Researchers work to develop real-time COVID-19 contact tracing app

Two UTHealth School of Biomedical Informatics faculty members are currently developing an app to aid in the fight against coronavirus disease 2019 (COVID-19). Associate Professor Xiaoqian Jiang, PhD and Associate Professor and Associate Dean for Student, Faculty, and Community Affairs Amy Franklin, PhD were awarded a COVID-19 research grant from the National Science Foundation (NSF). The grant funds are being used to build a mobile application (app) called REACT. The app will provide real-time contact tracing and risk monitoring via privacy-enhanced tracking of user locations and symptoms.

With the continued spread of COVID-19 across the nation, prevention and control are of the utmost importance. To aid in that process, contact tracing is an essential tool used for controlling pandemic disease outbreaks. It involves identification and follow-up of all individuals who may have come into contact with an infected person. By utilizing GPS-enabled devices, contact tracing becomes a more reliable resource. GPS-enabled Devices enable use of real-time location traces of users. When that is accompanied by self-reported symptoms and case status confirmation, contact tracing is a very promising social distancing tool. Furthermore, it provides a large-scale, data driven opportunity for contact investigation.

Despite the many advantages of contact tracing, one critical issue in collecting real-time location information is privacy. Recent news indicates that there is little usage of the social distancing app developed by Google and Apple due to user privacy concerns. With the urgency of the COVID-19 pandemic, Jiang knows this type of tool requires a careful balance of privacy protection along with public health benefit.

“We always believed that health benefits should not suppress personal privacy, even during a pandemic,” states Jiang. “Many social distancing apps provide utility by sacrificing personal privacy. Our research team knows that technology can serve as an enabler to reduce the infection risks in a privacy preserving manner so we decided to develop this unique patient-centric privacy-preserving contract tracing app to combat COVID-19.” Drs. Jiang and Franklin are working with other researchers from the University of Southern California (USC) and Emory University, located in Atlanta. The group envisions that REACT will include contact tracing for both individuals and locations. In a circumstance where someone has a confirmed case and was in contact with the user, she or he is informed – allowing for quarantine and decontamination as needed. REACT will also include individual risk monitoring in real-time based on the locations each user has visited and aggregated risks of other users they have come in contact with so they can be informed or alerted. Lastly, the app will include community risk monitoring and detection of early signals of community spread to prepare for larger scale infections.

While Institutional Review Board approval is ongoing, the development work is underway. “Our colleagues at USC have already started the development of the mobile app, which considers privacy protection by design,” Franklin noted. “Our group of researchers from USC, Emory, and UTHealth are comparing different architectures and alternative solutions to deliver the best performance and usability.”

Collaborating with other universities allows Franklin and Jiang to take advantage of each researcher’s unique contributions. While there are overlapping skill sets, the group also has harmonizing strengths. “The USC team is skilled in mobile data anonymization, the Emory group is renowned for their work in de-identifying spatial-temporal data, and we have strong capabilities in privacy preserving time-to-event models,” says Jiang. “We will work collectively to develop the front end tool (app interface) and everyone while take a lead on algorithmic development based on their respective expertise for the backend. This means the backend will be seamlessly integrated into the final contact tracing software.”

One unique, proposed feature of the app is the ability to choose the frequency of tracking or the option of manual check-in along with granularity of uploaded locations (e.g., generalized locations) as the user’s risk evolves. The researchers will also employ a two-stage approach to get more precise results. To accomplish this, the app will allow users to upload their original trace in the recent time window when they are notified of a “possible contact” with a confirmed case.