Online search data can help predict COVID-19 peaks and inform public health responses, according to experts from University College London (UCL).
The report from UCL says web search activity data may help form predictions of peaks and surges in COVID-19 cases as far as 17 days in advance, on average.
This approach of analysing internet searches and web behaviour is already in use for monitoring other infectious diseases and viruses, such as seasonal flu.
According to UCL News, researchers used the COVID-19 symptom profile and search terms identified by Public Health England (PHE) and the NHS, to develop ‘models of prevalence’ based on symptom-related Google search data.
They also took into account, and recalibrated for, the news media coverage impact on searches. Academics could then provide a more accurate model to official bodies to help with case surge predictions.
Dr Vasileios Lampos from UCL Computer Science, and lead author of the paper, told UCL News: “Adding to previous research that has showcased the utility of online search activity in modelling infectious diseases such as influenza, this study provides a new set of tools that can be used to track COVID-19.
“Our analysis was also among the first to find an association between COVID-19 incidence and searches about the symptoms of loss of sense of smell and skin rash. We are delighted that public health organisations such as PHE have also recognised the utility of these novel and non-traditional approaches to epidemiology.”
Findings and modelling from the study are currently being shared with PHE on a weekly basis, as part of the disease response. And the UCL COVID-19 tracker models for the UK and England can be viewed publicly online.
While the research, which also encompassed several countries in addition to the UK, was supported by organisations including the Engineering and Physical Sciences Research Council, Medical Research Council/ National Institute for Health Research and Google Health.
To find out more, read the full research paper, which is published in Nature Digital Medicine. You can also view the latest data from the calibrated model in PHE’s weekly recent surveillance reports.
