The digital medicine company Physiq Inc. has received a contract from the NIH to develop an artificial intelligence (AI)-based index that can provide an early warning that a patient with COVID-19 is in decline and needs medical treatment.

The index, called the COVID-19 Decompensation Index (CDI) Digital Biomarker, is a personalized and multivariate AI-based algorithm that will run on Physiqs existing Accelerateiq computing platform, analyzing physiological data from FDA-cleared wearable devices to create a personalized baseline for patients who have tested positive for COVID-19. The technology then indicates changes in vital signs and results of the personalized baseline, producing the index that is used by medical personnel.

Chicago-based Physiq was one of seven companies and academic institutions to be awarded an NIH contract to develop digital health solutions to help address the COVID-19 pandemic.

The contracts were awarded in two phases, with an initial phase to demonstrate feasibility and a second phase for development. If all seven projects advance to the second phase, the total value of the contracts would be $22.8 million. Physiqs portion of the contract is worth $6.6 million. All companies have one year to complete the entire project.

The thing thats unique about our approach is we learn the patients unique physiology. This is not a population-based test, Gary Conkright, Physiqs CEO, told BioWorld. Instead, an N equals one approach to learning physiology allows us to see very subtle changes over time.

Building on existing index

Conkright said he anticipates that the company will be successful in developing the CDI because it builds on the companys existing FDA-cleared algorithm, the multivariate change index (MCI), which helps track health changes in patients with chronic disease. That technology has accumulated 1.5 million hours of physiologic data that is expected to assist in the development of the COVID-19 specific system.

This is going to work, were just trying to make it more specific and theres value in doing that, he said.

To develop and validate the CDI, Physiq has partnered with the University of Illinois Hospital and Health Sciences System (UI Health) to monitor COVID-19 positive patients who are recovering at home.

During the development phase, COVID-19 positive patients who are sent home will receive a Physiq monitoring kit, which includes a smartphone with the Physiq app installed and wearable devices. The physiological information collected will be sent via cloud technology to advanced practice nurses from UI Health who can identify if a patients condition is deteriorating. Those nurses will contact patients who appear to be in decline and advise them about whether to go to the hospital for treatment.

The idea, Conkright said, is to get patients treatment before their condition becomes critical.

One of the horrible things about COVID is that it so subtly decreases physiological function that sometimes patients dont detect its getting worse until they are really in trouble, he said. By that time, it might be too late to prevent an intubation or worse.

During the validation phase for the technology, the company will evaluate lead time to event statistics, decompensation severity assessments and the ability of the CDI technology to predict decompensation severity.

During the development phase, the company plans to collect 400 patient datasets. They will collect an additional 1,200 patient datasets during the validation phase, Conkright said.

The goal there is to show how early and what kind of predictive power our index has, Conkright said.

The study, which is pending approved by the UI Health Institutional Review Board, is being designed to capture data from a large and diverse population to assess CDIs performance differences in subgroups, based on gender and racial/ethnic characteristics. Conkright said one of the reasons that UI Health was such an attractive partner for this project is that their patient population is racially diverse, with a large number of Black and Latinx patients.

Physiq is not new to the COVID-19 research space. The company has a separate study about to launch with UI Health to examine whether antihypertensive medications have a worsening effect on COVID-19 symptoms, specifically in Black and Latinx patients. Additionally, the company is working with the Department of Defense to collect data on military personnel with COVID-19.

Other COVID-19 digital tools

Along with Physiq, six other organizations received contracts from the NIH to test their COVID-19 digital health solutions.

Evidation Health Inc., based in San Mateo, Calif., is investigating a health measurement platform to analyze self-reported and wearable device data to both detect COVID-19 and differentiate it from influenza.

IBM, in Armonk, New York, is testing an integrated solution to support contact tracing and verifiable health status reporting.

Icrypt Inc., based in Santa Clara, Calif., is developing a smartphone-based platform to provide proof of testing, serologic, and vaccination status.

Shee Atik Enterprises LLC, based in Sitka, Alaska, is developing a smartphone-based platform aimed at supporting individuals in low-resource settings with a Bluetooth-enabled thermometer and pulse oximeter.

The University of California, San Francisco is investigating a GPS-based retroactive contact-tracing tool for COVID-19.

Vibrent Health, based in Fairfax, Va., is working on mobile applications, data integrations, and validated machine learning algorithms to identify COVID-19 and differentiate it from influenza, and to perform contact tracing.

Read this article:
Physiq begins development of AI-based COVID-19 early warning system - BioWorld Online