Project Open Air | Hacker News Reader

Project Open Air

projectopenair.org

84 pointsmhb28 comments

"Open source ventilators" are a really really bad idea.

Patients on ventilators need to be intubated first, which involves using a specialized tool to insert a tube down past the vocal cords and inflating a balloon that seals the airway. You can kill a person before you've even turned on your nifty machine.

Once you have a patient tubed, you have full control of their respiration. If your calculations are wrong you either over inflate the lung and they literally burst, or you under oxygenate and the patient dies of oxygen deprivation.

I know engineers look at something like a ventilator and go "you don't have enough machines? I can make machines!" This is very admirable and I appreciate the desire to help, but I can't stress enough how dangerous this is.

Edit to add: This study showed that trained paramedics were only successful in tracheal intubation 52% of the time. You have a 0% chance of doing this at home. https://www.jwatch.org/na47414/2018/09/07/maybe-paramedics-s...

Agreed. Need for ventilator is quite bad enough state for the human body. And it gets complicated further after intubation. Most patients have to be sedated to reduce the distress and anxiety caused by the endotracheal tube. Extubation is also a risky process and patient may have to be re-intubated when though they have recovered but stayed in for a recommended max period around 2 weeks and lungs got weak. When patient is on ventilator, it requires many control parameters to help a patient live longer - the sedatives have to be just right depending on the patient, antibiotics, secondary infection risks, intravenous needles risks, pre-existing illnesses, allergies, etc, etc, etc. All this assuming that the ventilator machines and peripherals are perfectly safe. On a large scale during a pandemic, there aren't just enough doctors, experienced nurses and heath care workers who that can deal with it. Engineers can build the ventilators, but the logistical part is still a nightmare.

This is not a fair assessment and I think it sidesteps the point.

"This is very admirable and I appreciate the desire to help, but I can't stress enough how dangerous this is."

Indicating that 'medical equipment requires specialization' is not a special insight that most people are going to miss.

Assuming that the product will be designed by those who lack credibility or that such equipment might not be of sufficient quality of design, construction, or that it may be misused is unwarranted.

'Open source' encryption and other systems are behind the entire world's financial systems, it has integrity partly because it's 'open source'. It's not designed by fools, and just because jokers and everyone else can 'fork and do something' doesn't mean it will be implemented anywhere that matters. Obviously regulations, security, risk etc. play a role 'where it matters'.

Any kind of medical equipment development is going to require some kind of oversight, regulatory approvals, and special manufacture, but this is all feasible within the concept of 'open source'. There are plenty of highly credible and knowledgeable people, possibly even those who have worked on 'closed source' ventilators can possibly contribute.

Sheppard through some kind of regulatory process, manufacturing setup, and possibly deployment and training, it's not infeasible at all that something like this could help.

Even the availability of high-quality, unlicensed equipment designs, with parts pre-designed for manufacture and ready to be made by qualified entities ... could feasibly make all the difference in the world.

For example, once a design is complete and approved, the medical wing of a poor country, or possibly an NGO could facilitate the actual manufacture and distribution to medical staff in various places.

There are massive margins (bordering on criminal) in medical equipment and supplies these days, hospital beds now cost more than automobiles (!), it's high time there was an open approach to such issues.

I can absolutely see medical researchers and PhD's at universities, teaming with Engineers, MDs, and manufacture/supply chain specialists to produce this kind of stuff.

Obviously there's a limit to complexity here, but I hope they figure out not only ventilators, by any number of other kinds of gear.

Though odds are it will fail, it's not a bad idea, it's a great idea.

> Assuming that the product will be designed by those who lack credibility or that such equipment might not be of sufficient quality of design, construction, or that it may be misused is unwarranted.

Given the number of software engineers who have written detailed and strongly opinionated analyses of the situation rather than relying on expert opinions... I'm not so sure we can assume that this project is operated by experts.

