Last week I left you with a link to a sane analysis of blockchain’s applications from The Correspondent. It wasn’t about healthcare or medicine, but rather something to tone down the hype about it. I found the article insightful, so I started exploring a little, specifically how blockchain can be used in healthcare.
With that said, I already wrote a little about blockchain in the 3rd issue, but it was one-sided and short. This time I want to make more expansive arguments for and against its use.
What and how?
First, let’s just look at what blockchain is and how it works, as many are familiar with the term but not with the concept. Blockchain is exactly what the word tells us, a chain of blocks. Every block contains some data, its hash and the hash of the previous block. Okay, what?
The data that’s inside the block is variable on the type of blockchain. For example, Bitcoin, implementation of blockchain in the form of cryptocurrency, contains data of the sender and the recipient of a certain amount of money. In healthcare, this data might be (as you’ll see later) genetic information, logs or actual patient information.
A hash is a block’s unique code that’s based on the data inside it, which is a form of cryptography. This also means that any changes made to the block will reflect in the change of the block’s hash. But every block also contains the hash of the previous block, which forms the “chain”. If we wanted to change the contents of one block, we would effectively have to change the whole blockchain for it to be valid.
The other feature that makes it secure is that all of this is stored on a public ledger. Think of a giant chat that every user has access to (has a copy of the whole blockchain). It follows that every transaction, change or addition can be seen by everyone in the chat. This prevents it from being centralised by a single authority (for example, a bank).
Okay, but in theory, you could take a powerful computer, change a block and change all the following blocks accordingly. This is prevented by a concept called proof-of-work, which slows down the creation of new blocks by having to solve a complex problem (that has 2^256 possible solutions). If you change one block, you’d have to solve all the problems of the following blocks once again. Not a superb idea.
For illustration, think of gold…someone has to put in work to mine it and the proof of this work is pure gold. In blockchain’s case, the proof is that a computer solves the problem. In Bitcoin’s case, anyone that solves the problem gets a reward in form of Bitcoin, which is currently at 6.25BTC (gets smaller approximately every 4 years).
For this issue, what you read is more than enough. But the details are in this video:
Why in healthcare?
I know that wasn’t in any way connected to healthcare. But it’s crucial to understand its basics to understand its applications.
Patient data safety
Quite a few applications of blockchain technology have popped up in healthcare, number one being securing patient data. Perhaps this is best explained in the example of Estonia. I didn’t know this, but they have one of the most digitally advanced systems in government and also in healthcare.
How did they apply it to healthcare? They set up an additional layer of security using blockchain. This does nothing else than "record and timestamp each instance of access or each change to a patient’s electronic records”. In other words, they have complete overview and control over what’s going on with the patients’ files. Anytime they detect a data breach, administrators can act immediately to limit the damage and secure the information. Another benefit is that this ensures that the most recent record is always used.
We see that they don’t secure patient records directly with blockchain, but the information about changes and accesses to the database. Read more in this article.
But some companies and startups are developing similar platforms. BurstIQ is focused on data management, primarily on how it’s owned, shared and trusted on a larger scale. Factom, on the other hand, is developing a system to safely store data. Not just in healthcare, but also the mortgage industry and the Internet of Things.
Simplifying patient care
This might be a bold statement since the technology behind simplifying patient care is not the simplest. Usually, different companies are approaching this by merging all the patient data under one hood.
For example, Coral Health puts patient data into a ledger that connects different stages of care - from doctors, pharmacists, public health authorities to scientists. One of their big advantages is that their system automates administrative processes. Another example is Robomed that gathers patient information from various sources (wearables etc.) and securely shares it with health providers.
Supply chains and drug safety
This is another interesting area where blockchain might be applicable. I briefly mentioned in issue #22 how Dell executives see delivery, storage and usage of medical material an administrative nightmare. No hospital in the world probably knows exactly how much material it has, uses and needs. Blockchain might be a solution.
