We focus on ten societal areas.
TNO cooperates with companies, the public sector and other organisations, to apply our knowledge and expertise with and for others.
We develop knowledge not for its own sake, but for practical application.
On TNO Insights you can read in-depth interviews and articles. Get inspired by TNO’s innovations!
TNO offers you the chance to do groundbreaking work and help customers and society with innovative, practical and smart solutions.
One of the most challenging aspects of the coronavirus is predicting its next move. From the number of people who will contract the virus to the number of needed ICU beds to the effect of lockdown measures, nations around the world are looking for ways to anticipate the course of the virus and take measures to control it. Maurice Hanegraaf explains how TNO developed a tool that can make accurate coronavirus predictions and support decision-making.
Want to know more about how data-driven prediction models can improve decision-making? Contact Jan Diederik van Wees for more information
When the virus reached the Netherlands, ICU beds filled rapidly. The Dutch National Corona Outbreak Team and various hospitals needed a way to predict the number of new cases they could expect each day, in order to respond effectively. TNO’s expertise in predictive models provided a solution. Using the basic algorithms of a model originally intended to predict earthquakes, a team of geology experts at TNO developed a predictive model for the coronavirus, together with medical experts from University Amsterdam Medical Center, Erasmus University Medical Center, and Leiden University Medical Center.
‘We adjusted the model to be effective at predicting aspects of the outbreak, and used data from China, Korea, Italy and the Netherlands to make the model more robust,’ Maurice says.
Early on, the model aimed to predict the need for ICU beds from day to day. But with more data, more parameters and more input, the model can now predict the outcome of lockdown measures, track new viral outbreaks, and help with decision-making for everything from treatment to reopening strategies.
Some of the features of the virus are consistent, like the onset of symptoms within two weeks of infection and the general course the virus takes in the body. When localised, region-specific data – including the date of the first cases and the lockdown measures taken – are entered into the model, users can extract predictive data to prepare for further outbreaks, new hospital admissions and decisions about reopening.
The tool shows users the drastic effects small changes can have. ‘Only a slight change in the parameters can profoundly affect the predictions of new cases, hospital admissions and deaths. It illustrates how crucial accurate data can be in fighting the virus,’ says Maurice.
The model is open source and available for download. But to get the most out of it, it must be properly calibrated based on country-specific data. ‘TNO can help to set up a predictive tool for any country,’ Maurice says. ‘Then, decision-makers can use the model to make accurate predictions in less than a minute, simply by plugging in the relevant parameters.’ The tool even includes a ‘cruise control’ to show how to keep numbers at a desirable level, and an ‘emergency brake’ that warns policymakers to immediately reinstate lockdown measures.
After hearing about the model, a large NGO in a developing country contacted TNO to see if the model could also be used there. TNO immediately replied that it could. As long as the right partners are around the table.
Initially, Maurice and his team ran some of the country’s data themselves, to offer initial predictions. They used data from three African countries to show proof of concept. But more work is required if the country wants to run the model themselves. ‘Users need to be trained on how to use the decision support tool,’ he explains, ‘and if they need a specific interface, we can work with an ICT partner to adjust the application. Together with medical experts from a specific country, we can help them to ensure effective data interpretation and the most impactful decision-making.’ The tool can be helpful for government policymakers, crisis teams, hospitals, NGOs or for daily prognosis in a public dashboard. Of course, once the organisation can effectively use the model, it can continue to use it to track other diseases and outbreaks as well.
Need accurate predictions to help keep the coronavirus under control? Want to know more about how data-driven prediction models can improve decision-making? Contact us today for more information.
About three times a year we let you know about our newest activities