Study tracks COVID-19 spread via cellphones
Public Health may not know where you are, but your phone does
“The virus only moves when people move. If you do not move, the virus cannot move.”
Newfoundland and Labrador Health Minister Dr. John Haggie said those words almost a year ago, and they’ve become a mantra of health officials ever since.
Now, research out of Ontario has vividly demonstrated how true that is by tracking mobile phone data over the span of the last year and comparing it to growth in COVID-19 cases across Canada.
The research was published in the Canadian Medical Association Journal last week.
A team led by Kevin Brown, an epidemiologist and infection control specialist with Public Health Ontario, took anonymized data sets from Google based on smartphone locational services, which most people have turned on for convenience.
The data is readily available to the public online.
They then developed what they call a mobility threshold, a calculation based on a variety factors that estimates the maximum amount of mobile activity that can be tolerated to contain COVID19 spread.
One of the main factors is that there’s a higher threshold in the summer, when people are outside more, and a lower threshold in the winter.
When actual out-of-home mobile activity data and COVID-19 growth rates are plotted over the same span of time, the numbers show a clear correlation.
EARLY WARNING
Brown says the research complements similar studies done in other jurisdictions.
“This study confirms that, sure enough, if you look at mobility as a proxy, it acts as a pretty good proxy for contact rates in that it’s very strongly associated with underlying COVID-19 case growth,” he said in a phone interview Tuesday.
Brown said there’s a lag time between mobile activity exceeding the threshold and an increase in COVID growth, which he attributes to the time it takes for the coronavirus to incubate and the time it takes to get test results.
“What we know is that mobility data and these contact rates pre-date case growth by around 11 to 12 days,” he said. “Depending on how quickly your mobility data is coming in, it does provide this early warning signal.”
A government science team in Ontario is already looking at the data as it informs Public Health decisions, Brown said.
SLOW TO REACT
The study also found that residents coming out of lockdown are actually slower to react to their new freedoms than they are to comply with new restrictions.
“You don’t see an immediate uptick in cases after you reopen,” said Brown. “It takes about a month or so for the transition related to reopening to start to occur. That complacency takes a long time to ripple through society after you’ve been in a lockdown.”
Unfortunately, the data would be useless for a province like Newfoundland and Labrador, where there isn’t a steady baseline of cases to compare to mobile data.
“There’s this sort of decoupling that occurs in places like a lot of Atlantic Canada,” he said.
But with cases climbing again in larger provinces as new variants take hold, Brown said mobility data can still act as an important bellwether until vaccination rates reach herd immunity levels.
NO PRIVACY ISSUES
Newfoundland and Labrador’s privacy commissioner says he sees no issue with the study’s use of mobility data, although its provision by a private company such as Google is beyond his oversight anyway.
“If information is effectively anonymized or de-identified, then it is no longer personal information,” Michael Harvey told The Telegram. “Often this is more complicated than it may seem, and certain anonymized or de-identified datasets are at risk of being re-identified. However, if we are talking about a population-wide set of data, such as cited in this article, then the risks of de-identification seem, at least on their face, low.
“The disclosure of such datasets for research purposes therefore may be socially beneficial while avoiding privacy risks.”