Since the University of Arizona launched Covid Watch Arizona – an anonymous COVID-19 exposure notification app – in August, the app has had relatively high download and usage rates, and university analysts estimate that it may have helped reduce the average number of people infected on campus by 12%.

Image

"That estimated 12% reduction in transmission matters. If you can get R(t) (the average number of people infected by a positive

person) to drop from 1.06 to .94, that translates from exponential growth to exponential decline," said Joanna Masel, a professor of in the UArizona Department of Ecology and Evolutionary Biology and a member of the university's BIO5 Institute. Masel led the development of the app's risk scoring component, as a member of the UArizona Covid Watch implementation team and a consultant for WeHealth, the company that distributes the app.  

The app is designed to anonymously alert users of exposure to COVID-19 and direct them to appropriate resources. Anyone can download the app, but its follow-up recommendations are tailored based on location. UArizona students, faculty and staff can select the University of Arizona as their region setting. The app is also in use at Northern Arizona University, where users can select NAU as the region.

Once installed, the app generates random, anonymous codes that are exchanged between app users' phones via Bluetooth. Users who test positive for COVID-19 can input a verification code from a lab, doctor or medical center; at UArizona, that code comes from Campus Health. The app then sends an exposure notification alert to other Covid Watch users whose phones were registered as recently being near the infected person's phone for a significant amount of time.

An estimated 14,000 people on the UArizona campus have downloaded the mobile app.

The app's technology was born from a research collaboration between Stanford University and University of Waterloo researchers, including Covid Watch Executive Director Tina White, who graduated from UArizona in 2007 and 2009 with bachelor's and master's degrees in aerospace engineering before pursuing a Ph.D. at Stanford.

New updates to the app, rolled out on Dec. 16, allow for more nuanced messaging about COVID-19 exposure. Previously, when app users received an exposure notification, it indicated either "no significant exposures" or "significant exposures." Depending on the length of exposure and other factors, users will now instead receive one of three exposure notifications: No exposure detected; low exposure, in which case users do not need to quarantine; or high exposure, in which case users are directed to self-quarantine and are provided information about testing.

Finding R(t)

The university's Covid Watch implementation team was able to estimate the impact of the app on campus based on responses to questions asked about app usage during manual contact tracing and the number of times positive diagnosis verification codes were issued by Campus Health in the UArizona region.

Nearly half (47%) of the infected individuals interviewed by UArizona contact tracers reported that they had downloaded the app. About a quarter (26%) of those individuals reported having entered a verification code into the app – a number the Covid Watch team verified by dividing the number of codes requested by COVID-positive people by the total number of positive tests in the campus community.

If 26% of COVID-positive app users entered a verification code into the app and 47% of the people they infected had the app and received the exposure alert, that would translate to about 12% of infected people receiving an exposure notification from the person who infected them. If they all quarantined in time to eliminate further spread, R(t) could have fallen by 12%, Masel said.

Image

"A 12% reduction in that transmission doesn't just mean that you have a reduction in that next wave of cases but every subsequent wave of cases," said Kacey Ernst, a professor in the Mel and Enid Zuckerman College of Public Health and member of the UArizona team that implemented Covid Watch. "If one in eight can benefit from this, that's pretty good. It's easy to use, and the benefit keeps increasing as more people use it."

It's possible that the reduction in R(t) is lower than the estimated 12%, Ernst said. The best-case scenario assumes, for example, that everyone had their phones with them when they were exposed, that the notifications were received in time and that people actually quarantined after receiving an exposure notification.

However, the reduction also could be higher than 12% if people changed their behavior following notifications of exposures that did not lead to infection, Masel said.

Bruce Helming, assistant director of medical services for Campus Health, says he has seen firsthand that students are responding positively to the app.

"Campus Health providers are saying that students like it. This app does the job that nobody wants to do, which is notify your friends and classmates," Helming said. "And, from their perspective, it's so easy. If you test positive, you enter a code anonymously. They're comfortable doing that."

App Works Best in Concert With Contact Tracing

The Covid Watch app was just one tool in a robust toolkit that allowed the university to offer in-person instruction for courses of up to 50 people during the fall semester. 