Ventilators are not lacking a design. There are many functional designs across a variety of medical device manufactures (most of which are sub-$1000). China is already ramping up production to assist other countries.

The only thing open sourcing a design does is encourage DIY home builds, which is wildly irresponsible in a time of panic.

Neat idea, though I think it will be quite some time before it produces something that people can build or use.

Another option might be to figure out ways to adapt existing ventilators to be used by multiple patients. If you find 4 patients with similar tidal volumes and body size, you can grab some Y tubing, attach them to the same ventilator, then set it to 4x the tidal volume and breathe for all of them.

This has been studied before[1] and was done in at least one mass casualty event[2]. The main disadvantage is the risk of spreading germs (though the air return hose is separate from the incoming hose, so that shouldn't be too much of a problem) and the fact that patients sharing a ventilator require more careful monitoring of blood gasses.[3]

I doubt patient outcomes will be as good as having one ventilator per patient, but if it can increase our capacity by 4x overnight it seems worth pursuing.

1. https://www.ncbi.nlm.nih.gov/m/pubmed/16885402/

2. https://epmonthly.com/article/not-heroes-wear-capes-one-las-...

3. https://www.resuscitationjournal.com/article/S0300-9572(08)0...

My understanding is that what a ventilator does is basically:

- supply pressurized air to patient - turn off the pressure, open exhalation valve and let patient exhale

Given that every hospital I've seen has compressed air outlets, wouldn't it be possible to get acceptable results with a pressure regulator and a timed set of valves? Or is the pressure required to fill a patient's lungs so high that the volume must be controlled?

My layman's understanding is that ventilation is a fairly complex art/science, and that modern ventilators have multiple modes that are effectively different control laws.

I think that the pressure required to fill a patient's lungs is quite different in a COVID patient than a normal patient, because the disease has caused huge flow restriction in the lungs. I think that generally you want to control flow rate, because you need to get a particular respiration rate, and the lungs can hold a certain amount of air. You then use whatever pressure gets that respiration rate.

If the pressure or flow rate is too high, I think that that can cause lung damage, so one option is to increase the oxygen concentration in the air you're delivering. My understanding is that the downside is that although this will lead to more oxygenation, it doesn't help any with removing CO2. If you reach this limit, the only option is then ECMO.

> My understanding is that the downside is that although this will lead to more oxygenation, it doesn't help any with removing CO2.

If your input is O2 (or O2+N, no CO2) and you run part of the pressurization cycle with the exhalation valve open, won't you displace at least some CO2 (not in tissue, just in the lung cavity)?

This is madness. You can’t scale up ICU based ventilatory support to keep up with a pandemic.

All resources should be going towards infection control measures, a quick and highly scalable diagnostic test, and antiviral therapy. In that order.

Exactly, you can't fix an exponential problem with a linear solution. If the caseload doubles every two days, shutting down cities two days earlier is better than doubling the number of ventilators (because even with ventilators, more infected means more deaths).

Note that Wuhan shut down when they saw 400 new cases in a day, and their hospitals were still overwhelmed. The USA is over 400 new cases daily already. That's comparing a country to a huge city, but in exponential terms that only buys you days at most.

See https://medium.com/@tomaspueyo/coronavirus-act-today-or-peop...

Also people don't seem to realise, a patient on a ventilator is extremely unwell. They are the most unwell people in the entire hospital. They have 24 hour 1 on 1 nursing, constant invasive monitoring of vital signs, blood tests, X-rays, CT scans, antibiotics. They are managed by a team of extremely well trained and highly experienced senior medical professionals. And this goes on for a week or more. Then they leave ICU and proceed to be the most unwell people in the hospital except for those still in ICU. They might have various other complications, and need several more weeks to recover, often needing inpatient rehabilitation.

This does not scale.

You can scale ventilator support faster than you can develop a largely untried branch of medicine. Antiviral drugs barely even exist. Hoping they come along fast enough to help with what's likely to happen in the next four weeks is a pipe dream.

