start	end	text
540	7920	Hello, everyone. Thanks to Nina and everyone at Biohack Village for having me. My name is
7920	13100	Anthony DeFranco, and I'm one of the founders of the Open Insulin Project. And I'd like to
13100	22020	tell you a little bit about the general problem that we're addressing and how we're approaching
22020	29980	it, and how you all can maybe get involved if you're interested. So let me begin.
32560	44240	The basic fact is that worldwide 50% of the people who need insulin to survive don't have access to
44240	59980	it. And when we take that figure into account, and also look at the population of people worldwide
59980	68720	with diabetes, and some bounds that we have on the death rate that's directly attributed to it,
68720	78540	and we also look at the insulin access statistic here, and we put all of the
79260	85400	information we have together to get a rough estimate of what the
87320	94260	impact of this situation is, where, you know, fully half of people who need insulin to survive can't get it.
95800	104580	A very rough order of magnitude estimate is that about 20 million people, you know, 10 to 20 million
104580	112340	people a year are dying due to lack of access to insulin worldwide. And to put that in perspective,
112340	119340	12 million people per year was the rate that people were dying due to World War II
120020	125380	from all causes, which was the deadliest conflict in history. So
129940	135360	this situation where half of the people in the world lack access to insulin
136540	147240	is really a crisis on a historic level, you know, it's something that it's hard to find anything that compares to this.
148460	158800	However, unlike a war or, you know, a very kind of complex problem like that,
159840	166600	the fundamental technical realities of making insulin and getting it to people are fairly simple. So
168020	171860	we're going to look a little bit at why, despite that,
172480	175720	there is a problem of this magnitude and how we can
176400	180200	try to approach and how we are trying to approach it in the Open Insulin Project.
183840	186540	So that's kind of the worldwide situation.
189360	194920	Here in America, where the many people in the project are based,
194920	198380	we know a little bit more.
199180	203560	Now, the worldwide situation is largely due
205100	215140	to, you know, supply chain issues and the kind of geopolitics around
215760	223080	different countries' access to markets. So those issues don't really exist within the U.S.
223240	228340	The U.S. has plenty of infrastructure to get insulin to people.
230560	235800	You know, the technology to produce is well established here. So
237420	243620	it's a little even more puzzling to note that there's still a significant problem in the U.S.
245140	248800	So we want to look a little bit at that.
250200	257540	So here are some figures on just the prevalence of diabetes in the U.S.
258180	267060	You know, almost half of people have either diabetes or a risk of developing it in a fairly
267060	279800	short term. Just diagnosed diabetes alone is a over $300 billion a year cost for everyone.
280180	287080	Having diabetes significantly increases overall health care costs.
291840	300400	And there's a bit more detail on how, you know, diabetes is contributing to the global
300400	304960	death rate. So this is another kind of data point, a little bit more specific about what
304960	313340	we were talking about earlier, where, you know, we can actually say that the deaths are directly
313340	327100	attributable to diabetes, also quite significant. So, you know, why is this eminently treatable
327100	336880	disease still causing so many deaths? And why is lack of treatment even in an industrialized,
338580	344160	advanced technologically and infrastructurally country like the U.S., you know, why is this
344160	350620	disease still causing so much harm and so much clearly preventable death?
351860	362180	Well, one of the factors is that the, you know, supply of insulin, even though it's very easy to
363140	368240	empirically speaking, just has a lot of trouble reaching the people who need it.
368720	373740	And the first step in that process is the manufacture of it. And
376420	385540	globally, three manufacturers control almost the entire supply of insulin.
385540	394900	And that is true in the U.S. as well as true everywhere. So, you know, already we're seeing
394900	402620	that at the very first step in the supply, there is a kind of a restriction that may prove
402620	415460	significant. And as a result, the list prices have risen on a, you know, roughly exponential
415980	426700	trajectory. So this has been part of a pattern of all of the major manufacturers
427900	434200	raising their prices at about the same time by about the same increment for decades. And,
436080	443080	you know, this has resulted in the list price becoming unaffordable for many people.
445320	450920	And because of the complexities in the U.S., for instance, you know, the U.S. kind of has a
451600	460820	strange healthcare system where, you know, some aspects of the prices are subsidized in various
460820	466340	ways by different parties who negotiate amongst themselves. You know, there's a very complex
466340	472120	system involving insurers and manufacturers and pharmacy benefits managers and pharmacies.
