Israeli scientists have launched a major international appeal for COVID-19 data, to power a new algorithm designed to improve home care for patients by detecting hard-to-spot warning signs in oxygen levels.
“We’ve developed a way to remotely monitor, continuously, COVID-19 patients from their quarantine location,” Joachim Behar of the Technion – Israel Institute of Technology told The Times of Israel.
He added: “This is a really exciting project because such an algorithm may enable us to provide a new way to monitor the health of individuals with COVID-19 remotely from their home quarantine.”
But to get the system up and running, he needs an initial batch of data on coronavirus patients and their oxygen levels over time, which he expects to gather quickly after issuing an international call for statistics through an open source medical platform. Behar hopes to get tens of thousands of case files from around the globe.
There is a strong push internationally to keep COVID-19 patients at home as much as possible, given the strain on hospitals and the infectious nature of the virus. But it can he hard for doctors to monitor patients’ health without seeing them — and there are concerns that they can develop sudden pneumonia-like symptoms, and fail to get to a hospital as soon as they should.
Behar’s innovation, developed with the help of masters student Jeremy Levy, is a system that gathers data on the oxygen levels of home-based patients which the latter can measure with a simple oximeter, and then enter into a website or app. The system will crunch the data using an algorithm that looks for specific biomarkers — measurable indicators — and tell doctors if the oxygen levels are following patterns often seen by those headed to deterioration. He hopes that the ability to pick up signs of deterioration early will make home hospitalizations safer.
Behar acknowledged that some patients already have oximeters, but said that readings are normally only used to verify that their oxygen levels are in a normal range, not to detect the type of complex patterns that his algorithm will find.
“For people who have the coronavirus but who are at home or in a quarantine hotel, it’s a way to monitor them and know if their condition is likely to develop further,” he said.
He said that once the data he needs is submitted through the crowdsourcing site, he will quickly activate the system for use by doctors. “With the right data, this system can be up and running within weeks, helping to give doctors a better picture of patients in home hospitalization,” he said.