Expecting data to defeat Covid-19 is unrealistic and risky, says Saba Mirza
The cheerleaders for big data seem to believe that any problem is solvable with enough numbers. The coronavirus is just the latest example of such a problem.
Data is certainly flooding in—about how the virus has spread globally, how it has multiplied within populations, the most common symptoms in different age groups, the effects of local health policies, the global death count, and how some countries have used expertise in big data to control the virus. Forecasting models predict how the virus can spread, and the consequences in sickness, death and economic damage of different policy options.
Governments and societies have put their trust in numbers, in the data generated by each person, hospital, country, academic institution and forecasting guru. However much there is, more seems to be needed to help the sick, identify the vulnerable and quarantine the infected.
In the UK, the digital arm of the NHS, NHSX, is reportedly working to launch a contact-tracing app, of the kind already deployed in some Asian countries. This will notify people if they have spent time close to a person who has tested positive for the coronavirus.
Apple and Google too have launched a joint initiative, as providers of the two main mobile operating systems, to enable third-party apps to develop contact-tracing. The companies eventually plan to do away with third-party apps and directly collect data through their platforms to mitigate risks of data leaks.
Who’s watching?
During the lockdown, these measures may seem essential to resuming normal activity. But they come at a massive privacy cost to individuals.
Besides having access to people’s health data, these apps will also record their movements and interactions with others. Just as Facebook knows its users’ online connections, Apple, Google and the NHS will effectively know people’s networks in the physical world—our coworkers, neighbours, those we commute with, and so on.
The NHS is focused on health data, but Apple and Google have troves of all manner of data on their users. These data, when combined with health data and location-based tracking, will massively extend the reach of these tech giants into their customers’ private lives.
There is also no telling if this data collection will halt once Covid-19 is no longer a global threat, or if it will become routine. As the data will have both commercial and surveillance value, stopping collecting it could prove less straightforward than it might currently seem.
This is not to deny that data is an essential component of the research effort on coronavirus. The quickest route to drugs and vaccines, for example, lies through sharing the results of lab studies and clinical trials as widely as possible. How willing pharmaceutical companies will be to do this remains an open question.
But an over-reliance on data brings its own problems. Oxford University’s Big Data Institute has estimated that for a contact-tracing app to work, 60 per cent of the population needs to sign up and report accurate symptoms.
If the app remains voluntary, reaching this threshold will be difficult. But making the app mandatory raises huge issues around consent and privacy.
It is also difficult to estimate how much such an app will reduce the risk of contagion. The system will certainly not be foolproof.
It might flag people to isolate unnecessarily if, for example, they come within three metres of one another but remain in separate rooms. Or it might miss cases, as symptoms vary considerably among the infected, not to mention those who are asymptomatic.
Clutching at technology
Even if enough people sign up and the app helps to contain the spread of the virus, for the sick and vulnerable, key issues remain. There are persistent reports that frontline healthcare workers lack adequate personal protective equipment across NHS facilities in both Covid-19 wards and beyond. Intensive care units have struggled to find enough ventilators. Not enough people are getting tested.
An app will not solve these problems, nor many others.
And however good the data, it will only help government decision-making to a limited extent. Forecasts of the progress of the disease and the effect of measures to curb it will still be influenced by the models’ assumptions and oversights. Different models and datasets will still give different answers.
There will still be trade-offs between containing the disease, the immediate hardship caused by the lockdown—last week, the Food Foundation estimated that 3 million people in the UK had gone hungry at some point since it began—and long-term economic damage. Meanwhile, countries that had successfully contained the virus are seeing a resurgence.
Until a vaccine becomes available, the best government strategy will not be a straightforward question of ‘following the science’, or even of choosing one scientific model over another based on data generated by an app or otherwise.
It will be a complicated and difficult set of choices with underlying ethical values and principles. Clutching at technological solutions only shrouds these value judgments and shifts the focus away from problems that can be fixed, such as spending on public healthcare and providing welfare support to those in need.
Saba Mirza is a research assistant in the department of Science and Technology Studies at University College London