Impact of data and modelling on COVID-19

Search

Loading...

News

Latest News

Jun 22, 2020
by Ruth Thorlby
Impact of data and modelling on COVID-19

Sciana members and foundation staff take part in third webinar in COVID-19 series

Photo by Mika Baumeister on Unsplash

The role of data and modelling during the COVID-19 pandemic in Europe was up for discussion in the latest Sciana webinar.

On May 26 2020, Sciana members and foundation staff convened for an online meeting held under the Chatham House Rule. Ilona Kickbusch, chair of the Sciana Network, and John Lotherington, program director, moderated the conversation.

Members focused on differences in the scope and quality of the data available, the interaction of data and modelling on political decision-making, and the implications for leadership more broadly, at institutional, local, and national levels. Amid the extraordinary changes brought by the COVID-19 pandemic, one of the most striking is the public importance of data about the cases, deaths, and likely course of the virus across the world.

Websites collecting and publishing data across countries have emerged, from academia and the media. Meanwhile, visualisations that make complex ideas simple have quickly gained a global public audience. The story of how data and modelling have been used at a national level to drive decision making and public debate has some variation by country.

In the UK, some believe the decision to lockdown links to modelling produced by Prof Neil Ferguson and his team at Imperial College London on March 16. This study raised the alarm if there was no action, citing the risk of hospitals being overwhelmed and intensive care unit (ICU) capacity needing to be expanded even with suppression measures.

In the end, much of the additional ICU capacity in rapidly built field hospitals has remained unused, provoking debate about the accuracy of the modelling at the national level. Modelling has become less prominent in the national debate as the pandemic has progressed. At the hospital level, however, it is still important to understand when further peaks might arrive, particularly as the NHS regularly experiences winter pressures from seasonal flu and other respiratory illnesses. Local modelling is considered more accurate.

There are daily updates of national-level cases and deaths in the UK (e.g., from the Department of Health and Social Care), Germany (from the Robert Koch Institute), and Switzerland (from the Federal Office of Public Health). In countries with federal systems, national data has been pulled together with a great effort from across the scientific and public health communities, as well as local authorities. In Switzerland, there are 26 separate health systems, each with its own medical director, and central data collection has had to involve collaboration across the cantons and navigate public sensitivities about data use. More than 300 scientists came together to form a task force and provide briefings and analyses of data.

In many countries, e.g., the UK, national-level data is presented to the public by government leaders. Still, webinar participants felt the public are not always aware that data may be far from complete. Data on deaths from COVID-19 are a case in point. The WHO has issued guidelines on certifying death from COVID-19, but not all countries are using the same procedures. Even within countries, data has changed over time, as more data has become available.

The NHS in England provided the earliest daily data, of the number of people who have died in hospital of confirmed or probable COVID-19.  From April 29, the Department of Health and Social Care started to publish UK daily data on deaths of people who tested positive for coronavirus, including outside hospitals. From April 28, the Office for National Statistics began to release weekly (England and Wales) data on deaths “involving” COVID-19 (including suspected cases) and also provides an analysis of excess deaths compared to an average of the previous five years. The UK government stopped publishing international comparative data on deaths on May 12, but there is evidence the changing approaches have confused the public.

Data has also been used in the UK to understand which sections of the population are affected. Analysis of COVID-19 deaths has been broken down by occupation and by ethnic group. Similar reports are not available in Switzerland, and this is a significant gap. Data on testing is similarly complex and variable between countries, not least because tests vary in their sensitivity, and participants agreed, overall, we are a long way from having robust numbers on this pandemic.

Underlying all this is the impact of the virus on the human body, which is still far from understood. Studies based on autopsies are emerging but rare. For survivors, the long term effects are also not fully understood. An interview with virologist Peter Piot illustrates the range and length of symptoms experienced by those who get a severe form of COVID-19 illness, and health systems will need to plan the aftercare for COVID-19 survivors.

While there are variations in how countries collect data about COVID-19 cases and deaths, there are also much larger variations in the availability and use of health service and other data sets to analyse the direct and indirect effects of the pandemic on people’s health. A collaboration between academic teams and data software companies in Oxford has linked an extensive GP data set with records of those who have died of COVID-19 in hospitals in England. Data linkages of this kind are possible but not straightforward in the UK, but harder in fragmented health systems, for example, Germany, where, despite legislation, there are few pooled data sets.

The capacity for local organisations and areas to generate and use data is vital. It allows flexible responses to local situations, even where national data is weak. An example from the UK is the resurgence of people arriving at emergency departments with alcohol, substance misuse, or mental health problems. National data on this will not appear for some time, but local health services can start to respond much sooner if they have the data.  

In considering the risk of any future pandemics, all actors will need to pay more attention to communicating effectively with the public and addressing the trends of misinformation and conspiracy theories that gain traction where there is uncertainty. The WHO’s Director-General called attention to this “infodemic” in mid-February. In response, the WHO has published a series of myth-busters for countries to use.

There was much agreement among participants about the need to learn how leaders at all levels could better communicate data, especially where there are ambiguities or absences of evidence. UK participants noted a tendency of politicians to assert they were “following the science” and not acknowledge the scientific evidence pointed in different directions. While it may be easier for scientists and academics to shift their positions in the light of new evidence, it is perceived to be harder to reverse policy decisions. Nevertheless, there is learning across borders in how national politicians have communicated uncertainty to their citizens without undermining their authority or avoiding their responsibilities.