Two new blockchain and healthcare business networks are launched
The Pharmaceutical Utility Network and Bayer's second federated learning initiative were revealed this week.
|Robert Miller||Feb 2|| 3|
Blockchain and healthcare updates
KPMG, Merck, Walmart, and IBM formed one of the pilots in the FDA’s DSCSA program. This group completed their pilot this past December and instead of disbanding they are continuing their collaboration as a new business network: the Pharmaceutical Utility Network (“PhUN”). There wasn’t a formal announcement for PhUN, which is a bit odd, but I noticed the above page on IBM’s website. The only known participants are the four involved in the DSCSA pilot. IBM lists a few supply chain use cases and the general category of clinical trials as areas of interest, it’s what you would expect a pharma consortium to focus on. I thought their description of PhUN’s approach was interesting:
PhUN takes an open source platform-first approach — a similar concept to the Apple App Store — to empower stakeholders to comply with regulations and to use them as the basis for innovation. At scale, PhUN will integrate regulatory requirements such as the DSCSA into the platform, allowing solution developers — pharmaceutical manufacturers, distributors, dispensers, software vendors and others — to comply with ease.
PhUN’s platform-first approach is in contrast with other business networks that have (initially) focused on a single use case, like the Synaptic Health Alliance. The Health Utility Network, another IBM convened business network that is payer and provider focused, is also platform-first. Notably they seemed to be figuring out their use cases half a year after their initial announcement and we still don’t know what they are focused on. It’s more difficult to reach consensus between members on what to do if all of those parties don’t join with the same vision for the consortium and the value it provides.
The focus on regulatory compliance is consistent with a platform-first approach. Their bet is that regulatory compliance is a capability, not a use case, and by integrating that into their platform 3rd party developers will be able to focus on adding value in ways other than regulatory compliance. It’s an interesting idea, but one that most use-case first consortium are also thinking about, so this likely will not be a unique differentiator.
Lastly, in the coming months I expect to see other organizations talking about the results of their DSCSA pilot, and eventually a final report from the FDA as well. Perhaps more interesting, if project had successful pilots they will almost certainly continue on as consortium or as solution providers. So, there will be a lot of competition in the supply chain space this year and we’ll get to learn a lot about the dynamics of the market and what the best approaches to blockchain solutions are.
You can find roughly the same content non-paywalled here, but it is in German so you’ll need to use a translator.
Bayer has two major federated learning initiatives: MELLODDY and a new collaborating with seven hospitals and research organizations. MELLODDY is the consortium using blockchain and federated learning to enable pharma companies to collaborate on drug discovery. The new collaboration is training an algorithm to identify a rare mutation in cancer, which would inform treatment of that cancer.
Rare diseases are a place where federated learning can add unique value. Each individual organization will have too few patients with a given rare disease to make any sort of statistically significant analysis. But several organizations combined will have enough data, and federated learning makes it easy for them to collaborate because no underlying data is shared. This is one reason why federated learning networks will have strong network effects.
Moreover, in a federated learning network there is a need for some kind of orchestration to decide what data to use and what algorithm to train. If there are multiple data sources it is problematic for one party to control this, which is why group usages of federated learning like MELLODDY and KCL’s leverage a blockchain, but the Mayo Clinic’s standalone usage does not.
The individual predictions are a bit long to list here, but a few people got together under Blockchain and Healthcare Today and published their predictions for 2020.
EVENT: Blockchain and Digital Transformation in Health 2020 (February 26th in Austin, Texas)
If you’re looking to hear from industry practitioners and researchers then this blockchain and healthcare event is a great fit. A number of my colleagues and friends in the space will be speakers, and I'd encourage you to check it out.
What I’m reading this weekend
Clemens has been involved in a lot of real world deployments of blockchain, first at R3 and now at ConsenSys, so he has a lot of insight into what the barriers to production are. The list is focused on finance but it’s a short and good read.
I found this to be a relatively approachable way of thinking about the economics of drug development and the various levers that policy makers have.
How to properly fund and provision public goods is a deeply important problem with far ranging implications. One place where it’s easy to see this in the blockchain space are developer’s contribution to blockchain protocols. These highly specialized individuals provide huge value to their communities but aren’t paid and often can’t get funding to continue their contributions. Some blockchain communities have tried to solve this by dedicating a % of all coins mined to a fund for developers, but there is a lot of disagreement over whether this is a sustainable or desirable solution.
A novel philosophy for how to organize decentralized ecosystems of public goods called Liberal Radicalism was proposed by Vitalik, Zoe Hitzig, and Glen Weyl and has since then been implemented by Gitcoin in a grants program. The whole thing is fascinating and Vitalik had a great review. This is a space I’m watching and one I expect to have insights applicable outside of the blockchain space too.
The Aave protocol provides something called “flash loans,” or loans that only exist for as long as the transaction in which they are issued. Put another way within a single transaction a loan will be given and closed out. That’s useful if you need some capital to execute an arbitrage opportunity and it’s a cool demonstration of the sort of thing that is only possible with programmable money.
If you found this newsletter valuable then you can click the button below to sign up for free.
If you’re an existing reader I would deeply appreciate it if you share this with people who would find it a valuable resource. You can share this piece on LinkedIn here, and there is a button for sharing on Twitter below. You can also “like” this newsletter by clicking the heart just below this, which helps me get visibility on SubStack.
My content is free, but if you would like to support me, you can do so on Patreon.