As the coronavirus began to take hold in the United States, there were fears of a ventilator shortage. In an effort to prevent such a catastrophe, the Defense Production Act was used by the U.S. Government to order the development of 200,000 ventilators from 11 different manufacturers with a cost totaling more than $3 billion for the Strategic National Stockpile (SNS).
According to data from the CDC as recent as February 9, there have been 26.9 million coronavirus cases and more than 462,000 deaths in the United States. The country’s seven-day average is more than 3,100 deaths per day, with the leading for COVID-19 patients being acute respiratory distress syndrome, or ARDS.
In the early stages of the condition, the smallest blood vessels in the lungs begin leaking fluid into the tiny air sacs where oxygen exchanges take place, called alveoli. This causes the lungs to shrink and become stiff, leading to difficulty breathing and hypoxemia. In the worst cases, oxygen starvation affects the brain and other tissues, leading to organ failure.
A study published in the Chest Journal last September mirrors my personal experiences as a critical care doctor working on the frontlines during this global pandemic. According to the study, patients “often” develop ARDS and severe hypoxemic acute respiratory failure. In the study, the functionalities of the SNS devices were compared, and the result? “Current SNS ventilators and those on order are capable of supporting most but not all COVID-19 patients,” the authors stated.
When the SNS ventilators began arriving, it became painfully obvious to medical providers across the country that many of the machines fall short when it comes to treating patients with ARDS. This is because they were built with quick deployment and mobility in mind, so they lack advanced features prefered by critical care doctors ,such as pressure-regulated volume control (PRVC) and bi-level ventilation, to treat ARDS.
There is no denying that more basic machines have a life-saving role in this pandemic—any ventilator is better than bagging a patient—but while every life saved is valued, no physician should be happy with settling for saving “most” lives if the option to save more exists. We must ensure that it does when the next global pandemic hits.
Beyond my desire to save as many lives as possible, I have a passion for technology. My parents were in telecommunications and I, too, have owned a technology company. I also had the opportunity to work for Robert Bartlett, M.D., a pioneer in the field of critical care. He developed the technology known as extracorporeal membrane oxygenation (ECMO). I served as a research assistant in his ECMO lab at the University of Michigan, and that opened my eyes to the things technology and hardware can achieve when combined with patient care.
Critical care hasn’t had many breakthroughs in recent years and in order to change that, I believe we need to put more of an emphasis on integrating technology—specifically artificial intelligence (AI) —into the field.
When other teams including those from MIT, NVIDIA and the University of Vermont began working on their ventilators, I too set out to develop a solution. By July, my concept, called the VMS-I, was finished but I set it aside. With the arrival of SNS ventilators, I realized it was time to take it back off the shelf.
The VMS-i includes a machine with capabilities that SNS ventilators lack, such as PRVC. However, it is the ‘MS-i’ of VMS-i that truly sets it apart from current ventilator management systems, because of its remote monitoring and AI-capabilities and, by extension, the way it could dramatically improve patient care.
The screens on SNS ventilators usually cannot be clearly seen outside the patient room. That means that when numbers need to be checked or adjustments need to be made, a member of my team or I have to put on full PPE and enter the room. There are several problems with this. First, it takes time. When providers are working upwards of 90 hours a week with record-breaking caseloads, every minute must count for something.
Bedside care also increases the chance of healthcare professionals being exposed to the virus. An analysis by The Guardian and Kaiser Health News published in December revealed that more than 2,900 healthcare workers died of COVID-19 last year, while they worked tirelessly to save the lives of the virus’s other victims. There was already a shortage of healthcare workers prior to coronavirus taking hold globally and we must value those who have entered the field enough to provide them with the proper tools to lower the risk of exposure. We must give them the resources they need to save patients —and themselves.
The VMS-i could be that resource for those working with ventilators. It’s AI-capable mobile app integration will allow providers like myself to check a patient’s stats and make adjustments to treatment from anywhere, whether it be a command center or a doctor’s cell phone. This will not only decrease exposure risks by reducing time spent next to a contagious patient, but will also save us the precious minutes that we need to give our patients the best possible outcomes.
In its current state, the app offers the ability to manage patients remotely but the ai-capability means there is room to add a cognitive system that is able to take data from multiple sources, find and flag significant events and offer more insight than the APACHE or SOFA mortality scores that machine learning currently offers critical care providers.
Right now, my system requires changes to be made by a physician, but I believe that in coming years as AI technology advances, the VMS-i could be developed to make critical adjustments for a patient immediately without the need for a physician. We have self-driving cars that are capable of making split-second life-saving decisions to ensure passenger safety on the road. In some cases, the process of the vehicle detecting the hazard and taking action to avoid it plays out before the individual in the car notices they are in danger. Imagine the lives that could be saved if an issue could be detected and addressed immediately, without the time it takes for a physician to be notified of the situation.
I designed the VMS-i to work with my machine concept, but also to be easily integrated into existing machines from other manufacturers, once standards have been agreed upon. This could allow providers access to the remote ventilator management system without the need to purchase a brand new ventilator—something that many facilities can’t do at the drop of a hat.
By embracing next-gen concepts like the VMS-i and continuing the conversation about AI’s place in medicine, I am confident we will save more lives. The COVID-19 pandemic has been an incredible tragedy. I believe we must view this as a wakeup call and do more to integrate widely available technology into patient care if we want to mitigate the impact of future pandemics. We cannot be caught off guard again. Our providers, and most importantly, our patients, deserve more than devices that were built quickly and are only enough for “most”.