I’ve been following the 23andme saga with dread and fascination and frustration since it hit earlier in the week. If you’re not up to date, David Dobbs has the canonical collection of stories to get you there.
This is a post about business models and culture clashes. It’s long. I should also start by disclosing that my former non profit project received funding from Anne Wojcicki’s philanthropic foundation, that I hold some shares in small startups that could benefit from the success of direct to consumer genomics (though not in 23andme) and that I now work at a non profit that is premised on the idea that personal access to data will fundamentally change research. My 23andme data is online for all to download.
So I could personally benefit if 23andme survives and thrives, and I personally admire Anne and her work, even when I criticize some of the choices she makes. It’s hard as hell to be a CEO and she deserves credit for icebreaking, vision, and perserverance.
That said, I am deeply frustrated by the simplistic narrative of OMG FDA BIG GUBBERMINT SILENCING DARING ENTREPRENEUR. It’s not that simple.
23andme has been trying to bring a technology business model, and a bayesian data culture, to the genetics world for six years. Both of those were gambles. And I think the problem they’re in right now is a direct consequence of those gambles not paying off yet.
First, by a technology business model, I mean the idea that you go get all the users you can get by giving your service away for free and figure out a way to make money later. This is no different than social networks business models. Facebook, Twitter, Instagram, Snapchat - all of them started with this concept of acquiring the most users and then working out a model for revenue. And all of them converged on advertising, because for all the promises of the new network, selling people sweaters worked better than anything else.
As a business model for genetics though, it’s different, because your cost isn’t just distributing an app and paying Amazon cloud fees. 23andme has been selling their direct-to-consumer genotyping kits for six years now, always at a deep discount to their actual price. They exacerbated that financial model when they dropped the monthly subscription requirement last year. If they can’t find a way to make money soon, they’re going to wind up in the pets.com box - you can’t sell a dollar for fifty cents and expect to make money on volume.
I would guess that the hope was that population genetics would yield a business model that wasn’t regulated. If you could get enough people in the door, then your internal analysis could inform efforts at pharma, at insurance companies, and more - without ever having to market your spit kits as medical devices in the regular business model. Their push to hit a million users indicate this remains a goal.
But since they’re not doing that yet, it seems they’re not at the magic number. That means going the traditional route, which is to develop and rigorously validate a test, get it approved by the FDA, and then market it at an incredible markup. That’s what 23andme tried to do in this run. They invited the FDA in the front door, and at least so far, haven’t been very good about serving them tea. We’ll find out more about why later.
But I posit the data culture that permeates technology companies is at the root of the frosty relationship between the FDA and 23andme. It makes it very hard to pivot from a model based on loss leadership and data mining to one based on regulatory submission.
That “traditional” submission to the FDA would be of a very specific kind of analysis based on randomized controlled trials. It is designed to keep bad things from happening to people, not to make sure good things happen to people. As one of my favorite papers lays out, parachutes would not receive FDA approval as a gravity-resisting device.
Modern tech culture doesn’t work that way. Bayes’ rule is about probability. It’s a different way of knowing that you know something, and it’s one in which there is far more tolerance for uncertainty than the FDA is accustomed to. And there’s been statements by FDA officials that they are deeply uncomfortable with that. Note in particular Janet Woodcock’s statement that causal inference needs a “level of rigor” - and the date in late April. It’s right before 23andme cut off communication with the FDA.
The FDA isn’t saying that 23andme needs to shut down, or stop giving people data, or stop providing spit kits. It’s saying their causal inference doesn’t hit the level of rigor that allows it to pursue a technology business model in the health regulatory space. That’s a gamble 23andme was always taking, that the FDA wouldn’t know how to regulate it in time for it to make money. It’s not an overreach, or a screw job on patients. It’s just business.
And I hope this isn’t the end of it. We need DTC screening. It helped me. It’ll help many others. But until the FDA learns how to deal with Bayes’s rule and its discomforts - and until DTC companies figure out a business model that isn’t based on massive loss leadership - we’re going to keep coming back to this clash of culture and business models. Both sides need to make some changes if we’re going to avoid doing this over, and over, and over.
Well, yeah. I’d agree with daring.
As of today, the FDA has absolutely smacked down 23andme’s attempt to sell spit kits for genotyping as medical devices. I encourage you to read the letter itself, it’s an unusually scathing tone for a rather studiedly gray agency.
23andme has been leveraging a gap in the market for a long time - the idea that their spit kit that creates a file of individual genetic variation is recreational, not a medical device that increases health. Over the past year their rhetoric around health has ratcheted up, and they’ve very loudly engaged with the FDA since the middle of 2012. They were proud of this fact. It set them apart from other DTC companies.
Since last year, I for one have been eagerly awaiting the outcomes. Nobody else has the sample size of 23andme. If 400,000 sets of genetic variations are in one place, with smart people studying them, what can come out? It’s the first look at a lot of genotypes at scale, by an agency whose job it is to tease out whether or not there’s a health effect, and if so, if that effect is positive.
That’s a big deal. Whatever the answer, that’s a big deal. Somebody got to look into that particular box, and knows the answer. The whole point of the FDA process is that you have to submit the answer for review.
But here’s the thing, there’s no answer at all yet. 23andme hasn’t called back…for a while.
If the answer is inconclusive, there’s plenty of money to be made fighting it through the normal process. That’s the bread and butter of the biotechnology industry - fighting the FDA over the fine details of efficacy and safety to get devices and drugs onto the market.
