Physicists Mobilize to Combat Coronavirus

Sponsored by the APS Forum on Physics and Society, the session included speakers discussing three areas where physicists are already making a difference. Reiner Kruecken, deputy director at TRIUMF (Canada’s national particle accelerator laboratory), spoke on behalf of the Mechanical Ventilator Milano (MVM) project, a large collaboration focusing on producing low-cost and reliable ventilators to meet a global shortage. Stephen Streiffer, Deputy Laboratory Director for Science (Interim) at Argonne National Laboratory (ANL) discussed the ways in which DOE labs are lending their expertise and infrastructure to projects such as modeling COVID-19 spread, imaging the virus in search of drug discovery, and more. Savannah Thais, a post-doc at Princeton University, explored big data and machine learning projects aimed at mitigating effects of the pandemic that physicists can get involved in, even from home.

International Collaboration Fights a Ventilator Shortage

The MVM project was born out of the need to ensure ventilator access for those who need them during the COVID-19 pandemic. One way of meeting this need is increased production of high-capacity moderate-cost (that is, cheaper than typical high-end intensive care unit equipment) ventilators, a design specification met by MVM. From conceptualizing the MVM to manufacturing a working production model only took the collaboration four weeks, kicking off March 19 (see Physics article).

“There was a tremendous effort to bring this forward,” said Kruecken at the April Meeting. “We are now in the final stages of getting the system approved by the FDA, and plan on starting manufacturing as soon as we have approval—realistically in early May.” The MVM received FDA approval on May 5.

The MVM was conceived as a low-cost, easy to operate ventilator, based around readily available parts. To achieve this, clinicians and experts in Italy, the United States, and Canada, along with manufacturers, gave their input on the final design, especially keeping needs of COVID-19 patients in mind. While some ventilators operate by pushing a certain volume of air into the lungs, MVM belongs to the pressure control class of ventilators, which is more suitable for patients with lungs damaged by COVID-19.

During the design period, pre-production models were sent to five different locations for testing and improvements: Naples, Valencia, Fermi National Accelerator Laboratory (FNAL), Canadian Nuclear Laboratories, and TRIUMF. The final MVM design, which can be operated in two modes, one for patients who are sedated in an ICU and another for patients who are transitioning off of the ventilator, is estimated to cost just around $5000—substantially less than regular ICU ventilators.

Kruecken attributes the rapid development of the MVM ventilator to the power of collaboration across nine time-zones and the ability to leverage physics expertise in gas and electronics systems used in particle physics projects.

“People want to help, and people bring whatever expertise they have to the table,” said Kruecken. “The end product ... demonstrates how that expertise, taken together with a large international effort, can ... accelerate development in a marvelous way.”

National Lab Infrastructure Supports Coronavirus Research

The Department of Energy (DOE) national labs are also applying their expertise and infrastructure in a number of areas, from structural biology research to epidemiological modeling. Each of the 17 national labs, which focus on areas from basic science to national security, have members participating in a new initiative: the National Virtual Biotechnology Laboratory. This consortium of the national labs is using DOE user facilities to tackle a variety of challenges in responding to COVID-19.

ANL, Oak Ridge (ORNL), Los Alamos, and Sandia National Labs are all collaborating on modeling COVID-19 spread in an effort to address key questions about the pandemic. By utilizing supercomputing capabilities, these models can anticipate how changes in behavior, like adhering to stay-at-home orders, modify pandemic impacts. ORNL is also one of several labs involved in addressing supply chain issues by applying advanced manufacturing techniques to produce critical medical supplies.

DOE user facilities, like the Advanced Photon Source at Argonne, have been employed to study the structure of the SARS-CoV2. According to Streiffer, using the Advanced Photon Source for this kind of research is not new, as about a third of the facilities’ users were already conducting life-science research.

The National Virtual Biotechnology Laboratory is also tackling computational drug discovery by mobilizing artificial intelligence capabilities to find molecules that can interfere with one or more of the coronavirus’ 28 proteins. Using data-mining with machine learning, about 30 potential therapeutic molecules have been identified for further testing.

“The work that we are doing…is only in many cases a minor pivot away from work the DOE was already doing, and we’ve been able to apply that really quickly to the COVID-19 pandemic without really redirecting work that people would’ve been doing otherwise,” said Streiffer. “The DOE plans to continue this work—it’s one of the reasons they established the National Virtual Biotechnology Laboratory to have a framework that could be sustained after the current crisis.”

Open Science Recruits Physics Expertise

To conclude the session, Thais provided an overview of open science initiatives that will allow physicists to apply their skills during the COVID-19 pandemic. As most of these projects require familiarity with large data sets, advanced analysis models, and distributed computing, many physicists already have the skills to provide support. These open science projects range from epidemiologic modeling to non-invasive diagnostics to projects aimed at fighting the spread of misinformation about the virus.

Thais mentioned several modeling projects to track the spread of COVID, like HealthMap and and CDC Flu Models that help predict hospital burdens. Other expanded modeling projects like COVID-Care integrate hospital resources into disease models to better predict risks in a specific area. Non-invasive diagnostics projects are applying technology to lessen testing burdens, such as using audio data to distinguish COVID-like coughs with a phone app.

Other projects Thais highlighted include: leveraging AI to improve telehealth to at-risk populations, supporting hospital systems by giving them a model to predict how supplies will be utilized (CHIME), and “info-demiology,” tracking the spread of unreliable public health information online.

Thais urged attendees to consider ways where their skills might plug in to projects—even if not in a “super technical” way. She also encouraged “thinking locally and globally” to identify what needs exist both in one’s community and in dealing with the pandemic at-large.

For more about the Virtual April Meeting visit april.aps.org.