472700	479320	And there's a set of complex and mostly secret agreements among them as to, you know,
479320	485100	what people are actually paying. But anyone who doesn't have insurance, which, you know,
485100	491040	is tied to having a job largely in the U.S., you know, they're likely to have to pay that list
491040	501500	price and to be unable to afford it. So this is, you know, implicated in what the problem may be
501500	509880	as we come to understand it more. And then the result of that has been that one in four people
509880	519860	with diabetes in the U.S. have reported rationing their insulin and not taking as much insulin as
519860	529100	you need puts you immediately a much greater risk of the complications of diabetes. In the short
529100	539480	term, you can develop diabetic ketoacidosis, which is potentially deadly within hours or days.
540040	544360	And in the longer term, if you're not taking as much insulin as you need,
544360	552500	you can develop the longer term complications of diabetes, which are many of which are listed here,
552500	559080	you know, blindness, cardiovascular disease, and amputations, for example.
562760	571960	So it's a rather grim situation, but it's kind of difficult to understand why this would be the case
571960	579500	since insulin in some form or another has been produced as a medicine for almost a century.
579500	588180	And even the state-of-the-art forms of insulin, you know, the most modern forms that we have are
588740	598340	largely off patent. So why is it that there's still this access crisis? Why are only, you know,
598340	604480	three manufacturers still controlling almost the entire market and able to price the insulin out
604480	613880	of reach of many people? So there have been some investigations into this where people have looked
613880	623100	at the price increase patterns and are, you know, finding evidence that there may be collusion. So
623800	630820	some of these cases are still in progress here. But, you know, the fundamental reality is that,
631420	637500	you know, whether or not there is explicit collusion here, there is this pattern of
637500	642440	price increases and it has become unaffordable. And
645400	651160	at the technical level, there's really not that much of a challenge to making insulin. So
652940	657200	that's what we're focusing on now in our project is how to make insulin.
661040	664840	So this is part of a broader problem in
667760	677240	how health care and the economy around medicine is, you know, thought to
679300	686640	what the kind of dominant philosophy and how to manage it is. There was a
688480	695460	kind of economic theory that gained a lot of credibility in policy circles in the mid-20th
695460	704900	century that said that, you know, making medicine is kind of different from
706300	709180	other economic activities in that
710620	719180	subsidies are needed to do all of the research involved and that a monopoly is
719180	729420	an expedient form of subsidy. So that, you know, that would be a sensible policy to tolerate and
729420	737420	even encourage monopolies around medicine. And these claims actually collapse fairly
737420	746560	quickly under scrutiny. You know, most of the R&D involved in making medicines is already done
746560	756140	with public money and universities and the companies themselves spend much more of their
756140	762000	own budgets on other things like marketing and executive compensation than they do on R&D.
762000	768960	Also, this belief that monopolies are essentially harmless has been, you know, roundly debunked
768960	778840	since then. But nonetheless, that's the system that we've inherited. So the result of this is that,
779340	785160	you know, we've had this complex system of agreements grow up around it, you know, with
786120	790720	all those other, you know, parties like pharmacy benefits managers that we'll look a little bit
790720	798760	more at later in the presentation. And, you know, they kind of partition up the market and
800960	808860	it's very difficult to do anything outside of one of these kind of proprietary partitions,
808860	817320	or to go between them, you know, even to the point that some people who are allergic or don't
817320	821600	tolerate one brand of insulin are unable to get it covered by their insurance because the insurance
822480	828800	only covers one brand at the preferred tier. So even when there's clear medical necessity,
828800	838760	it still doesn't matter in this system. So, you know, it's not much of an economy in the end,
838760	849520	because it's so thoroughly, you know, shot through with monopolies and monopolistic arrangements.
851940	859040	And we see, you know, the problems that are typically associated with this.
860440	867060	And, you know, one of the most notable is pricing the poor out of
869020	872560	the market, but also, you know, you can see that
874540	880840	progress that would threaten an existing
882900	889100	line of business is discouraged. And, you know, one example of that was that,
891060	897220	you know, an analyst here at Goldman Sachs is asking, is curing patients a sustainable business
897220	905480	model? You know, the answer in an economy where competition is essentially impossible is no,
906160	913480	because, you know, the profits are already being maximized by chronically treating things,
913480	918800	and so curing them doesn't make any sense if people don't have any alternative.
919200	925220	And so that's kind of the unfortunate result of this. And this is something that, you know,
925220	938760	many have researched this. One of my favorite theorists of this is Mr. Schmitz at Minneapolis
938760	947460	Fed. So, you know, he has looked at how monopolies actually do cause the harms that we've discussed.
