Coronavirus: Don't forget data evolution and context

The rest of this series will look at how digital technology applies to the the outbreak of a novel Coronavirus (nCoV-2019) in China. How is engineering being deployed to help medical professionals combat its spread there and in the rest of the world?

But as coverage and understandable public interest – and concern - about the outbreak has intensified, it seems wise first to look at how the event is exposing some of technology’s limitations both in and of itself and how we interpret what it can do for us.

Three themes emerge.

Technology is not immediate. There is a tendency to believe that the recent rapid advances in processing power, algorithmic development, machine learning and artificial intelligence can now deliver almost instant answers to everything. The digital landscape has changed massively since the SARS outbreak in 2003. But while we may be able to do things more efficiently, some things still take time and that is particularly true of a virus. There is no ‘one size fits all’ here and, indeed, there are real dangers in believing that one analytical strategy and its particular use can be applied universally.

Second, as viruses spread and evolve, so do the data sets around them. We are not dealing with a situation of simple ‘garbage in, garbage out’ but rather one where new information is emerging daily (and there is the further complication that there were delays in early reporting of the outbreak). That means that the results even the wisest medical experts can produce are also going to change. Also, as you might imagine, their methodologies vary.

Third, a lot of data is being placed into the digital sphere, across social media, specialist sites and elsewhere. Here, the comparison with 17 years ago is especially profound. But the intentions behind various postings are many. Anyone might be able to access the data via a web browser – and its free availability is laudable – but its intended audiences can be very different. In some cases, experts are looking to inform the public but in others they are looking to contribute to and stimulate the debate taking place among themselves. These goals are laudable and necessary, but they do not always coexist harmoniously.

Let’s look at some examples in each case.

Perhaps one of the most alarming stories to spread across the media in the last couple of weeks – though arguably ‘alarmist’ is the better word – suggested that nCoV could lead to 65 million deaths worldwide. The experts are nowhere near being able to make that kind of projection, especially those behind the simulation from which that number was taken.

That simulation was an exercise called Event 201, organised for the World Economic Forum and The Bill and Melinda Gates Foundation in October 2019 by one of the leading specialists in pandemic modelling, the Johns Hopkins Centre for Health Security. Part of the purpose behind Event 201 was to present a series of challenging scenarios not in search of a doomsday number but rather, as Johns Hopkins has now been forced to clarify when the results were wrongly picked up, “to highlight preparedness and response challenges that would likely arise in a very severe pandemic”.

“Although our tabletop exercise included a mock novel coronavirus, the inputs we used for modeling the potential impact of that fictional virus are not similar to nCoV-2019,” the Center has added.

That qualification by Johns Hopkins introduces the second issue of data sets and how there are circumstances where they are indicative rather than definitive.

For the media and much of the public following nCoV-2109 online, another ‘worrying’ set of analyses has concerned the so-called R0 for nCoV-2019, the notation for its basic reproduction number. This gives an estimate of the average number of people to whom an infected individual will pass the virus. But a large part of the immediate intention behind the current release of this data is again being misunderstood.

About 10 or so experienced groups have released R0 estimates, though the first things to note are not only their acknowledged differences in methodology (“Tools are universal, methodologies are not,” as an EDA colleague likes to say), but also that these are still early results based on early data, and that the findings have not been peer-reviewed. Moreover, these groups are themselves updating R0 as more information becomes available. Heavy caveats.

As things stand the, the most important conclusion to be drawn from a consensus range, with a couple of outliers, that is between 2 and 3 is that the R0 for nCoV-2019 is greater than one and this means that the virus is continuing to spread So, authorities and medical professionals should continue to work towards better public health controls. More settled data will gradually emerge (and I’ve tried to avoid specifics here precisely because of that) but this is the main conclusion to be drawn from all these proliferating numbers right now.

Finally, let’s look at the environment in which all this information is being shared, changing, targeted.

Public awareness is important. The promotion of greater basic hygiene is, for now, the primary goal for the information that is being universally distributed about nCoV-2019. That is what you and I can do. But the dangers of being predisposed to a bit of paranoia while armed with a web browser should be obvious.

On Twitter right now, for example, you can follow a range of very smart experts in tracking and characterising these types of outbreak. A good number of them are doing their best to counsel against misinterpretation of their findings – they certainly are not peddling ‘fake news’. But you would prefer they did not spend too much of their time engaged in that.

As well as providing a public forum, social media serves as a useful way through which the virology community can exchange findings, data and ideas among themselves. It is worth noting that the battle against nCoV-2019 is very much a global one, with specialists toiling not just in China, but across the globe, all trying to bring knowledge, expertise and research into new technology-driven methodologies to the fight. China itself is trying to be far more proactive than in the past about getting useful information out there for foreign experts to work with. But you need to see the context in which information is being shared as equally important as the information itself.

You might think all this is obvious. But look at some of the panic that is spreading, too often in this case fuelled not by malice (although there are some idiots out there) but misunderstanding.

Rant over. To conclude, just consider one of the more positive aspects that this series will now move on to consider. Digital technologies have supported huge innovation in medical analysis and that which has taken place has built a framework that should help mitigate the nCoV-2019 outbreak far more quickly and effectively than has been the case in the past.

Next, let’s start to see what that framework looks like.