If you’re involved at all in the micromanufacturing, hardware hacking, or open-source hardware communities, or interested (as I am) in their potential for economic relocalization and for undermining corporate power, you’ve probably seen a story going around about makers in Italy 3D printing valves to keep ventilators running for COVID-19 patients in critical condition. According to an article in Fast Company, a hospital in Chiari, a small town in Lombardy, couldn’t keep ventilators going for its coronavirus patients because it had run out of replacement valves and the manufacturer — Intersurgical — couldn’t supply any more on short notice. Cristian Fracassi and Michele Faini, using equipment at the digital manufacturing lab FabLab, printed out 100 replacement valves at a cost of around one Euro. Since the manufacturer — citing legal and safety issues — refused to release the design files, Fracassi and Faini reverse-engineered the valve design.
Although the tone in maker communities where the story has been circulated has been largely celebratory, it occasioned some debate over safety issues. One person with a background in biomedical engineering, who goes by @turzaak on Twitter, raised the question of whether the materials were biocompatible and whether they were sterilizable. She suspected based on the valve’s appearance that it was printed from Polylactic Acid (PLA) or Acrylonitrile Butadiene Styrene (ABS), two common feed stocks used by 3D printer hobbyists. According to @turzaak these materials may be difficult to sanitize, decompose when exposed to moisture, or outgas toxic chemicals.
The actual information available on the safety of the 3D printed substitute is mixed. The Fast Company article may (or may not) provide at least partial assurances on the sanitary and biocompatibility front: “Though this was the first time Lonati SpA has printed something for the medical sector, Faini says the company’s SLS 3D printers can print with PA12, a material that can be sanitized and used for biomedical purposes.” Still, as @turzaak explained in response to my further query, PA12 isn’t used in biomedical devices because it’s a type of nylon and nylons soak up water.
In any case, according to the manufacturer the actual physical design of the printed valves may be suboptimal. “The original pieces work better than the 3D printed ones… because the intricate designs and small holes are difficult to 3D print. Fracassi also denied that the commercial valves were as expensive (10,000 Euros) as reported in some media outlets, although he did not indicate what the actual price was.
A March 14 article at The Verge reported that the unauthorized replacement valves have been used successfully on ten patients so far.
The same general caveats — sanitizing, biocompatibility, toxicity, etc. — also probably apply to older open-source ventilator projects like the Pandemic Ventilator prototype posted at Instructables twelve years ago during the avian flu (see also this). The discussion of materials devotes little or no attention to safety issues, and might well raise any number of red flags based on @turzaak’s long thread on biocompatibility issues involved in respirator design (“Biocompatibility. Basically it means you want materials that a) don’t fall apart, and b) don’t poison/kill when you use them in something that’s gonna interact with the gooey meat stuff.”). This part of the thread, in particular, made me think of the discussion of microcontrollers in the Instructables post:
Say you want to put some fancy electronics in here. What do you reach for? If you said “my spool of wire”, congrats, you possibly will kill someone, unless that spool is gold wire. Copper is toxic. You had better be REALLY SURE about encapsulation or not use it at all….
That means your adorable little RPi or Arudino control? Keep it far away from anything bio or dunk it in potting material. So much bare copper.
Naomi Wu (@RealSexyCyborg), a prominent figure in the maker community, raised similar concerns regarding 3D printed masks. The materials lack the flexibility needed for proper facial contact, she said, and the designs are mostly a gimmick to promote 3D printing and “sell unicorn filament.” Sewn cloth masks with bendable wire frames are far more appropriate.
In the debates over open-source ventilator designs and 3D printed parts, some critics have gone so far as to generalize that all regulatory design specifications exist purely for safety reasons, and that disregarding any of them is unethical. I’ve also seen claims that the elevated price of the commercial respirator valves (although what that price actually is seems to be up in the air, as we saw above) is owing entirely to the cost of biocompatible materials and not to patent markups.
Anyone familiar with the centrality of intellectual property to corporate profit models, or to the share of prices made up of embedded rents on intellectual property, will no doubt be skeptical of such assertions (not to say charmed by their naivete). It doesn’t take much of a tinfoil hat to suspect that many of the regulatory design specifications in any industry have more to do with grandfathering in the products of incumbent firms, or suppressing competitive pressures for design innovation, than about safety as such. Consider how much of building codes are written by lobbyists for incumbent contractors to prevent competition from cheaper, vernacular methods, or how zoning and other local regulations are used to protect brick-and-mortar businesses from competition by things like food carts or home-based microenterprises.
The real difficulty facing any would-be open medical hardware project lies in determining which of the specifications reflect genuine safety concerns, and which reflect corporate interests in enforcing artificial scarcity and using artificial cost thresholds as entry barriers to restrict competition. The best way to do this is to eliminate the stovepiping, and improve communication, between the engineering and hardware communities on the one hand, and the medical user community and biomedical hardware specialists on the other. All these stakeholders need to be integrated into a cohesive p2p community at every stage in the design process, with constant feedback from the end users. That’s supposed to be one of the strong suits of commons-based peer production.
The supply chain disruption from the coronavirus pandemic is demonstrating the necessity of industrial relocalization and production for local use. The high cost of medical equipment and short supply of replacement parts is also showing the usefulness of micromanufacturing and open-source design. In both cases, micromanufacturing can be a lifeline to provide import substitution and increase community resilience in cases where the corporate logistic chains have failed us. But if the maker community jumps in to act as saviors without adequately considering the needs and concerns of its end users — particularly, in this case, safety — or listening to their own technological knowledge when it’s relevant to design, it will have wasted its opportunity.
Note: I am not an engineer or a maker. I didn’t know PLA from ABS until I Googled them. Nothing in this op-ed should be construed as a firm statement of opinion on my part regarding the safety issues around 3D printed ventilator valves or open-source medical equipment design more generally. This is simply my good faith effort to repeat the concerns or arguments of everyone, on multiple “sides” of the debate, who all know far more about the subject than I ever will, and to spark further discussion. I hope I pulled together information from the parties involved in a way that facilitates such further discussion.