What is the healthcare blockchain but a dream within a dream?

November 7, 2016
Luke Turner Staff Software Engineer

“Blockchain” has become a familiar buzzword in healthcare IT. While I think the interest in this amazing new technology is fantastic, I also think that some sense of perspective has been lost in the process of gaining recognition.

The blockchain design pattern is a great idea, but technology exists to solve problems, and the blockchain pattern is a low-level technology that requires a lot of implementation-specific care to actually do that.

Defining a system by its use of a blockchain is like defining a system by its use of a SQL database—it’s not wrong, but it encapsulates a forest of different possibilities, each of which has its own quirks, variations, guarantees, and requirements.

Why do people keep talking about it?

Everyone has their reasons, but the fact is that the whole idea of the blockchain is relatively new, and (let’s be honest) extremely exciting. The spark of genius and broad applications of the idea make it easy to hype, but simply getting excited about the prospective achievements of the blockchain won’t usher in a new era of actually using it in healthcare.

Instead, we need realistic expectations, steady experimentation, and some breakthrough insights to crack open the tough, outstanding problems that dim the promise of decentralized computing. I might not be able to help with the paradigm-shifting insight—not in this post, anyway—but I can at least take a shot at clarifying some expectations.

For starters, know that the true power a blockchain offers over traditional databases (like Oracle or what have you) is that a blockchain is designed to enable a decentralized, peer-to-peer architecture.

This is probably the most important characteristic of a blockchain, because decentralized architectures are highly desirable: they have no individual failure points, and even if parts of the network are separated, they can operate independently and recombine when the network is healed.

Also, they are typically harder to attack: if the system is large enough, it becomes very difficult for malicious actors to damage it permanently.

Remember, each blockchain system is different, and they serve different markets and offer different guarantees. For example, blockchain systems can be public like Bitcoin, or they can require authorization, like a traditional database. They can have all peers be equal, or they can have some “super” peers that have extra control. They can use smart contracts and privacy-preserving computations—or not.

Each new configuration has to be evaluated on its own merits. The field is ripe for experimentation, but let the buyer beware—the technology does come with some faults. When it comes to health IT, blockchain faces some special difficulties in comparison to other markets. For example, the financial tech industry (fintech) is a good place for decentralized technology to break in, because the underlying account data is relatively simple.

That’s not to say fintech is easy, but Bitcoin, Altcoin, Litecoin, and others all take advantage of the simplicity of the underlying transactions. Basically, you can add and subtract. That’s it. This makes validation simple, because all you really need to do is compare the account balances and the added and subtracted amounts and make sure they match up. These transactions also have another great property, which is that they’re commutative—you can record your transactions in any order and still get the same result. This is a huge simplifying characteristic.

In contrast, medical data validation is difficult because the data is complex and doesn’t relate to itself well.

Say a new diagnosis is put on a patient’s hypothetical blockchain-chart. Is that transaction valid? Is it consistent with the rest of the chart? This question may have no certain answer, but every user of the blockchain needs to be able to decide it accurately in order to either reject or accept the transactions.

Not only does validation pose a problem, but healthcare workflows are complicated and often have real-time and in-order requirements that limit options for database processing. No matter how you look at it, healthcare data validation is just hard, and using a blockchain is only going to shine a spotlight on this problem.

We do have solutions, but unfortunately they often result in some degree of re-centralization of control, which diminishes the overall benefit of the approach.

Bitcoin’s novel proof-of-work idea—requiring difficult computations in order to do validation, so that a malicious actor would need tons of computational power in order to invalidate existing data—just doesn’t make sense in most other contexts, including healthcare.

How are my transactions validated?

This is probably the most important question that needs to be answered by any new blockchain system, and we just don’t have a good answer for that yet in healthcare.

Another problematic aspect of blockchains is that they can be unruly in ways we aren’t used to dealing with. This is a common property of “self-healing” or “self-managing” systems.

Each client usually has a copy of the chain, and it may not be in-sync with the others. Transactions may take seconds or minutes to get where they need to go. The chain itself might get too big, since it contains all historical transactions.

Exercising control in the system can become challenging, because there is no “super-user” with admin access to do whatever he wants, and updates are decided with a democratic process, automatically, among machines.

From a business perspective, this can be scary, because it may be difficult to undo mistakes. But the fact that these types of administrative actions are so hard is part of what makes a blockchain secure, so one way or the other, a tradeoff has to be made: do we reduce security, or do we sacrifice control?

The hurdles for blockchain in healthcare are all about the brass tacks. The general concept is solid, but as soon as you roll up your sleeves and start to apply it, you realize how quickly the theoretical beauty can be weakened or even completely compromised by real-world considerations.

The idea of the blockchain is so powerful and dynamic that I think it can evolve to transcend this dilemma.

What I suspect will happen next is a period of experimentation with different types of blockchain solutions. During this phase, I hope our industry will hit on powerful ideas for solving the hard problems of authentication and validation in new ways that are unique to the healthcare space. Then we’ll probably need to wait for the business side of the industry to accept these innovations, which will have a very different architectural profile with a totally new set of risks and rewards.

And then, finally, the true dream of blockchains in healthcare might be realized.

Check out our webinar on Blockchain by clicking on the banner below!

Blockchain and healthcare: Is this the future?

Stay in the know! Subscribe to our newsletter.