19 March 2018

Is blockchain a good fit for health?

KnowCents Technology

Just a few short years ago, the technology known as blockchain was really only being talked about in niche technical journals.

These days, it is being promoted as the most disruptive technology since the internet, with the potential to transform many industries, including healthcare.

But few people, when pressed, can really say what the technology is, or explain how it will affect their lives.

Some may recognise blockchain as the technology that underpins the cryptocurrency called bitcoin, but beyond that, it all gets a bit fuzzy.

Put simply, blockchain is a database. Trust in the veracity of that database is created through complex algorithms, enormous computing power, and democratic consensus between multiple parties.

Blockchain is specifically designed to make record keeping as laborious as possible to stop people cheating the system.

It forces its users to solve tricky mathematical problems before they can alter data. Even then, changes will only stick if a large number of people agree that the transaction record is valid.

Blockchain could thus be described as an incredibly slow database, which has lots of barriers in place to stop the quick transfer of information, but also one which engenders great trust in the accuracy of that information.

There’s a great deal of excitement around blockchain applications outside finance, but it’s not entirely clear if this technology, which operates on the principle of radical transparency, could be of any use in a healthcare context.

The public demands absolute confidentiality in health. And, in circumstances where information is shared between trusted providers, data must move instantaneously, particularly during a medical emergency.

Every computer using the blockchain has a full copy of the data, which arguably makes sensitive information held on the blockchain less secure than if it was stored elsewhere.

It’s the perfect technology for a currency such as bitcoin, which may partly explain why the recent rapid increase in the value of the cryptocurrency. For example, it is impossible to spend bitcoin twice without someone noticing. Every bitcoin transaction is scrupulously documented and copied onto thousands of computers worldwide.

But the proof of work needed to unlock and edit a blockchain is a very high overhead to pay to record health information when a mix of cloud encryption, software, and centralised databases can essentially do the same job.

It might not seem like a perfect fit at first sight, but interest in using blockchain in healthcare has never been higher.

Out of every sector investigating applications for blockchain, health is by far the most active.

IBM estimates that around 16% of healthcare executives globally are experimenting with blockchain technology, which is a higher proportion than in finance and banking.

Startups are emerging in every field of health, including genomics, electronic medical records, medical registration and real-time prescription monitoring.

The government of Estonia has employed the company, Guardtime, to create a blockchain for its national medical records.

Even the Australian Digital Health Agency sees blockchain as a promising development while acknowledging the immaturity of the technology.

There must be something about blockchain that is very different to existing technology, and which makes entrepreneurs and governments overlook its current limitations.

So what is so special about blockchain?

Joining the conga line

For starters, jumping on the blockchain “hype wagon” comes with major financial and public relations benefits.

More than a few companies are riding the blockchain wave by incorporating this new buzzword into their branding. The silliest example was the Long Island Iced Tea Corporation, which re-labelled itself as Long Blockchain Corporation last year and saw an almost 300% bump in its share price in one day.

The current blockchain frenzy is “the most insane bubble in history”, mirroring the dotcom boom of the late 1990s, says Dalibor Frtunik, the Australian head of digital health record company Sorsix.

While the hype lasts, everything blockchain touches turns to gold. And it seems that nothing is out of its grasp. “Blockchain journalism” and the use of blockchain for gun control are two of the weirder examples.

Frtunik is not convinced that blockchains are superior to centralised databases in the health sector. “The idea that the blockchain makes things permanently secure is bullshit branding and not true,” he says.

Myths around blockchain are less common now. But at the start of the craze, some companies made up reasons to do blockchain-based projects because they wanted to be seen as thought leaders, says Mark Staples, a principal researcher at CSIRO’s Data61, a data science joint venture between CSIRO and GE.

Every technology comes with trade-offs. Blockchain is typically slower than conventional computing, and it can’t store big data, Staples says.

Blockchain creates multiple copies of information, which gives data thieves options. The network effect means that first users of blockchains end up paying more and getting less benefit than late starters.

The latest iterations of blockchain address some of these drawbacks. But technology can only do so much. The biggest challenge in health is actually determining what level of privacy people should be entitled to, Staples says. Blockchain can’t fill a policy void.

Point of difference 

Blockchain is functionally quite similar to existing technology, but it has unique non-functional properties, according to Staples.

Traditional digital platforms depend on a single company or a government department to maintain a centralised database. Blockchain, by contrast, is operated by a collective.

“That’s the main benefit, or the reason why people are using blockchain,” Staples says.

By sharing digital infrastructure, no single actor can compromise the integrity of the system or act as a data bottleneck. Each party has an equal say in how the platform operates.

Like a digital game of chess, the rules of engagement are written in code and cannot be broken. Every node on the network has its own copy of the data, making it hard to delete information or commit fraud.

Blockchain believers argue that this trust-by-design promotes transactions that otherwise might not occur, such as the sharing of health data currently trapped in silos.

It’s these features of blockchain that make it appealing for administrative purposes.

It takes on average about 45 to 60 days for a hospital to credential a physician in the US. This cumbersome procedure is repeated every two years.

“So this is a never-ending process of constantly reviewing this data,” says Corey Todaro, the COO of Hashed Health, a blockchain consortium based in Nashville.

This year, Hashed Health is planning to speed up this process using a blockchain-based provider credentialling exchange platform.

“What the blockchain does is provide a guarantee that the artefact comes from the source that it claims to come from, and that it has not been altered in any way,” Todaro says.

Instead of each hospital having to get an independent verbal confirmation that a doctor has graduated from a certain university or has worked at a particular clinic, they can use consensus verification. The blockchain records every credentialling event, so somewhere down the line a hospital is going to see that a set of educational history has been independently verified 16 times by other hospitals, and can decide whether they really need to call again, Todaro explains.