I think it's blue sky, too, what they're doing. But what the hell, maybe it'll help. It's not as if that Slack channel would be synthesizing remdesivir by the ton if it weren't doing this, after all.

> You can’t scale up ICU based ventilatory support to keep up with a pandemic.

It's almost like people think individual lives matter and don't have control over the larger response?

No real information. Marketingy-pictures. This seems like total fluff? I feel like the other HN post about DIY ventilators is a lot more interesting.

The point is the gathering of a network of people to tackle open engineering needs.

Let's be positive. Though it's a long shot and probably unlikely usable in the developed nation context, there are 7 billion people on planet earth many of whom might be stricken in the coming months (and years), if someone can put together a solid, reliable, simple piece of gear that works, it's not unfeasible that 'someone' could pay for rapid manufacture (this is definitely possible) and distribution in far-flung places wherein 'basic' might just be a lot of help. Or maybe not, but it's maybe worth trying.

It 'starts somewhere', so I hope it works.

Or you could just order one on Alibaba.[1]

More likely to work than something from a website-first operation with no product and no manufacturing capability.

Now that the coronavirus epidemic is winding down in China, demand for those units is probably down in China, so getting delivery is probably not too difficult.

[1] https://www.alibaba.com/trade/search?fsb=y&IndexArea=product...

ICU ventilators are surprisingly complicated machines, we've just brought some HAMILTON-C6 machines at work if you want an idea of the top of the line

I would consider the following a bare-minimum feature set for a COVID patient ventilator, any less and it would do more harm than good:

- cycle between an inspiratory and expiratory phase

- during the inspiratory phase, deliver an adjustable volume of gas (in the region of 6 ml per kg of patient's body weight) using as little pressure as possible, with an adjustable upper limit of pressure (in the region of 30 cmH2O)

- during the expiratory phase provide an adjustable pressure against exhalation (in the region of 0-30 cmH2O)

- allow blending of air and oxygen to deliver an adjustable inspired oxygen fraction

- allow the timing of the inspiratory and expiratory phases to be independently varied, thereby allowing the respiratory rate and the ratio of inspiration to expiration time to be controlled. Permit respiratory rates in the range of 8-60 breaths per minute

- measure and display the pressures and volumes within the respiratory system

- allow adjustable alarm-limits for pressures and volumes, and provide clear audible and visual alarms if these values are exceeded

For added patient safety and benefit, the following would be helpful

- measure inspired and expired oxygen and carbon dioxide content, and display both on a continuous waveform graph

- allow the patient to initiate the inspiratory phase by sensing patient inspiratory effort and providing pressure support for inhalation; ie sense when the patient inhales and deliver 10-15 cmH2O pressure for 0.5 seconds to augment inhalation

So somebody built a website, a google spreadsheet, a github repo with a readme, and a slack channel. I am skeptical. This seems like a promotion for the company hosting the project.

I keep wondering if existing CPAP/BiPAP machines can be software adapted to serviceable ventilators. Some BiPAP machines are identical to units produced for NIV.

Yes. Very easily.

Has anybody done it? Is there any effort to collect them incase ventilators get scarce, I wonder?

An anonymous spinup website. For some reason, some feel the urgency to put affairs into order on when its trending. Not before that, though?

Wouldn't hospitals, WHO, UN, whoever be able to come together with their capital and just buy out a medical company?

Wouldn't this have to also get past regulations? I'm not too sure if I want a kickstarter for medical equipment. I'd love to see medicines and medical equipment made open and maintainable though!

I wouldn't be surprised if this is the reasoning behind Apple and the likes preventing any new apps with the corona theme into the app store unless they come grin reputable sources like WHO.

All the activity is on slack, so pay no attention to what’s on the site. :)

As to whether people are serious, here’s a document from one of the channels. https://docs.google.com/document/d/1RDihfZIOEYs60kPEIVDe7gms...

How about make apps that make contact tracing easier?

That train departed a few weeks ago.

Reply on news.ycombinator.com