— Dr Erich Schulz, MBBS, FANZCA and Robert L. Read, PhD
For about the last six weeks, engineers have been scrambling to design ventilators that can be produced quickly in huge numbers to address an anticipated shortfall. At the same time, major manufactures have ramped up production and formed partnerships with automobile companies and others to increase production. Because the scope of the pandemic depends on contentious social policies, the future is uncertain in wealthy nations and doubly uncertain in the less wealthy nations.
From the outset there has been uncertainty regarding the best treatment protocols and overall case mortality.
The one thing that has been clear from early on in the pandemic is that ventilators were going to be in short supply.
Amidst all this confusion, hundreds of engineering teams, from 40-person teams staffed by biomedical engineers to 1-person projects with little experience, have been trying to self-organize and do the best they can with the resources and talents available to them.
An important guideline has been the UK’s Rapidly Manufactured Ventilation System (RMVS). This has laid out one nation’s evolving vision of what represents a good compromise in a pandemic ventilator designed and rapidly manufactured to address the hopefully short-lived COVID-19 crisis. On April 10th, it was updated to version 4, and this quote is important:
Update on recent clinical experience 10th April 2020
Clinical experience in the UK of COVID-19 has developed rapidly in a way that impacts on the requirements of the ventilators needed. [First …] Second, the duration of intubation is longer than ‘normal’ and so the relative number of ventilators needed in different categories is changing. While a mix of transport, simple mandatory ventilation and complex full featured ventilators is still needed, a greater proportion of these need to be capable of supported spontaneous breathing modes to provide resources for the latter portion of intubation episodes.
This is a conclusion which one of us (ES) had already published in living document “A brief for engineers, by a doctor, on hacking a ventilator for surge capacity in Covid19 patients.” This four-volume work remains open to feedback and comments.
A dumb bag-squeezer simply won’t do the job we need doing in this pandemic. Others, in addition to the UK experts, agree. Managing COVID19 patients with acute respiratory distress syndrome (ARDS) and allowing them to “wean/liberate” themselves off the ventilator requires a more sophisticated approach.
Essentially, we need the electronic controls to synchronise with the patient (see section on importance of smart ventilation). If our current crisis was something like polio (where the lungs were healthy but the muscles were not), or a mass neurotoxin poisoning, then it would be different. If ICUs have to heavily sedate (or even paralyse) a patient to keep them ventilated then we’ll drastically increase the number of ventilated days we need to wean/liberate patients form the ventilator as they get better. Struggling health-care systems will rapidly run out of staff. We are better off with a smart ventilator that gets people in and out of ICU as fast as we can.
Another way of thinking about it is from the point of view of logistics. If we, the community of engineers, especially those of us working on open-source projects, pivot slightly to support at least one “spontaneous breathing” mode of ventilation, the pandemic ventilators we hope to produce are far more likely to be deployed and actually mitigate the pandemic.
This is simple: some patients will eventually require intubation (and likely have very high mortality.) Far more patients will benefit from supportive care before they reach this stage, and may even avoid the most severe symptoms. By building a large surplus of ventilators that can double as bi-level positive airway pressure support (BPAP) machines and be useful outside of an Intensive Care Unit (ICU)/Critical Care Unit, we can support hundreds of thousands of symptomatic patients at home, in field hospitals, and in clinics when the ICUs are full. The work of exhausted medical staff is simplified if they don’t have to swap out the ventilator for a patient often. Such a machine that can also function in an ICU is a combination of a BPAP machine and a critical care ventilator; in the past, we have not needed a special category for such a machine. Let us call them “pandemic ventilators’’.
It IS harder to support spontaneous breathing. You must sense the patient’s initiation of inhalation by an inflow or drop in pressure in the airway. This is algorithmically harder and requires sensors that make the machine a little bit more expensive. There is a world-wide supply-chain shortage of flow sensors right now, but there are several alternative open-source designs that can be made from common parts. (Our spreadsheet lists three on the “Flow Sensors” tab.) It also may require a more powerful and agile gas-delivery system, because not only must the ventilator deliver the required volumes, it must do so rapidly in order to maintain synchronisation with patients’ own efforts. Then the ventilator must be able to rapidly release pressure when it senses that the patient is trying to breathe out.
Engineering is about compromise. Adding sensors to rapidly developed ventilators, for example via our VentMon project, while adding complexity, dramatically reduces the risk of patient harm. The reward of expanding the mission is a pandemic ventilator that is far more likely to actually save lives.
The good news is that we are closer to the goal of making smart safe ventilators than you may realise. We have all the components we need and many clever designs. Our challenge is to synthesise and integrate the efforts of multiple teams in a way that extracts the best from each. Ventilators break down neatly into various components: electronic controllers and interfaces, oxygen-air blenders, pressure regulators (or pumps or bellows), inspiratory and expiratory manifolds, scavenging units. Each component has multiple subcomponents such as valves and sensors that can be addressed individually.
The quickest pathway to usable solutions is to start identifying and analyzing the best candidates for each particular component or module. Hopefully this is an effort that we can address collectively. Such component level reviews, if accompanied with links to complete open designs, are the modular, reusable building blocks of the solution we need to simplify rapid pandemic ventilator production.
To this end we have started laying out an initial framework for this analysis here and collecting modules and components on tabs of a spreadsheet. We encourage you to assist us with these efforts, or to provide links to those that are either under-way or complete.