The fact that 23andme basically cut off communications in May - and addressed this warning letter on Facebook - is a strong sign that the data are at this point conclusive.
But since 23andme won’t tell anyone, we don’t know which way. This fits into a general pattern of espousing open science while not practicing it for the company. So the data are either really bad, and releasing them would forestall the chance to get more people enrolled and wait for Metcalfe’s Law to kick in, or they’re so good that they’re worth picking a public fight with the federal agency least likely to be disrupted by a startup.
And the early messaging, also reported by Fast Company, doesn’t indicate strength - instead, it appears the first line of response is not “look how awesome our data is” but “the health insurance companies are terrified of us.” I tend to think that Kaiser could drop the $100 per patient and sequence 5.6% of their enrollees for free, within a year, if they were terrified. I think the data aren’t there yet.
I’m a deep believer that genetics generally, and personal sequencing in particular, will drive a marked and permanent change in health. But 23andme are either there today or they aren’t. They dealt the play last summer when they started this process, and advertised it. They’ve upped it by framing themselves as a daring company. But they’re going after the FDA with this strategy. And if you come at the king, you best not miss.
Just a quick note to put out a thought. I’ve done a lot of work with informed consent, and am one of the people pushing the idea that in its current form it isn’t serving our needs as a society, as patients, as scientists, as people. That’s pretty much the whole point of my TED talk. And I’m a supporter of projects like Reg4All, which embrace fair information services practices as their basis. There’s a real role to play for that approach, especially in aggregate level data, or in recruitment of people for studies that will in turn require informed consent.
But I am noticing a disturbing trend in the conversation around consent, one that I intend to combat fiercely, which is the idea that informed consent itself is the problem.
I don’t think it is. I think it’s the way we do it that’s a problem. When it’s used as a shield against patient involvement, or as an excuse to deny people their data, that’s a problem. When people don’t understand the forms and are rushed through the process, that’s a problem. When data can’t be integrated into larger and larger studies because of consent, that’s a problem.
But the idea that we should be informed about the risks and benefits to ourselves and society before we decide to share our data? I don’t think that’s a problem that needs disrupting. We need to embrace better design for informed consent. We need to bring it into the 21st century, to make it compatible with distributed computing, with large-scale mathematics, with the internet. That’s clear.
None of these problems, together or collectively, mean we should treat the issue of consenting people into research as a simple transaction cost. It’s not about reducing the “friction” of “participation” to a few clicks here or there. If we treat informed consent as a transaction cost to be minimized or eliminated, we may solve the problems of data liquidity.
But data liquidity, as we’ve seen in national security, social media, and more, doesn’t always come alongside beneficence, justice, and respect for people. That’s the point of informed consent, not liquidity of data. Health data is one of the only places where informed consent - not just simple clicking - is part of our cultural heritage.
I think that’s a feature, and not a bug. If we can’t find a way to balance informed consent, and its ethical heritage, with data liquidity, that failure is on us.
Today is the White House’s Champions of Change event celebrating Open Science. You can watch and join in at 1PM US Eastern time today.
But on this day, I wanted to remind everyone that “open” is a word that comes with history and, more importantly, with meaning. It’s not a word that someone can take and sprinkle on their work and claim without understanding the history and the meaning. If someone does that, they are going to get bullshit called on them - like the Yale-Medtronic supposedly open data agreement, which is anything but open.
So let’s get this clear. Just because you’re making something available that wasn’t previously available doesn’t qualify as open. Just because you’re reducing the transaction costs of access to something doesn’t qualify as open. Just because you’re involving more people than before doesn’t qualify as open.
Open means that there is no prejudice against any kind of user, anywhere, any time. Open means that commercial use is allowed, in advance. Open means that new works, new research, new products are allowed, in advance. Open means following the Open Definition.
Here’s the Definition. It’s short and sweet.
“A piece of data or content is open if anyone is free to use, reuse, and redistribute it — subject only, at most, to the requirement to attribute and/or share-alike.”
The reason for a definition is so that the word open actually means something. Open is a great word. But it has to mean something for it to mean anything. And it’s hard to meet this definition. It means you can’t be granular, or differential, or non-commercial, or academics only, or only for some kinds of uses in science and not for other kinds of uses. It means a bright line test that a project either meets or doesn’t meet.
The Yale-Medtronic example is a classic case of misappropriation of the word open. It’s anything but. Yet there goes the science establishment trumpeting it as such.
If “open” is going to mean something in science the way it means something in culture and in software, those of us who’ve spent years toiling in the trenches have to collectively examine, carefully, every claim to open science. And we have to speak truth to power, to hold those who would claim openness as their mantle to the definitions and call them out when they fail to meet them.
Being open - really open - is hard to do. It doesn’t give the people what they want. As Bob Young noted in his keynote at this year’s Sage Congress, what people wanted out of Red Hat software was for it to run Microsoft Office.
He didn’t give them that. He gave them what they needed, which was the right and the power to control their own operating systems. And people thought he was crazy for not giving them what they wanted, for not compromising on open, right up until Red Hat was so successful that everyone agreed it had been obvious all along that open was the smart choice.
It wasn’t easy being open in software, then. It isn’t easy now in science. But the value comes over the long term from the decision of a small set of people being unreasonably committed to making true public goods. Today’s event is a small but pivotal step in recognizing those people, and in carrying those true public goods forward.