947600	954720	And in the case of medicine, the harms aren't just economic, they are directly related to
955900	964040	you know, lost lives and, you know, really tragic and tragically avoidable health consequences for
964040	974760	people. So one of the things that is noted in general that we've seen in insulin is,
974760	981440	in the case of insulin in particular, is that, you know, we have
985100	991460	seen people, you know, turning to crowdfunding because the cost of insulin when they lose
992480	1000780	health insurance or health coverage of some sort is, you know, more than many people make in a
1000780	1007740	month. It could be many thousands of dollars a month. So this is really a quite egregious example
1007740	1015200	of how, you know, people get priced out of the market fairly quickly. And, you know,
1015200	1025600	it's not just a subtle effect, it is a really, you know, a very large difference in prices that
1025600	1032080	people have to bear to the point where, you know, it'll be exceed everything else that they're
1032080	1042360	paying for in life by quite a bit. So just to make that a little bit more concrete, the cost
1042360	1049360	of manufacture of insulin is about $5 per vial, but the list prices have risen to the point where
1049360	1059580	they're around $300 to $600 per vial in some cases. And, you know, these conditions have
1059580	1066460	persisted for decades, and they've only gotten worse. There hasn't been any successful pushback
1066460	1077760	against them yet. So what we are trying to do is kind of invert everything that we've observed in
1077760	1088660	our project. So instead of relying on proprietary development and mechanisms of monopoly, we are
1088660	1095840	sharing everything that we develop in a commons, and we are looking to produce insulin in
1097540	1102100	low overhead distributed networks of small scale manufacturing sites.
1103060	1110240	And instead of having this really strong conflict of interest between the producers who are
1110240	1118200	motivated by profits and to deliver profits to investors, instead we're looking to have the
1120020	1129380	people who use the insulin actually be directly involved in the governance of the organizations
1129380	1137360	that are making the insulin so that they'll be able to make decisions as to how everything is
1137360	1144800	managed in the interests of themselves and each other and the people using the insulin.
1144800	1153600	And a helpful side effect of this will hopefully be that we'll be able to be a lot more nimble and
1157340	1162740	actually make a lot more improvements and not be incentivized to keep old things around just
1162740	1170260	because they're already profitable and actually move things forward and hopefully make progress
1170260	1185220	And in part, we're inspired by other successes in the diabetes community.
1185960	1195900	So although we're the first effort I know of to make insulin itself open source, for
1195900	1205680	about 10 years now, people have been working on artificial pancreas systems, which is essentially
1206020	1212900	a software algorithm that controls an insulin pump to determine the timing and the amount
1212900	1218240	of the dose of insulin automatically, usually based on feedback from a continuous glucose
1218240	1227920	monitor. And this is something that was already possible with technology from the 80s, but
1227920	1238120	the manufacturers just never seemed to be able to bring it to market for whatever reason.
1238300	1245260	But once the technology to do this became more widespread, you know, once
1246660	1254740	embedded computers became cheaper and easier to work with, and knowledge of software and how
1254740	1262360	to write these kinds of algorithms became more widespread, individual diabetics and small groups
1262360	1268740	were able to just pick up this work and bring it to completion within the space of a few years.
1268740	1277560	For almost a decade, there have been DIY artificial pancreas systems that, you know,
1277560	1282440	many people have been relying on and trusting with their own lives and the lives of their children,
1282440	1289320	and, you know, they're working very well. And this was done without any, you know, large R&D budgets
1289320	1294880	or institutional overhead. It was just people collaborating in these horizontal networks to
1295480	1300360	solve their own problems and each other's problems. So we're working on
1301940	1306020	enabling this for the insulin itself so that we can kind of complete
1307600	1317120	the open source economy for treating diabetes. And it turns out that the same trends that
1318520	1327680	made the DIY artificial pancreas systems possible are also going to help us.
1328940	1339140	The, you know, more pervasive and cheap computing technology is going to play some role in letting
1339140	1346080	us automate different aspects of the production and purification, and it also just makes it easier
1346080	1353700	for us to collaborate with each other. And so we're seeing signs that we can reproduce this
1353700	1366260	kind of success. And, you know, the result of that will be ideally to move this activity from
1366800	1373620	these, you know, proprietary fiefdoms and into the commons. And this is something that is
1374400	1380460	good for people. It's very good for business as well, you know, judging by how it's gone
1381240	1388840	with, you know, open source software businesses. But, you know, major difference is that it
1390860	1400920	actually respects and promotes competition and innovation instead of sabotaging it and,
1401840	1406660	you know, trying to go on needlessly suffering and dying for no reason.
1408120	1419420	So that's our goal. And we're working on developing both the technology and the organizational forms
1419420	1425540	that are associated with this to, you know, create a comprehensive alternative for the status quo.
