Healthcare on the Blockchain, Day 2: Drug Management, Machine Learning, & Private vs. Public Blockchains

Day 2 of Insight Exchange Network’s “Healthcare on the Blockchain” event began with sessions on using blockchain technology for secure verification and changing the architecture of consumer decision making (see our coverage of Day 1). Morning speakers included David Houlding, Director of Healthcare Privacy & Security at Intel, returning to moderate a session on blockchain security with Jeremy Grant, Managing Director of Technology Business Strategy at Venable LLP and Debbie Bucci from the Office of Standards and Interoperability at Health and Human Services.

Doug Emery, Director of Benefits Innovation Solutions and the the Center for Value in Healthcare at Altarum, and James Maldonado, CEO of Trestleworks, presented a session which encouraged some new points of view. For example, the speakers discussed evolving from fee-for-service to risk-adjusted systems as the unit of account and transitioning from revenue cycle management to value cycle management through computable, composable payment and benefits contracts.

After a morning networking break, Jennifer Georgino, Contributor to Blockchain Healthcare Review, returned to moderate a session on managing data on the blockchain for cost-savings and error reduction. Panelists for the discussion included Tapio Tolvanen, Co-Founder and CTO of BetterDoctor, Michael Marchant, HIE and Integration Manager at UC Davis Health, and Srini Attili, Life Sciences and Healthcare Principal at Deloitte.

The panel began by highlighting challenges in healthcare that could be addressed by blockchain technology. Michael commented that positive patient identification is the biggest impediment to sharing data between healthcare data silos. Millions of dollars worth of software management tools and personnel time across multiple departments are used on patient identity management today, which could be reduced or replaced by moving this effort to a blockchain. More broadly, Srini commented that blockchain technology is best applied to processes running across multiple stakeholders in an ecosystem. Blockchains have the potential to eliminate the friction points in those systems but not necessarily by replacing the process. Rather, Srini believes the value of blockchains comes from sitting on top of legacy systems to enhance them by improving speed while reducing errors and risk.

Similar to the rubric Nicole Carthcart shared on Day 1, Srini shared Deloitte’s twenty questions to use when determining if blockchains are an appropriate solution to a problem. Two of the questions Srini highlighted were: (1) Does the problem fall within the four walls of an organization? If so, blockchains are likely not the right solution. Too often companies think that because blockchain technology is the shiny new penny, it can solve any technology issue, which is not true. (2) Do you have a network or consortium of stakeholders interested in solving the problem? If not, again blockchains might not be the right solution since a network of participants is required to allow a shared ledger to work. If multiple stakeholders are interested in participating, such as multiple health insurers, questions like “who is going to pay for the technology implementation?” and “how can a proof of concept scale into a sustainable business model?” follow.

The first question led to some interesting back and forth between the audience and the panel regarding where “the four walls of the clinic” exist. While there was consensus on the fact that blockchain can solve challenges within an organization, there were differing opinions on whether existing technology solutions not using a blockchain ultimately achieve the same thing more efficiently.

Before breaking for lunch, Bob Celeste, Founder of the Center for Supple Chain Studies (CSCS), spoke about how blockchain technology can be used for managing drugs. Bob painted a stark picture of the impending challenges facing drug management as a result of the Drug Supply Chain Security Act (DSCSA), a piece of legislation enacted by Congress in 2013. Today, about 59 million packages representing 30 billion individual drug products move around the country each year. The DSCSA mandates that by 2019, all drugs need to be trackable, by all trading partners, back to the manufacturer. Bob believes that what blockchains comports to do can solve many of these challenges, especially in managing transactions between trading partners. As an example, if a drug is produced by a manufacturer, there is a shipping event which transfers ownership to a new organization, like the wholesaler. All the steps that it takes to get the drug from the manufacturer to the wholesaler are important, and need to be documented, but the final state of the drug is what matters.

