As we confront the pandemic, we’re flying blind, like pilots in 1930s aeroplanes flying through fog. That, at any rate, is what I take away from reading John Ioannidis, a Stanford professor who is an expert in epidemiology, population health, biomedical data science and statistics. Covid-19, he writes, in a startling article, “has been called a once-in-a-century pandemic. But it may also be a once-in-a-century evidence fiasco. At a time when everyone needs better information, from disease modellers and governments to people quarantined or just social distancing, we lack reliable evidence on how many people have been infected with Sars-CoV-2 [the official name for the virus] or who continue to become infected”.
Draconian countermeasures are being implemented all over the place. If the pandemic dissipates – for whatever reason – such extreme short-term measures may be bearable. But what if the thing just keeps going? And how can policymakers know they’re doing more good than harm?
A key principle of control engineering is that you have to be able to measure the variable you’re trying to control. In the case of Covid-19, we currently have no way of accurately measuring how we’re doing, because we’re not able to do enough testing of the population. Dammit, we’re still not even testing frontline medical staff.
I know, I know: this is hard; this thing came out of the blue; we can’t just magic up the resources needed to do extensive public testing out of thin air; etc. But at the same time, every sentient being in the government must know by now that we must find some way of measuring the thing we’re trying to control. How else will we know – other than by counting the number of desperate cases who show up needing intensive care – whether that curve is being flattened or not?
We need a magic bullet. And, miraculously, we seem to have one. It’s called a smartphone. Almost everybody has one nowadays. And – as security agencies, internet companies and mobile network owners know – these devices track your every movement, 24 hours a day, seven days a week, 365 days a year. They’re a bit like those electronic tags we put on some criminals to make sure they stick to the terms of their parole conditions. Except that the phones are smarter than tags. We can run apps on them that make them serve our public health purposes.
Interestingly, we were doing this in the UK way back in 2011 – in the “FluPhone” project, run by Prof Jon Crowcroft and Dr Eiko Yoneki of the Cambridge Computer Laboratory. They created an app that monitored flu-like symptoms by questioning the phones’ owners and then subsequently logged their physical proximity to other people they met. The pilot study, conducted by volunteers over a few months – which in an odd twist of fate coincided with the outbreak of swine flu – demonstrated convincingly how smartphones can provide data that would otherwise be unavailable to public health authorities.
So, in a way, this approach to monitoring behaviour during an epidemic is old hat, technically speaking. Not surprisingly, the Covid-19 crisis has led people in a number of countries to develop similar tools. In early February, for example, South Korean geeks wrote “Corona 100m”, an app that allows people to see the date a coronavirus patient was confirmed to have the disease, along with that patient’s nationality, gender, age and the places the patient visited. The person using the app can also see how close they are to coronavirus patients. It was launched on 11 February and had a million downloads in the first 17 days.
It looks as though the Chinese authorities have been using similar approaches almost from the outset of the pandemic. And the Israeli government authorised the country’s internal security agency to use mobile phone location data to help combat the virus. According to a New York Times report, the data will be used to retrace the movements of individuals who test positive for the virus, and identify others who should be quarantined. And I bet there are people in 10 Downing Street (and not just Dominic Cummings) who are actively thinking along those lines.
You can now see the quandary heading our way. We have a terrifying emergency on our hands. There’s a technology that could be really useful in providing a real-time measure of the effectiveness (or otherwise) of public policy. But it’s so intrusive that in “normal” times we would be very hesitant to adopt it. Present times are not normal, though, so we will probably adopt it – perhaps sensibly – “just for the duration of the crisis”. And then?
You know the answer because we’ve been here before – with 9/11. That emergency led a panicked US into building the surveillance state that Edward Snowden eventually exposed in 2013. Once a government commits to this stuff, it seems that there’s no going back. Hard choices lie ahead.
Glen Weyl and Jaron Lanier have written a striking essay for Wired which neatly ridicules the pretensions of artificial intelligence evangelists.
What goes up…
“Don’t feel sorry for the airlines” is the headline and gist of legal scholar Tim Wu’s terrific New York Times article arguing that companies that do share buybacks don’t deserve bailouts.
Pandemics from the past
Read Isaac Chotiner’s fascinating – and timely – New Yorker interview with Yale historian Frank M Snowden, who wrote the book on the history of pandemics.