“That’s a decision of the institution, but we can provide a platform where someone can rely on the collective performance of other constituents in the network.”

Blockchain isn’t a panacea, but it does allow multiple parties to verify information, without creating a centralised database, and that is something new.

Data police

Estonia’s use of blockchain for its national health records went live in 2012, making the small eastern European nation the earliest, and (as far as we know) only adopter of the technology for such a purpose.

The rationale is simple, says Marten Kaevats, the national digital adviser at the Government Office of Estonia. The whole information ecosystem relies on trust and blockchain provides trust services.

The blockchain doesn’t secure the data itself; it works like a speed camera to catch people breaking the law. Health data is stored off chain, but changes to it are tracked through a series of hash values. Hash values act like fingerprints as they leave a unique mark automatically each time the data is altered.

Instead of minor data tampering going unnoticed, red flags go up in seconds. The system detects unauthorised use, and automatically alerts users to any unusual activity.

All citizens in Estonia can go to a government webportal and see who has been looking at their health data and why. And, yes, it is cheap, Kaevats adds.

This is a minimalistic approach to blockchain design. Other groups see blockchain taking on a much bigger role in securing the health data itself, rather than just policing data breaches.

Robert Chu, a member of Linux’s open source blockchain community, Hyperledger, is championing a hybrid architecture, where data-light electronic medical records (such as diagnoses or prescriptions) are stored on the blockchain, while larger files, such as CT scans or X-rays, remain off chain.

His New York-based company Embleema aims to reconstruct an individual’s medical history by bringing together fragmented data from different sources.

Once the data is captured by the blockchain, healthcare organisations can hand complete control of data back to the individual, who can use passwords, smart contracts and encryption to regulate which parties can do what with their health data, and exactly when that access expires.

For example, a patient could authorise read-only access for three minutes, allow access for a particular observational study, or choose not to share parts of their medical history.

Certainly someone with malicious intent can still take a photograph of the data or steal a password, but blockchain gives patients much more fine-tuned control over who sees their data in normal circumstances than traditional health records, Chu argues.

Smart contracts, or the ability to switch on and off a range of personal privacy protections at will, are a central selling point for blockchains in health.

But, as with all new technology, we will have to wait for blockchain to be tested before these security benefits can be validated.

The initial findings aren’t promising, however; a recent study by Singapore researchers found that around 3.4% of smart contracts on Ethereum were vulnerable to attacks.

Storing a copy of the health record at every node makes it close to impossible to compromise the data’s integrity, because someone would have to meddle with more than half of the nodes for any changes to be authorised, Chu says.

Duplication of data does create vulnerabilities, however: “The probability you can be hacked is multiplied by the number of nodes,” Chu adds.

It’s for this reason that medical doctor and entrepreneur Dr Tal Rapke is storing all private prescription data off the blockchain for his Australian start-up company ScalaMed.

His company, which intends to create smart, paperless prescriptions using blockchain, wants to avoid creating a “honeypot of data” by storing health data in a decentralised database, Dr Rapke says.

Dollars in data 

Once patients have total control over their health records, they become the rightful owners of a very valuable resource: data.

A new marketplace emerges, and this is where blockchain really comes to life.

A Florida-based company, called Encrypgen, is one of a number of start-up companies  building blockchains that will allow individuals to trade their genomic data in exchange for cryptocurrency. Genomic data is stored off the blockchain, but data custodians can choose to upload certain metadata (such as gender, ethnicity, age, and the brand of genomic test) to the blockchain to create a secure, searchable online shopping platform.

Research organisations, pharmaceutical companies and other entities must gain permission to access the system and pay for that access in by helping maintain the blockchain, and by handing over cryptocurrency for each genome purchased.

But is this simply using new technology for the sake of it? Why not create an Amazon for genomics, supported by Paypal, and bypass the blockchain altogether?

It would be possible to create such a platform, but there are quite compelling reasons to use blockchain, according to David Koepsell, the CEO of Encrypgen.

Paypal can take up to a day to complete transactions, whereas a blockchain hosted by the Ethereum platform takes just seven minutes.

Using a unique cryptocurrency makes transactions cheaper and removes the need for foreign currency exchange, he says.

It’s true that the blockchain does not secure genomics information stored off chain, but blockchain transactions do hide the location of that data, or the IP address, which cloud services wouldn’t do.

Data theft is much harder when hackers don’t know where the data is being held,   Koepsell says.

Moreover, Amazon, Yahoo and Google cloud services are major targets for hackers, so simply avoiding these services improves security, he argues.

The marketplace will determine the value of a single human genome, but the price will probably settle at around a few thousand dollars, according to Koepsell.

Some populations will have much more valuable DNA, however. Koepsell says that part of the inspiration for Encrypgen was the story of Henrietta Lacks, a woman who gave the first immortal cell line to science and died in poverty.

“We want to create a future where the next Henrietta Lacks is rewarded for their contribution to science more directly,”   Koepsell says.

Some big players are starting to show real interest in blockchain; IBM Watson Health recently announced its partnership with the US Food and Drug Administration on a blockchain project involving oncology-related data from electronic medical records, genomics, clinical trials and wearables.

It’s easy to see the appeal of blockchain in health. By acting as an independent source of truth and by generating binding codes of conduct, blockchain seems to constrain human behaviours that degrade trust, including lying, stealing, cheating and making mistakes.

Lack of trust is a major obstacle to the collaborative use of health data.

So, should we all be looking to sprinkle some blockchain magic into our health services? Or is blockchain, as Frtunik puts it, a “solution without a problem”? At this early stage, it’s anyone’s guess.