PALO ALTO, Calif., Oct. 29, 2020 /PRNewswire/ -- In a recent preprint, scientists at Stanford, Oxford, and the Broad Institute leveraged 3.6 million anonymous health data submissions from Enya.ai's FeverIQ to demonstrate how secure multiparty computation (SMC) can optimize for Covid-19 symptom screening without compromising user privacy.
The information needed to manage Covid-19 is some of the most sensitive imaginable — medical symptoms, test results, where you work and sleep, and your closest contacts. New research based on FeverIQ, however, shows that privacy and Covid screening can be reconciled. FeverIQ is a Covid symptom tracking and screening service created using the Enya.ai secure computation platform. With the help of FeverIQ, the study collected 3.6 million submissions of Covid symptoms and test results and analyzed them using secure multiparty computation (SMC). The FeverIQ risk model for predicting Covid health risks outperformed others by up to 2.7 times — and, as the consortium's study states, "to protect the participants' privacy, no identifiable information was requested, collected, transmitted, or stored, and geolocation data were downsampled prior to leaving the user's device."
Privacy and health screening accuracy are two sides of the same coin. Unfortunately, they have often stood in conflict with one another. While personal health information is needed to assess an individual's Covid risks accurately, health data are often stolen, and concerns for misappropriation of private user data within the business world are increasing. The consortium's study elaborates on this issue in relation to the status quo of Covid risk assessment: "The role of privacy has not been broadly addressed in tracking projects, which collect participants' symptoms, test results, and metadata. Especially for longitudinal studies, in which symptoms and geospatial data accumulate for each study participant, it can become possible to re-identify participants even when identifiers such as names and email addresses are scrubbed from the data."
In addition, nearly every week, our understanding of the virus — including its symptoms, effects, and treatments — is shifting as the world learns more about its characteristics and their implications. For example, we have learned that temperature checks may be an inadequate litmus for coronavirus, and ophthalmological research suggests something as seemingly unrelated as wearing eyeglasses is linked to significantly reduced risk of infection. Furthermore, the virus is mutating, though more slowly than seasonal influenza. Therefore, an effective Covid screening model needs to adapt by continually incorporating the latest findings in the field and from the medical research community. As the consortium's study states, the FeverIQ risk model is "able to validate the diagnostic power of newly reported symptoms without needing to receive unprotected granular health information from participants." Such a framework is precisely what is needed to build adaptive solutions as the coronavirus and our understanding of it evolves. The research was funded in part by the Bill and Melinda Gates Foundation and involved researchers from Stanford, Oxford, and the Broad Institute of MIT and Harvard.
Enya.ai is the only secure computation platform optimized for edge devices such as mobile phones, helping organizations derive differentiated insights without the risks of exposing sensitive data. Trusted by millions of users, Enya.ai operates the largest secure computation network for healthcare. Enya.ai is also among the leading Stanford alumni scientists and physicians participating in the StartX Med COVID-19 Task Force, mobilized at the onset of the pandemic to provide critical solutions for the prevention, diagnostics, and treatment of Covid-19.
For more information, visit https://enya.ai.
Get started with FeverIQ: www.feveriq.com
FeverIQ by Enya.ai
[email protected] | (650) 418-3042