1425540	1435700	And we're trying to prioritize the people who currently lack insulin and need it most in
1437770	1444960	how we develop things so that we can most directly address the most urgent problems
1444960	1453800	that people in society have. So this is a closer look at the existing supply chain
1456300	1462280	just so that we can contrast it with what we're setting up. So I mentioned earlier that there are,
1462280	1472120	you know, these multiple middlemen who operate in these complex networks in between the
1472720	1478240	manufacturer and the consumer. And so I won't even really try to go through this in any detail,
1478240	1485340	but you can see that there are, you know, quite a few complex patterns of interrelationship
1485340	1492540	among them. And most of these are done in secret, so there's no transparency into how
1493020	1501860	this works in any detail. And, you know, it involves kickbacks being passed up and down
1501860	1512200	the chain and, you know, all kinds of things that just are in themselves a great challenge to
1514000	1519340	making any sense of things. And so when the system produces the nonsensical
1520940	1527800	result of, you know, denying people insulin that they need to survive that costs
1528540	1533060	very little to make, you know, this is kind of how that is able to
1535040	1539600	grow and that kind of problem can grow and sustain itself in this environment of
1539600	1545000	artificial complexity. So we're going to address that by
1546600	1554840	making things as simple and as direct as possible. So right now we're working in our community labs
1554840	1563180	to develop the microbes and the technology, and then we're going to publish the information
1563180	1571160	associated with them. And then using that public information about how the technology works,
1571160	1577940	we're looking at setting up a system of cooperatives where the insulin users can
1577940	1586580	participate directly in the ownership and the governance of those organizations. And
1587380	1594940	we may also be partnering with hospitals and local pharmacies that already do work that is
1595780	1602560	similar in some technical aspects to this, and they already have, you know, people who
1603580	1611580	would need the insulin, they have relationships with them, and we may also be looking at
1611580	1620960	collaborating with local and state governments. And, you know, aside from these fairly small
1620960	1630020	and simple low overhead organizations, it's going to be as direct to the users as possible.
1632140	1638360	So a closer look at some of the technical work that we're doing.
1639720	1647880	And this is, you know, this is just a general look at how protein production
1647880	1650440	purification works and how this applies to insulin. So
1653160	1660620	right now we've designed some genetic modifications to some microbes. We have
1662080	1665240	actually three systems we're working on now, this is two of them,
1665240	1670920	but we have two yeast-based systems for producing insulin glycogene, which is a
1670920	1679180	long-acting insulin, and one E. coli-based for producing LISPR, which is a rapid-acting insulin.
1679180	1689980	And so we have ongoing work with the microbe engineering. Some of it has produced some
1689980	1694320	small yields already, but we're working on getting the production to be reliable in the
1694320	1701680	practical level. And then once the microbes are engineered, we're going to focus on
1702400	1710920	a small-scale purification system, initially using proprietary equipment, but we're also
1710920	1722220	working on making open-source equipment for that. And we've also started to look at the
1725760	1732640	verification and quality assurance questions of making that product suitable for use as a
1732640	1739000	medicine. And so once we've got all of these aspects in place, we've got all the technical
1739000	1747380	problems for making insulin in a small-scale setting solved. So here's a look at
1749680	1759600	a kind of demonstration version of a bioreactor. So it is essentially just a bit of plumbing, and
1759600	1767520	aside from having to keep it sterile, there isn't anything terribly technically complex about it.
1767520	1777860	You're going to have some, you know, some pumps and some valves and temperature sensors,
1777860	1786940	pH sensors, and that sort of thing. And then for the purification, you're going to have,
1786940	1793720	you know, perhaps a column, which is a, you know, just a tube with certain ports filled with
1797140	1803620	some materials that may be a little bit costly, but for the most part, this is something that,
1803620	1809060	you know, you can, as these people have for this demonstration unit, they fit everything you need
1809720	1817320	to do it onto a tabletop. So, you know, we can work in a similar envelope.
1818820	1822940	You know, the corner of a room could be enough to produce insulin for
1823980	1830620	you know, a small city if we are, you know, running at a reasonable efficiency.
1832300	1839580	And as a result, we can bring the cost of insulin back down to
1840680	1848600	what the cost of production already is. So our own projections for what our cost of production
1848600	1854780	be are in line with what they are for the large-scale production that dominates right now.
1855320	1862060	So based on our preliminary model, we could be providing this to people for seven dollars a vial,
1862060	1867900	which is, you know, very close to the three to five dollars a vial that it's usually produced for.
