Many health and life sciences organizations are now bought into the concept of blockchain and are currently researching suitable use cases to implement this technology. Below I share five key characteristics I am seeing in successful blockchain implementations of use cases.
1. It's About the Network, Not a Database
Blockchain is a type of B2B middleware. It is not a replacement for a database within a given organization. It is relatively quick to come up with a blockchain use case. It takes longer to come up with a business network around that use case. That is why near-term opportunities with blockchain center around existing B2B networks, for example, clearinghouses, or drug supply chains, etc. Longer term, new and revolutionary use cases for blockchain will emerge, with new types of participants and networks, but these will take longer to establish. Existing B2B networks may have an intermediary. Rather than disintermediate the intermediary what is more likely to occur is for blockchain to change the role of the intermediary from a potential bottleneck of transactions and single point of failure to more of an enablement, facilitator, and support role.
2. Clear Definition of Use Cases
An essential ingredient for success with blockchain is a clear definition of use cases, participants in the B2B network, and value props to all stakeholders. In contrast, an approach of "let's put all the data on blockchain and figure out later how to use it" is likely to fail for many reasons, including a lack of business buy-in, privacy and security concerns, regulatory/data protection law/compliance requirements, and potential performance and scalability challenges for large datasets. A lot of the value props asserted for blockchain tend to be idealistic. Businesses need to have real business value props and ROI in order to incentivize and motivate their participation in sourcing and consuming transactions to/from a blockchain. These incentives, motivations, and ROI are already established and well quantified in existing B2B networks which is again why existing B2B networks are better near-term opportunities for blockchain. Regardless, business teams need to be engaged early in blockchain initiatives.
3. Throughput and Scalability
Blockchains typically have throughput in the hundreds, or thousands, of transactions per second range, depending on particulars of their implementation, including in particular the consensus algorithm. When analyzing potential use cases for blockchain, do a "back of the envelope" estimate of peak and sustained transaction throughput for your use case and see if your use case estimated throughput is in the right order of magnitude range for blockchain. For example, if you require millions of transactions per second then blockchain may not be the best option. Sometimes techniques such as queuing transactions, expressing urgent and high priority transactions through the queue, and batching transactions into blocks can be used to increase throughput. Sometimes, as in the case of batching, this can trade-off latency for higher throughput.
4. Minimal But Sufficient Data
Data put on the blockchain "B2B middleware" between organizations in a B2B network goes outside of the perimeter of your healthcare organization and is exposed to new risks to its confidentiality, integrity, and availability. Data on blockchains may also encounter new regulatory or data protection law compliance requirements depending on the type of data put on the blockchain and locations of the blockchain network nodes, or endpoints where the decentralized ledgers and data live. Large data types e.g. large images, genomes, etc could have a considerable impact on the performance of your blockchain and therefore are generally best left off-chain, referenced by pointers and hashcodes (of the off-chain data to protect integrity) placed on the blockchain. These concerns can have major impacts on the use cases and data on the blockchain and therefore can't be dealt with as a "band-aid" after the blockchain is implemented. Privacy, security, and compliance teams need to be engaged early in blockchain initiatives. See my earlier blog Healthcare Blockchain: What Goes On Chain Stays on Chain for more on this subject.
5. Stable Data
Blocks in blockchain are immutable. That means that once a block is written it (including the data within it) stays as is for the life of that blockchain. It cannot be changed or deleted. This is one of the foundational characteristics of blockchain. It protects the integrity of the data on the blockchain and is a key characteristic of blockchain that helps build trust in a blockchain network. If your use case involves data that changes frequently or may need to be deleted, then consider whether blockchain is the right solution for the use case. "Amendments" can be done to data on a blockchain by adding a new block to the blockchain with amended data that effectively supersedes the original. Similarly, "deleting" a block can be done by adding a new block to the blockchain that marks an earlier block as "deleted". However, in both of these cases, the old blocks continue to live as-is earlier in the blockchain.
Regardless, a use case that involves data that changes or needs to be deleted frequently is going to be a very chatty blockchain and this can have performance, throughput, and scalability ramifications. If you are subject to "patients right to be forgotten" type data protection laws and requirements, this may compel you to put some specific types of data off-chain including PII (personally identifiable information) so that in the event that a patient on the blockchain needs to be forgotten, the PII off-chain can be deleted and the opaque, unique identifier, or pointers to the "forgotten" patients PII that remain on-chain (and cannot be changed) is effectively de-identified and anonymized.
What kinds of challenges and best practices are you seeing for picking and developing use cases for blockchain in healthcare?