Interview with German Ventilator Manufacturer

I'm not so sure. We have an ongoing toilet paper crisis that is in it's third week now and the government is unable or unwilling to do anything about it, despite ensuring everyone that there's plenty of toilet paper somewhere in some storage of some company. If a somewhat trivial problem like coordinating a few trucks is beyond their ability, I have no hope for coordinating planes and import/export.

You might say "but that's toilet paper, who cares", but empty shelves in super markets in Germany are a great way to make people remember the past and go into panic mode.

What people do right now is buy masks individually from Chinese vendors on Amazon, and they are getting shipped. If you want them quickly, you can pay for airmail. China can deliver, but for some reason, Germany doesn't take them.

If you're interested in the complexities associated with retrofitting CPAP machines, I've written a couple of articles about it:

https://blog.plan99.net/cpap-for-covid-d47886bf978c

https://blog.plan99.net/more-cpap-for-covid-b6911f806c89

The primary difficulties are establishing medical consensus on what kind of pressure therapy is best (continuous vs bi-level) because that affects the availability of machines a lot and there's some disagreement. Secondary is how to stop patients on CPAP machines spraying virus into the air through the masks, however, at least one CPAP machine manufacturer thinks that concern is actually overblown and points to a study showing CPAP masks don't aerosolise to the extent current medical consensus seems to think.

Something very clear is that doctors are still figuring out treatment for this and there's some disagreement over what the right way to do things is. Disturbingly it may be that current medical consensus makes things worse for COVID-19 patients rather than better i.e. bi-level pressure therapy hurts the lungs in that state rather than helps (perhaps contributing to a high death rate?).

(Not an expert on ventilators, but I have been working on software side of medical devises.)

Producing semi-mechanical Bag-valve-masks (BVM) would be relatively easy, but you need professional to sit there and monitor the patient. That's not ICU ventilator. BVM is very crude tool for short term use. If the rules and legislation are relaxed so that hospitals could use non-certified crude devises that help. You would almost certainly see ventilator-induced lung injuries.

Ventilator-associated lung injury https://en.wikipedia.org/wiki/Ventilator-associated_lung_inj...

Ventilator-Induced Lung Injury Review (Part 1 of 2) https://pulmccm.org/review-articles/ventilator-induced-lung-...

If the system is automated life support, it generally has to maintain homeostasis of the patient and _dynamically_ adjust. If you are just tiny amount off from the optimal and there is drift in the dynamical adjustment it accumulates over time and you can kill the patient or cause immense pain or discomfort.

Oxygen is poison if you get it too much and if you get it too little, you die. Pushing air into the lungs can damage them and the oxygen mix should adjust. If lungs are filling with fluids, you might need more oxygen and less volume. The system must be connected to instrumentation that monitors the patient's state continuously.

There is the question of medical staff using the thing correctly. Medical UI/UX is safety critical in the same level as it is in aviation. If just 5% of users use it wrong it, can negate all benefits from using the device.

Generally you need deep knowledge transfer to scale up production of critical components or testing and calibration, just emailing the specs is not quick enough. You need to test it with real environment and refine the device to see how it works in practice. I'm sure that ventilator manufacturers have already outsourced every component that is easy to manufactured and produce to others. There are some bottlenecks that involve testing, special machines, materials and calibration that is not easy to scale quickly.

(edit: the time-frame of ICU overflow seems to be between May-Sep in the most countries in most models If you ramp up deliveries 3 months from now, you could catch the tail end of the need)

Any government could clone Twitter if they must. It's not rocket science and Twitter is actually quite open about their architecture and technologies choice.

You can use that mentality on things that are not safety critical.

How could anyone argue that once in a while there won't be enough beds to save everyone's life?

The closest things I can think of non healthcare is that airbags are designed for 5, 50, and somewhat for 95%ile individuals. But I think it's actually impossible to develop for everyone due to mechanical restrictions.