1867900	1878480	So, you know, we're seeing that there isn't really any economy of scale here and that,
1878480	1889600	for the limited supply that we see right now. So as I mentioned, we are still working on
1891060	1896220	the microbe engineering, but we have started to work on the open hardware as well. So we're
1896220	1906300	making progress on the technical aspects of this process and we're looking forward to
1908300	1915820	reaching some of the milestones on that fairly soon. And we've, you know, begun
1916980	1923500	working not just on the open hardware but also on some of the regulatory and organizational research
1923500	1936180	to make that hypothetical system of cooperatively organized production a reality. So,
1936180	1942700	that's going to involve some outside-of-the-box thinking and we're, you know, getting into
1943370	1950020	looking at a lot of interesting precedents like ideas from personalized medicine and
1950740	1959730	other forms of cooperative medicine and health care and state-level frameworks for things like
1961190	1966970	direct public offerings, you know, financial mechanisms, and so it's really
1969710	1979070	taking us on both a deep and a broad look at a lot of very promising alternative ways of
1979070	1986930	doing things that we hope will apply not just to insulin but to perhaps medicine in general
1986930	1995390	and business in general in the near future as well. So here's a closer look at the bioreactor
1995390	2009750	design that we have just recently completed the first look at. So,
2010750	2017070	as you can see, the bioreactor, when we break it down to the specifics, as I was saying
2017070	2026110	before, it's kind of some pumps and some sensors and, you know, an air pump and some heating pads
2029430	2037710	and this is a design that was based on one that another fellow in the community,
2037710	2044170	Sebastian Kosioba, was working on. So he was kind enough to show us around his design and
2044170	2051450	we have a group of interns that's working on this right now. So this was made by Paige Hanoka and
2052630	2059470	it took her just a couple of weeks to come up with this and the bill of materials for this comes out
2059470	2071810	to $115.80. So this would have a capacity of a few liters and a capacity of a few liters if you run it
2074170	2080750	fairly close to continuously that can supply insulin for a few thousand people. So, you know,
2080750	2087370	this is still leaving out the purification and the QA aspects but, you know, those don't add
2087370	2092090	an order of magnitude more cost to this. So you can see that the, you know, capital requirements
2092090	2102570	for doing this are negligible. So we've talked a bit about how we're looking to organize the
2102570	2110210	production ultimately but we are also organizing the R&D work that we're doing right now
2111110	2118070	in a similar way. So we're a fairly horizontal open source oriented
2121510	2127210	organization. We're taking cues from other open source foundations that are stewards of different
2127210	2135650	kinds of technology. So we have groups working on the wet lab R&D, the hardware
2137230	2143530	supporting software for our organization, developing the business plans, doing the legal and policy
2143530	2152630	research, and doing communications with our supporters and touching on fundraising as well.
2154050	2161250	And here a few folks from the team, you know, we are a loosely organized network of volunteers so
2161250	2168670	people kind of come in and out. We weren't able to list everyone but here are a few of the more
2170130	2179070	regularly involved folks. You know, Jan is our chief scientist working on the microbe stuff
2181190	2188450	and David and Max are also working on that here in our Oakland lab. Alex Hay recently developed
2188450	2194930	an open source glucometer which we have a blog post up on our blog about which is very interesting.
2195530	2203830	Nicole has been helping us with the legal work that we're doing. Yifei has helped with software.
2204410	2212650	Louise is working on communications and fundraising with the interns. Jessica is also working on
2214630	2220190	the microbe engineering in the wet lab and with an intern group as well.
2220190	2223930	Remy has helped us with some collaborations and
2225450	2232390	Kadi is actually one of our international collaborators in Senegal who is getting her lab
2233430	2241130	set up now to join us and perhaps help us look at some new possibilities involving
2241130	2246110	plant-based production. So, very interesting potential there.
2248010	2255090	Within our group, we are fairly horizontal as I said. So, you know, we
2256450	2262610	kind of go on the open source credo of, you know, loose consensus and
2264330	2271110	working protocols we could say and just take on tasks on our initiative. For the most part,
2272070	2280730	so we've had to collaborate remotely. So, you know, we're doing things in online meetings,
2280730	2288130	which anyone is welcome to join. And, you know, if anyone is interested in that, you can go to
2288130	2294650	our website, openinsulin.org. And we have some posts up there that will give you some contact
2294650	2305210	information and information about the meetings. So, you know, with that, I'd be happy to hear
2305210	2309890	from any of you if you're interested in learning more about what we're doing or, you know, helping
2309890	2317010	with any aspect of this. And the more the merrier and the more people are helping out, the faster
2317010	2323610	we're going to be able to solve this very urgent problem. So, it would be a pleasure to hear from
2323610	2331870	you. And, again, thank you for having me here. And thanks to Nina and the Valhack Village folks.