The university's Test, Trace, Treat strategy also included providing COVID-19 testing for students and employees, providing isolation housing for COVID-positive students, and engaging in traditional contact tracing efforts. Additional strategies included testing campus wastewater for the presence of the virus and rolling out a blood-draw antibody test, developed at UArizona, that can detect whether a person was previously infected. 

Image

Kristen Pogreba-Brown, an assistant professor in the Mel and Enid Zuckerman College of Public Health and BIO5 Institute member, led contact tracing efforts on campus, which she said complemented the Covid Watch app in an important way.

"Covid Watch works best for anonymous or casual contacts you might have with people. On the other hand, manual contact tracing traditionally gets at the more high-intensity contacts you have, which are people most likely to give you that infection or for you to give the infection to," Pogreba-Brown said. "It's important to have both contact tracing and the app so we can make sure everyone who is exposed is contacted."

Pogreba-Brown leads the Student Aid for Field Epidemiology Response, or SAFER, team, which provides training for UArizona students doing manual contact tracing for Pima County, Maricopa County, Pinal County and Yuma County.

What began as a small program to discover and curb the spread of illness – most commonly foodborne illness – has grown to a nearly 200-person effort to help mitigate the spread of the COVID-19 and misinformation about the virus. The operation will continue to expand in the spring semester, said Erika Austhof, an epidemiologist and lead coordinator for SAFER.

"At the beginning of the year, we had two paid employees – Kristen (Pogreba-Brown) and myself – and five graduate students working for our team on various projects, mostly unrelated to SAFER," Austhof said. "Before COVID, SAFER completed between 500 and 700 interviews for foodborne illness every year."

Today, SAFER includes eight full-time staff, 13 graduate research assistants, 20 graduate student workers, six undergraduate workers and almost 200 volunteers, including 80 psychology undergraduates. The team has conducted interviews for about 5,000 COVID-positive cases since April.

When positive COVID cases are referred to the SAFER team, team members call the COVID-positive individuals and ask them about symptoms and recent contacts. They then call and email those contacts and asks them to complete a secure survey about symptoms, where they've been and who they may have exposed. The team continues to trace the chain of exposures and follow up with people who tested positive up to three times to see if they have any questions.

While contact tracing is a vital tool, cooperation from the public is still lower than health officials would like. Some challenges contact tracers face include distrust, survey fatigue and people simply forgetting who they've been around.

"Some people also just don't know who they were in contact with," Pogreba-Brown said. "If you think back to a party from your undergraduate days, can you tell me 25 other people at the party? Or who were the 10 people standing closest to you and give me their names and phone numbers and email addresses? Probably not. That's the advantage of Covid Watch."

Some people worry the contact tracing process might breach privacy, but Pogreba-Brown stressed that the entire process is anonymous.

Students Step Up as Contact Tracers

As COVID-19 case numbers swelled in the community, SAFER was able to quickly rise to the challenge because students were willing to step up, Pogreba-Brown said.

In addition to public health faculty and students, about 80 psychology undergraduates joined SAFER's efforts, providing nearly 4,000 hours of calls.

"I realized that we have a lot of undergraduates who do research projects in the department as a part of their degree requirements," said Mary-Frances O'Connor, an associate professor in the Department of Psychology. Since many labs were no longer available during the pandemic, she proposed psychology students help with contact tracing for course credit.

Case investigation and contact tracing offer valuable real-world data collection experience for students, O'Connor said.

"Psychology majors are uniquely good at making these calls because of their background in understanding human communication and psychological distress," she said. "The students are calling community members and college students who are ill. Some are scared, and some are not having any symptoms and don't quite believe test results. So, it's a huge range of experience they're getting. They're learning how to get reliable data from people and also how to be a compassionate human being."

The students meet with faculty weekly to discuss their experience and research on psychosocial responses to the pandemic. Students also provide valuable feedback on how SAFER can improve its messaging and calls, O'Connor said.

Jacob Ybarra was among the psychology students to volunteer.

"I was attracted to the project because of the psychosocial impacts of the disease," he said, "but I also just wanted to make a positive impact on our community."

And that impact has the potential to be quite significant.

"By cutting off COVID transmission, you're cutting of a whole social network chain and all those cases down the road," Pogreba-Brown said. "The thought is that some of those cases could end up in the hospital, and some could die. You're not just keeping the next few people (you might have infected) safe, but way more."