When revolutions take place, things happen quickly. The year of Covid-19 is likely to be an example of how sudden pressure can lead to massive changes, and not necessarily all for the worse once all the consequences are taken into account. The pandemic and associated lockdowns have provided a wake-up call for the world in how things are done at every level.
We have seen in government responses a recognition that the market on its own does not provide. To ensure society does not suffer shocks even worse than a rise in excess deaths, governments around the world have decided economic help is important and even those who used to argue for the discipline of austerity have changed their minds.
The need to deal with an urgent problem has brought groups together to try to solve problems in ways they would not have considered before without the external pressure. Such situations provide a laboratory for ideas where it is possible to see quickly how things work and what problems seemingly good ideas can run into when they encounter reality. Take for example the issue of taking into account human behaviour when assessing the performance of epidemiological models or mobile apps designed to help limit the spread of the virus.
One problem that has taxed public policymakers when trying to analyse how to ease the lockdown is the realisation that techniques such as infection testing may still fail to be useful in managing transmission when other parts of the system do not work. At the same time, big changes to the way that research is carried out combined with the ability to see the impact of work in near real-time provides opportunities to see what is most effective.
The lessons from 2020 will take years to digest but even now we can see ways in which the people of the world can use the experience to move forward and take heart that much of what is achievable is down to cooperation rather than conflict.
Lesson 1: Open science
As the pandemic started to hit a coalition began to form, or rather an entire network of coalitions. Scientists around the world operating across disciplines from AI to genomics said they wanted to help. For example, the researchers working for the European Laboratory for Learning Intelligent Systems (ELLIS) saw a way of bringing together various groups using a series of online seminars to spread the word.
At the first ELLIS virtual meeting in early April, which kicked off with a thousand viewers, Bernhard Schölkopf, director of the department of empirical inference at the Max Planck Institute for Intelligent Systems in Tübingen, noted that an event had been planned that day at the Royal Society in London to celebrate the formation of a network of AI research teams across Europe. Instead, it had become a virtual meeting focused on the pandemic. “There’s a huge creative energy and urge to contribute to tackling this disease,” he said, pointing to what at the time were early-stage projects to look at the reasons for high mortality in different groups, and track infections without compromising privacy.
Other groups, such as the Usher Institute at the University of Edinburgh, have been running online seminars where healthcare experts around the world talk about their experiences and the lessons they have learned in responding to the pandemic.
Alongside online events, data sharing has also taken centre stage. Although there is already a long-term trend in favour of open-source data sets for use in research, the pandemic has accelerated their use. The European Bioinformatics Institute brought forward the launch of an online portal for anonymised genetic data from patients so that medical researchers could get a better idea of the crossover between susceptibility to the worst symptoms and their genetic make-up.
It joined other international efforts such as GISAID, which has collated genomic data from a variety of viruses since the emergence of H5N1 bird flu 15 years ago. The data proved instrumental in tracking mutations of Sars-Cov2 and in turn monitoring transmission patterns around the world.
Lesson 2: Fast therapy
At the start of August, Matthew Hallsworth, head of external relations at the National Institute for Health Research, claimed: “UK health research during the coronavirus pandemic has changed in ways which would have seemed impossible only six months ago.” He called on the sector to learn from the change induced by the rush to create treatments and vaccines for Covid-19 to improve the way in which medical research is conducted in the future.
It might have been a bit early to make that call. Even now, the medical-research community is still trying to evaluate what has worked and what has not worked so well or even failed outright. Driven by an influx of offers of help from many areas, particularly IT and AI research, the six months following the virus moving worldwide turned up many possible avenues of action, only some of which have borne fruit. That was to be expected as many aspects of the behaviour of Covid-19 in humans and why some drugs seem to work while others fail remain unclear.
The biggest change that seems likely is the increased use of adaptive trials like the UK’s Recovery. This evaluates multiple treatments at the same time and monitors results so that drugs showing no positive effect can be dropped in favour of promising alternatives.
Speed was of the essence in drug design. Although some modelling effort has been applied to trying to find as yet unknown compounds, the bulk of the research has gone into repurposing old drugs, on the basis that their side effects are known and they do not need as much testing. By the end of September, the European Medicines Agency alone had cleared more than 160 potential drugs for trials, practically all of them on existing drugs, although a number of them are fairly experimental, such as synthetic interferons.
To try to discover which of the multitude of pharmaceuticals might make a difference, various groups turned to machine learning, on the basis that these models could pore over medical databases and online archives of published papers to try to find connections with coronaviruses.
Some tried to find molecules that might interfere with the proteins that Sars-Cov2 uses to invade victim cells. However, in mining the literature, a worldwide team of scientists led by a group based at Northeastern University in Boston found that just one drug out of 77 promising candidates targeted any of those proteins. The rest all interacted with different parts of the massively complex signalling network that drive activity in living cells. Such networks form the backbone of what researchers call systems biology. The drugs identified can change the way in which proteins are produced in the cell which in turn bind with DNA to alter the behaviour of other parts of the network. This is important with a disease like Covid-19 because it can treat the conditions that lead to death, such as the cytokine storms that are the result of over-signalling pushing the immune system too far.
The ongoing problem for researchers is trying to work out which candidates are likely to show the best results given the enormous range of possible targets. One study of connections across the literature identified the drug Liponavir as a possible unsung hero in the fight against Covid-19. But, ultimately, the Recovery trial indicated no benefit from Liponavir when applied to Covid-19.
The situation for medicines like Liponavir may turn out to be similar to the one emerging for hydroxychloroquine, a treatment that has been on the good-news, bad-news rollercoaster since spring. The reality may be that its efficacy relies on what is known as stratified or personalised medicine: some recent computational genomics research has indicated that only people missing a certain genetic marker may see a benefit.
