Computing COVID-19
What if your phone knew you had been exposed to COVID-19 before you did? According to computer science and engineering professor Shashi Shekhar, co-author of the recent book Spatial Computing, data collected by the cellular devices we keep in our pockets could be a key weapon in the fight against the novel coronavirus.
Shekhar, a Distinguished McKnight University Professor, analyzes and studies large sets of spatial data, which is any data that has a location attached to it. Common examples include the United States census, daily satellite imagery, and location data from GPS-equipped devices such as vehicles and smartphones.
“Smartphone location traces, also called trajectories, are probably some of the most valuable big data,” Shekhar explained. “There have been many people excited about it for the past 10-12 years.”
Shekhar is a pioneer in the field of spatial computing, and he aims to use this emerging big data in ways that can benefit society.
In the past, Shekhar has studied GPS trajectories for various transportation industries—his research group is currently analyzing maritime ship trajectories to create collision risk maps and monitor locations of open-water vessels. He also works under a National Science Foundation (NSF) grant to find ways to reduce emissions in trucks and other vehicles, a project in collaboration with CSE mechanical engineering professor Will Northrop.
Now, Shekhar says that by using location data from smartphones, researchers can not only track and control the spread of COVID-19 but also evaluate how effective government responses to the disease have been.
Going mobile with contact tracing
For years, health workers have been tracking—and eventually containing—the spread of diseases through a method called contact tracing. Essentially, when patients test positive for a disease, health professionals will interview them to find out who they have been in close contact with and where they could have spread the disease.
While hospitals have been conducting this process manually for COVID-19 patients, Shekhar said using mobile location data would speed up the process significantly.
“COVID-19 is so fast-breaking and so large, you can see that it is overwhelming the manual contact tracing system,” he explained.
“That’s why today we need to help the contact tracing professionals with new tools, and one of those tools is smartphone trajectories,” he said.
Shekhar proposes a system in which medical records and spatial data work together. After compiling a dataset of patients who have tested positive for COVID-19, researchers could cross-check that data with smartphones that were close to that person during the infected period. Then, public health organizations could contact those who may have been infected and provide medical advice or recommend that they quarantine themselves for 14 days.
Countries like Israel and Singapore have already made use of this mobile contact tracing, sending texts to citizens alerting them of potential COVID-19 exposure. This would be harder to achieve in the United States, since current privacy norms impede the release of such specific data. However, considering the threat the novel coronavirus poses, Shekhar believes the U.S. government should make an emergency exception and harness this data for good.
The government already uses a similar technique called wireless emergency alerts to warn individuals in areas at risk of hurricanes or flash floods. Plus, to address HIPAA concerns, the U.S. Department of Health and Human Services announced in February 2020 that organizations may disclose protected health information about a patient in order to treat another patient. Recently, Minnesota Governor Tim Walz signed an executive order allowing the Minnesota Department of Health to share locations of COVID-19 cases with first responders while safeguarding patient privacy.
“COVID-19 is probably the biggest national emergency in almost 100 years, and during these times we need to think a little differently,” he said.
“In the case of emergencies, sometimes we need to share data more when it’s in the benefit of everyone,” Shekhar said.
Another option to address contact tracing digitally is through opt-in peer-to-peer apps. Google and Apple recently announced a partnership to develop an opt-in app for users to share positive test results via a public health agency. The University of Minnesota is currently working with HealthPartners to develop something similar—an app called SafeDistance, which would ask its users to voluntarily share whether they’ve tested positive for COVID-19.
#StayhomeMN
Contact tracing just scratches the surface of the insights spatial data can provide. Months before COVID-19 arrived in the U.S., the medical intelligence community detected drastic changes in spatiotemporal patterns of life in Wuhan and raised an alarm about the grave dangers. This analysis was likely based on high-resolution satellite imagery and other spatial big data.
More recently, researchers have begun using smartphone data to study the impact of both statewide and nationwide measures such as social distancing and shelter-in-place orders.
“Now a big question is, ‘How well are these measures working?’” Shekhar said. “Is the public complying or are they getting frustrated and is the compliance going down?”
By using mobile data to compute each person’s “daily range” over time—in other words, how far people are traveling from their homes—researchers can determine how effective shelter-in-place orders have been in reducing movement. Then, this information can be used to inform public policy.
“If you see the range of movement starts to go up, maybe it’s an indicator that people are getting frustrated [with the restrictions],” Shekhar said.
“And at that point, policymakers can start thinking of reaching out to people more, communicating more, or putting some other measures in place,” he added.
Similar studies have been done to identify gathering places and assess the effectiveness of social distancing in metro areas like New York City. Researchers can also use this data to create improved COVID-19 transmission dynamics models, which policymakers can use to estimate the number of infections and deaths.
Shekhar is currently working on obtaining smartphone data and high spatial-resolution satellite imagery from companies so that he can start analyzing it in relation to the local community in Minnesota.
The long run
Shekhar is also thinking ahead to the future. He said location data can be used not only to prepare for future pandemics but also to help us recover from this one.
Once the peak of COVID-19 passes, researchers can use spatial technologies to identify low-risk areas across the country. If adequate COVID-19 testing resources are available, governments could perhaps lift restrictions in these areas first, allowing more people to return to work and start repairing the economy.
Smartphone-based contact tracing could then identify and isolate the infected and shield the most vulnerable to quickly control further spread and reduce risk of subsequent outbreaks.
Spatial data can also prove beneficial in creating risk maps to predict areas where viruses are more likely to infect humans, such as regions with high concentrations of human-animal contact and coronavirus reservoirs like bats. Researchers can then monitor and study these areas to identify other highly contagious coronaviruses that could pose threats to society.
Shekhar said there are numerous other ways science and engineering fields like spatial computing can help, and he encourages his fellow researchers to start thinking about the ways their work can help if they haven’t already.
“I think people, in our college and others, need to take a look and see that many ideas they already have can make things better,” he said. “In this COVID-19 crisis, if science and engineering can help society, we need to do it.”
Story by Olivia Hultgren
The University of Minnesota’s Office of the Vice President for Research (OVPR) is encouraging researchers outside of the UMN Medical School to submit proposals for coronavirus-related research under its COVID-19 Rapid Response Grants fund. NSF has similar programs for researchers with and even small businesses with ideas to help understand or mitigate the effects of COVID-19.
For more information about NSF’s Rapid Response Research and other programs, see the NSF resources for the research community and small businesses.
Read Shekhar's Scientific American essay on using mobile data to stop the spread of COVID-19.