This is not deep learning related, but there are a lot of medical experts in this community. This is probing ideas and feasibility.

When people talk about the lack of ventilators, what kind of ventilators are they talking about? New York Gov. Cuomo mentioned the lack in his press conference a few days ago, and John Oliver on Last Week Tonight. This comment on a Hackaday thread clarifies a ‘high-flow nasal cannula NIV’ (non-invasive ventilation).

What are the required functions of a ventilator/cannula? Airflow and Humidification should be relatively simple with material from hardware stores and repurposed electronics. How important is Oxygenation? Apparently medical manufacturing companies use filters and sieves to concentrate O2 instead of chemically producing it.

The motivation is: For local businesses doing emergency manufacturing, could it be feasible to build simple ‘ventilators’?

Thanks Jeremy for the COVID-19 segment of the lecture; I wasn’t aware of the just-in-time manufacturing efforts underway around the world.


I share the sentiment. How hard can it be? But even for the simplest of designs, medical device manufacturers are met with an extremely high regulatory bar. After all, they are used in contexts where patient lives are at stake.

In the US manufacturers are subject to Code of Federal Regulations Title 21 as regulated by the FDA. Similar laws are enacted worldwide. The recognized approach here is always to start, not with design technicalities, but with patient needs. Regulators require manufacturers to specify the intended use of medical devices, whereafter they will classify the device in one of three risk categories.

A new ventilator will likely be considered a Class II device by the FDA. Given that, the odds of it being helpful in the context of COVID-19 are slim to none. One way to think of it is; if you are manufacturing butter knives your controls may be lax. But a much stricter set of rules will apply, to the exact same knife, if it is intended for use in heart surgery.

1 Like

I came across this last week:

The news is only in Italian so far, but will post as soon as it becomes more spread. In one of the most affected part of Italy they have got an idea, prototyped and tested in 72 hours (thanks to the help of a company that produces ventilators) ventilators that can successfully sustain two people rather than one. Local government has ordered a bunch of them.
Will keep posted, as this is potentially going to be a massive help

In case of interest, especially for physicians in this forum who haven’t read this article yet, please see

Publications mentioned in Dr. Babcock-Irvin’s video found in the article above:

Also, here is a relevant Twitter thread by @BonKu, an Emergency Medicine Physician and Director of the Health Design Lab at Thomas Jefferson University