There are other issues for AI in medicine. The massive neural networks developed by the social-media are designed to work with masses of fuzzy data. Medical data on the other hand is sparse and highly susceptible to tiny changes in detail, which is why many of the trials for the more famous antivirals seem so contradictory. Researchers into computational medicine point to small differences in dosage, demographics and treatment protocols between trials as major causes for seemingly contradictory results. This may help explain why possible connections when identifying candidates lead to therapeutic dead ends. Ultimately, the sheer volume of data that AI-focused research teams will have to pore over in the wake of the pandemic will likely show which techniques lead to positive results.
Lesson 3: Present dangers
There are inevitably jobs that rely on physical presence. Short of controlling robots with the help of virtual-reality headsets, there is little that will change that. But a well-known curse of the office environment, and one to which British workers are often subjected more than their foreign neighbours, is that of presenteeism: the need to be visibly at a desk even though the amount of productive work does not increase with the hours spent in place.
The pandemic has done more to shift attitudes over how much work needs to be done in a shared office than numerous reports put together by management-advice groups. Surveys indicate that few people want to work from home all of the time, and for many it’s an uncomfortable experience given that they may live in conditions where there simply isn’t space to create a home version of the office. Even so, the surveys suggest there is a strong demand to cut down on commuting by working from home at least a few days a week. But now the question is over how well these more distributed teams are being managed.
The 20th report on health and wellbeing put together by the Chartered Institute of Personnel and Development found that pre-pandemic almost 90 per cent of members had observed presenteeism in their organisation over the past 12 months but just a third had taken steps to tackle it.
Research by RAND Europe suggests that presenteeism is mainly driven by stress, suboptimal mental health, lack of sleep and poor financial well-being. According to Christian van Stolk, executive vice president at the research group, these conditions explained more than half of presenteeism cases before the pandemic struck.
Working remotely does not in itself stop presenteeism manifesting: it just moves it around. A survey carried out by Opinium for Canada Life in the summer found almost half of British workers operating from home during the spring lockdown felt pressure to seem to be present and a third working despite feeling unwell.
Though it may take time, a focus on better metrics and techniques to monitor productivity and motivate staff without physical presence or ‘online now’ signs on teamworking software should improve the work-life balance as the pandemic eases.
Lesson 4: Makers make a mark
As the pandemic hit Europe, countries there found they were running low on protective clothing and masks for medical and care staff. Only weeks after donating PPE to China for use in Wuhan, another batch was air-freighted back to Italy as the load on intensive-care wards in the north of the country spiralled upward.
Demand quickly outstripped capacity and hospitals had to look for alternative sources of masks. Companies with access to machinery that could be easily retargeted switched to making equipment such as face masks as well as equipment to help treat patients with breathing difficulties.
In the UK, companies such as electronics distributor RS Components as well as numerous independent makers joined an initiative organised by the National 3D Printing Society to make the fittings for face visors, with final visors assembled by Igus UK before being sterilised and sent on to the NHS. The projects to deploy 3D printing were replicated around the world with organisations like the World Economic Forum trying to link suppliers to the initiatives.
Devon-based Luminous Show Technology, which no longer had a customer base via the stage shows it supplied with special effects, worked out a method for making the face-shield fittings using a combination of laser cutters and mechanical punches.
The emphasis has now shifted towards higher-volume production as injection-moulding companies stepped in, although 3D-printing specialist Photocentric says its technology is now able to mass-produce face shields and similar components.
The campaign has shown how volunteer groups can engage with a sector that has stringent standards. RS Components says it is continuing to work with the Office for Product Safety and Standards to clarify the guidance on requirements such as heat testing and the liability issues for volunteer groups so that if a similar issue surfaces, the process will be more straightforward.
Although the circumstances of the pandemic were unusual, the rapid response in manufacturing may indicate a route for onshore manufacturing to evolve: using a combination of robotic automation and flexible tooling to adjust to changes in demand rapidly.
Lesson 5: Economic reboot
In the shadow of pervasive disinformation, it is perhaps one of the most unfortunate names the World Economic Forum could have come up with for a plan to reshape the way economies work. The ‘Great Reset’ can just as easily conjure the same authoritarian overtones as the neoconservative New World Order promised by President George Bush in the wake of the 1990 Gulf War as it can convey an idea that conventional ‘neoliberal’ capitalism has run its course and needs a thorough shaking out in favour of a more balanced approach.
“The world must act jointly and swiftly to revamp all aspects of our societies and economies, from education to social contracts and working conditions… Every country, from the United States to China, must participate, and every industry, from oil and gas to tech, must be transformed. In short, we need a Great Reset of capitalism,” claimed WEF executive chairman Klaus Schwab at the July launch of the proposal.
Similar to the Agenda 21 programme devised by the United Nations decades earlier, those words have become a lightning rod for sceptics of anthropogenic climate change and various other groups who point to the role of the WEF, World Bank and the IMF in promoting the globalism that has in turn triggered numerous nationalistic movements around the world.
Taken at face value, the Great Reset is meant to be seen as an attempt to deal with social inequality as well as climate change by trying to adjust how capitalism works and not just address the problems exposed by the pandemic but those caused by the financial crisis of 2008. By 2014, even the IMF had accepted that its call for austerity programmes in the wake of the crash was a mistake and now recommends governments spend more on economic stimulus packages in order to recover from the pandemic, though it has not extended that advice to developing countries that are unable to borrow easily.
As robotic automation takes hold, some of it driven by social-distancing requirements, a deeper debate is going to have to take place over the role of work in society whenever fewer people will have the ability to earn a wage through full-time employment. Whether the Great Reset will help in that direction remains to be seen, but the pandemic has helped accelerate the process of change. The first summit, to take place in January 2021, will at least show the direction of travel.