” There are also public health measures (identifying and containing routes of transmission) that can impact the numbers. According to an NIH paper (Time-Varying COVID-19 Reproduction Number in the United States) the reproduction number in the U.S. for Covid-19 was 4.02 in March but declined to 1.51 only a month later. By comparison, the highly contagious smallpox virus (fortunately extinguished) had a historical reproduction number between 4 and 6 in susceptible populations. Kucharski does acknowledge the limits of mathematical modeling since even if the model is good the data are often insufficient to plug reliable numbers into it until after an epidemic has run its course, but even a limited benefit is still a benefit. 

Strangely (or perhaps not so strangely), very similar analytical methods can be applied to social contagions. This explains Kucharski’s hop from banking along with the move of marine ecologist George Sugihara to building predictive models for Deutsche Bank, and the move of Gary Slutkin from epidemiology in Africa to modeling crime in Chicago. Social behaviors from fashion fads to riots to labor strikes to rumors to gun violence to (oddly) obesity all are contagious to varying degrees and can be described in terms of disease models with special attention to routes of transmission and susceptible populations. Kucharski describes the financial contagion of 2006-2008 in these terms. 

Online memes so obviously follow similar rules that the term “going viral” dates to the very beginning of the internet. Offensive memes and tweets reproduce better than inoffensive ones. Microsoft learned this the hard way in 2016 when its conversational Artificial Intelligence named Tay went online on Twitter. It was designed to learn how to increase its Twitter following by interacting with other users and adjusting its conversational style accordingly. In only 16 hours Tay had to be shut down because it had quickly learned that being an offensive hateful jackass maximized its following. Fake news, unsurprisingly, has a higher reproduction rate than actual news since it can be tailored to appeal to (infect) susceptible people who will pass it along. Real news stories also can be infectious, however, if they share similar traits to the highly infectious fake ones. Hence, while there are 20,000 homicides in the US every year, just a handful dominate the news at any given time while the rest are all but ignored – those handful just touch the right buttons. 

None of this would have surprised Edward Bernays (Sigmund Freud’s nephew) whose 1928 book Propaganda remains influential in advertising and political circles. (Bernays intended no negative connotation with the word; he was just describing how it’s done.) Modern modelling simply helps make it more effective. 

Kucharski’s book is a good first step in understanding contagion in its literal disease sense. It also (hopefully) is useful for inoculating oneself (at least to a degree) against social contagions by conveying an understanding of the mechanisms. 

New Years Day – Epidemic