Source – https://www.healthcareglobal.com/
DataRobot’s AI platform makes it easy to build predictive models to identify infection hotspots
A new data-driven project is launching with the ambitious aim of resolving the COVID-19 pandemic in 60 days.
ContagionNET has been designed by DataRobot, a global enterprise AI company that enables easy building and deployment of predictive models. Aimed at solving the data sharing problem that has plagued the COVID-19 pandemic response, the not-for-profit initiative will help determine when a patient is at their most contagious.
ContagionNET combines affordable antigen checks performed at home, anonymous data collection, and an AI platform. With these tools DataRobot are aiming to identify the most contagious areas in a community and encourage people to take preventative action. The goal is to identify people earlier than traditional testing (i.e., when they are most infectious) to disrupt the chain of transmission.
ContagionNET can also predict how many people need to participate on a county-by-county basis by measuring the impact of different participation levels and testing frequency. This localised approach allows those with the highest viral load to self-isolate earlier, preventing further spread of the virus.
“The reality is that those most likely to spread COVID-19 aren’t usually aware they are contagious” said Sally Embrey, DataRobot’s VP of Public Health and Health Technologies.
“With the proliferation of frequent, easy-to-use at home antigen checks, ContagionNET is poised to inform the most contagious of their risk and help them make lifestyle modifications that will significantly reduce spread of the virus.”
DataRobot has been involved in pandemic response since early 2020. Working with the US government, DataRobot’s forecasting helped vaccine manufacturers select the right participants and prioritise enrollment in the highest risk locations over each vaccine trial period.
The organisation has also used modeling to improve testing distribution, equity, and diagnostic reporting, all of which have informed its approach to ContagionNET.