A pox on virtual plagues

'Corrupted Blood': valuable scientific tool, or a waste of journal space?

Bill Hanage 30 September 2007

Carnage: players dropping like flies (Screenshot courtesy of Tweakers)

While arguably less interesting, the real world is probably more relevant to questions of how to handle epidemics within it

Sometimes when I look at my behaviour in the spirit of cool, rational assessment proper to all scientists, I am forced to conclude that I am showing alarming curmudgeonly symptoms.

Over the summer, Jenny (’s presiding angel) drew my attention to an article in Lancet Infectious Diseases. On reading it, I promptly dissolved into a heap of dudgeon, spluttering “I don’t be-lieve it!” – hardly the measured response one would hope for.

The subject was an outbreak of a disease called ‘corrupted blood’ arising from a giant winged demon thing called Hakkar. Originally confined to the remote area of Zul’Gurub, it rapidly spread to population centres where it wreaked havoc, causing enormous mortality. In fact, some unfortunate individuals were killed several times by it.

This last intriguing and unusual feature was made possible by the fact that this is not a real disease, as the eagle-eyed among you might have noticed. The entire outbreak afflicted characters in the World of Warcraft, one of the Massively Multiplayer Online Role Playing Games (MMORPGs) which have largely replaced and amplified games that, in my youth, led to no self-respecting nerd being without a 20-sided die at all times. In this pretend world, death is not final and players can start again. And an epidemic can be halted, not by vaccination or other control measures, but by rebooting the servers involved.

I actually heard about this outbreak when it was happening (in late 2005) from a student who, as the flipside of his remarkable programming ability, seems to exist largely in online form. We had a giggle about it and idly wondered how one would go about estimating the essential parameters that determine the course of epidemics, such as the basic reproductive ratio R0, which is defined as the average number of new infections caused by a single infected host in a wholly susceptible population.

While we then put such notions aside, the authors of the Lancet Infectious Diseases article, Lofgren and Feffernan, were clearly excited. The upshot of the paper is the potential of MMORPGs for modelling real world epidemics. And to put their money where their mouth is, they have even tried to estimate some such parameters, although they report them slightly strangely. (For example they state a reproductive rate of 100 per hour, apparently confusing this with R0, which is not the same thing at all.)

There are superficial similarities with some real-world properties of epidemics. For instance, some non-player characters were infectious but could not be killed, becoming effectively asymptomatic carriers. The differences, however, are huge. There is of course no acquired immunity, and contact patterns in the game world are, to say the least, somewhat different from the real one. Moreover, epidemiologists already devote a great deal of time and effort to simulating the real world, rather than one containing demons, wizards and resurrection. While arguably less interesting, the real world is probably more relevant to questions of how to handle epidemics within it. For example, Ferguson and colleagues carefully simulated the entire population of the US and UK in order to understand the impact of different interventions on the course of a new ‘flu pandemic.

According to Lofgren and Fefferman, studies like Ferguson’s are flawed because they use, of necessity, data on human movements that are derived from the pre-outbreak situation – whereas a major epidemic will almost certainly alter human behaviour. This is not an unreasonable caveat. However, to suggest that we can gain insights into plague-besieged human behavior by looking at what happens in a fantasy world is, I think, ridiculous. As they say themselves, “ ‘resurrection’ most certainly allows riskier behaviours”. Well, no shit. I can see no reason why we should place more credence on observations gained from World of Warcraft than on those of human behaviour gleaned during previously well-studied real epidemics such as SARS.

This paper disturbs me for another reason. There is a place for such speculative and undeniably entertaining riffs on science – springs to mind. Granted, Lancet Infectious Diseases did the article under its ‘Personal View’ strand. But to me, in my irascibility the flaws are too great. It was not apparently reviewed carefully; for example, someone should have picked up the confusion over R0 which, I’m afraid, also makes me question the authors’ competence). I ask, not for the first time, what the Lancet family of journals want to be: broadsheet or tabloid. A more egregious case of the Lancet publishing total and utter nonsense in the pursuit of a cheap headline was a letter suggesting that SARS could come from outer space (in which case, pray tell, where are the complex eukaryotic organisms out there beyond the atmosphere in which such a virus could replicate?). More famously, and with direr consequences, was the much-discussed publication of research which led to the MMR controversy.

One really has to question, yet again, the Lancet’s decision to publish this ‘corrupted blood’ article. There is carefully done, laboriously parameterised research performed on disease in the real world which kills real, non-resurrectable people, and that research is rarely well understood by clinicians. So for me the decision to report this nonsense in a journal which purports to be among the best in its field reeks of decadence.