News Coronavirus Researchers develop algorithm to predict coronavirus outbreaks
Updated:

Researchers develop algorithm to predict coronavirus outbreaks

Share
Twitter Facebook Reddit Pinterest Email

Scientists have developed a mathematical algorithm to predict when the next coronavirus outbreak will occur.

Concerned about new epidemic waves as the US began prematurely easing restrictions, a team based at Harvard University sought to develop a COVID-19 early warning system.

Led by machine intelligence experts Mauricio Santillana and Nicole Kogan, the team developed an algorithm that could raise the red flags two to three weeks before an outbreak is expected to take place.

Being able to monitor in real-time the mobility data from smartphones as well as what people are searching on Twitter and Google, for example, has enabled them to make such a prediction.

“What we’ve looked at is what we think are the best available data streams,” Dr Santillana said.

Their algorithm can also be used to predict when there will likely be an exponential growth in COVID-19 deaths.

The study, which was published on Friday (Australian time), likens the new algorithm to a thermostat that can guide health authorities to know when to activate or relax public health interventions.

“What we’re doing here is observing, without making assumptions. The difference is that our methods are responsive to immediate changes in behavior and we can incorporate those,” Dr Santillana, director of the Machine Intelligence Lab at Boston Children’s Hospital, told the New York Times.

Instead of relying so heavily on pharmaceutical remedies, researchers have also shown that simple “next-gen data sources” can be used to “provide early signals of rising Covid-19 prevalence” to stop the virus in its tracks, said Lauren Ancel Meyers from the University of Texas.

“Particularly if confirmed case counts are lagged by delays in seeking treatment and obtaining test results,” she added.

Their paper concluded that “these efforts represent an initial exploratory framework”, and while promising, require “continued study of the predictive power of digital indicators as well as further development of the statistical approach”.

It comes as the number of confirmed coronavirus cases per day in the US has climbed to a record high of more than 50,000, with the infection curve rising in 40 of the 50 states.

An alarming 36 states are witnessing increases in the percentage of tests coming back positive for the virus.

“What we’ve seen is a very disturbing week,” Dr Anthony Fauci, the government’s top infectious-disease expert, said in a live-stream with the American Medical Association.

The surge has been blamed in part on Americans not covering their faces or following other social distancing rules as states lifted their lockdowns in the past few weeks. Fauci warned if people did not start complying, “we’re going to be in some serious difficulty”.

The US recorded 50,700 new confirmed cases, according to Johns Hopkins University. That represents a doubling of the daily total in the past month and is higher even than what the country witnessed during the most lethal phase of the crisis in April and May.

All but 10 states are showing an upswing in newly reported cases in the past 14 days, according to COVID Tracking Project data.

The outbreaks are most severe in Arizona, Texas and Florida, which together with California have closed again or clamped back down on bars, restaurants and cinemas.