Furthermore, it could make production, transport and usage of pharmaceuticals more transparent. This is a great way to put it:
Once a ledger for a drug is created, it will mark the point of origin (i.e. a laboratory). The ledger will then continue to record data every step of the way, including who handled it and where it has been until it reaches the consumer. The process can even monitor labour costs and waste emissions. - BuiltIn
As DNA sequencing is becoming cheaper and cheaper it’s useful to think about how to safely store all of this data. If you think about it, in an ideal world, you would have the DNA information of many people without being able to connect it with their identity. In this way, research groups would have access to all the data, but be unable to reveal patient’s identity - simply because it would take too much time to reverse-engineer the hash that was generated by the blockchain. In other words, it would be a pseudonym for the patient. Not ideal, but better than otherwise.
This also has the potential to solve one of the biggest problems AI faces today (briefly touched on in issue #21). We already collected huge amounts of data, we just aren’t as good at using it. Blockchain may be another solution to streamlining this and driving research in various medical fields forward.
More about different companies applying blockchain to healthcare in this article.
Why it might not be such a good idea?
Until now, everything looks great. You might even think of blockchain as the saviour to all the problems we face today. Not so fast. A great quote that gives you an idea why that’s so is:
“You could also use a forklift to put a six-pack of beer on your kitchen counter. But it’s just not very efficient.” - The Correspondent
This is especially the case in “simplifying” patient care. Do we need a blockchain to integrate all patient records and securely access them? I doubt it.
As discussed earlier, one of the main benefits of blockchain technology is that it’s transparent. Everyone can see what’s going on at any time, which can exclude (or help detect) and wrongdoings. However, when companies and other institutions adopt blockchain, they sometimes don’t use a public, but rather a private blockchain. In other words, they are then writing and verifying each transaction or change in their system. This doesn’t make it decentralised nor public. It reminds us more of the system we have today. Read this article for more.
As mentioned briefly in blockchain in genomics, patients aren’t anonymous, but are pseudonymous, since they are linked to a hash. It’s extremely unlikely that someone would be able to reverse-engineer this and link it to a patient, but the danger is still there.
Another interesting argument comes from Bitcoin and Ethereum (another cryptocurrency). Remember when I was explaining the “proof-of-work” concept? Well, you need a huge amount of energy for this. Both networks combined “use the same amount of energy as the whole of Austria”. This usage is also only going to grow as “mining” cryptocurrency is getting harder and harder. We should ask ourselves if it’s worth using more energy to solve a given problem. Especially, if it could be solved differently.
There’s also the issue of hype. When Bitcoin rose to its all-time high, people were in a frenzy. It seemed as if anyone could make their solution for a “problem” in form of a blockchain or cryptocurrency. Sometimes this problem is completely made up. For example, what value exactly does DogeCoin bring to the world? Yes, I’m also perplexed.
The same goes for healthcare. Many times, blockchain is used because it’s a buzzword. And many times it’s considered the solution for everything.
So, does blockchain have a future in healthcare? Probably, but with good measure.
Links to keep an eye on
Paper: Machine learning models to predict atherosclerotic cardiovascular disease risk in multiethnic, real-world populations. The authors developed a machine learning (ML) model to assess if it can detect atherosclerosis in a broad and diverse population. We know that it can predict it in patients that show some signs, but not if we apply the ML model to a broader population. Overall, they showed it’s possible and comparable.
Article: Contact tracing apps have a major design flaw. An article discussing the weaknesses of COVID-19 contact tracing apps. There are some legitimate arguments as to why they may not be as effective as we think. I agree, the app by itself cannot end the pandemic. But an app that’s installed on much of the phones in a population, together with appropriate measures, can make it a lot easier.
Online Event: e-Estonia Digital Discussion: at the front line of digital health. This is an event I came across when researching for this newsletter issue. I’ll be covering it in detail in issue #36 and will also make my raw notes available. Just as in issue #21.