In order to test the application of blockchain technology to this example on a larger scale, CSCS used virtual simulations. While eighteen different approaches were tried, three that Bob shared were (1) putting all information about the transaction on a blockchain, (2) using the blockchain as a directory pointing back to relevant information, and (3) using a blockchain to monitor the state of the drug product. The first model ran into a challenge of governance which takes more time and money, the second reduced governance but brought up the challenge of avoiding fraud and requiring users to use addresses to lookup information. The last proved most successful while also incorporating smart contracts to implement rules, developed by consensus, around the sixteen transactions it takes to get a drug to a provider. The result is an efficient process management infrastructure that alleviates the time around basic decision making and allows participants in the supply chain, like doctors, to focus on more important things, like patient care.

Following lunch, the event resumed with a session on the relationship between blockchains and the internet of things moderated by Maulik Bhagat, Managing Director at AArete, and featuring panelists Tatyana Kanzavelli, Founder and CEO of the Open Health Network, and Michael Beck, CEO of DxChain. Both panelist’s businesses are building tools using blockchain technology and neither are storing data on-chain. For example, Tatyana spoke about Open Health Network projects where patient users have control over their metadata, which she defined as high-level, non-identifying factors like gender, that they can choose to share with payers, providers, and researchers. Those stakeholders can use their access to this data to engage patients about various programs without needing to know everything about them.

Returning to a recurring question of the event, Michael gave his perspective on when it makes sense to use blockchain technology. His take on the question is that blockchains are incredibly good when decentralizing trust or dealing with a distributed data store where a central custodian does not make sense. Michael made an important point on the fact that traditional data custodians not using blockchain technology have made investments in ensuring data immutability, such as logging, backups, and redundancy, to achieve some of the same goals that blockchain technology promises. Just because these technologies are not new does not mean they are not appropriate for many of the use cases for which blockchain technology is being considered.

Also adding his opinion to the previous discussion regarding using blockchains within the four walls of an organization, Michael commented that businesses who are concerned about the breach of data or who need to track the voracity of that data may find blockchains interesting to confirm the data was not accessed inappropriately. He also suggested that the definition of “four walls” might include an organization’s trading partners who have existing contractual agreements to work together. However, when interacting with stakeholders whom the organization does not have a formal relationship with, blockchain technology to manage sharing begins to make more sense.

During the panel, one audience member asked the question about how to differentiate between the dozen or so companies seeking to tokenize healthcare interactions. Maulik, Michael, and Tatyana agreed that since there is not much distinction between those businesses’ technology today, the decision point comes down to the unique advantages of a given platform. Tatyana stressed the fact that blockchain is a middleware and therefore one way to distinguish between platforms is by identifying those that connect to tools which successfully drive engagement with end users. Michael mentioned the need to distinguish between platforms using intrinsic versus non-intrinsic tokens, or those blockchain systems where tokens are used as an incentive to verify and secure information compared to those used as a crowdfunding tool.

Combining two hot tech topics into one session, the next panel, moderated by Heather Flannery, CEO of Obesity PPM, focused on the intersection of blockchain technology and machine learning (ML). David Houlding, Director of Healthcare Privacy and Security for Intel, returned to the stage for the panel accompanied by Joshua Rubin, Program Officer of the Learning Health Community. From the discussion, two important takeaways related to the relevance of blockchain technology to artificial intelligence (AI) and ML emerged. The first was that AI and ML are “data hungry” technologies. Using data from only one source, or in this case one healthcare organization, results in a specific model developed with a certain degree of inference and, as a result, a certain error rate. Blockchains open up the possibility of allowing AI and ML systems to source data across multiple organizations in order to reduce error rates and avoid ending up with a biased model.

When it comes to medical devices, the FDA allows for updates and patches provided the changes are tested to make sure the originally approved technology still works as intended. The second takeaway was that this same approach should be taken with AI and ML models to determine, at each stage of learning, whether the quality of the model has been preserved. However, models are easy to believe when the output is expected. Real value comes when an updated model outputs a new insight. So how do you know when to trust when the unexpected output is still accurate? Validation of AI and ML results is similarly challenging within a single organization but can be more easily achieved when comparing the results across a network of organizations leading to faster trust of new insights.

Transitioning from moderator to presenter, Heather Flannery closed the second day of “Healthcare on the Blockchain” on a pragmatic note highlighting the ongoing, iterative evolution still taking place as blockchain technology matures. Some of the steps that comprise this cycle include important themes we’ve heard discussed throughout the event, such as educating healthcare professionals to make blockchain more approachable and accepted as a viable technology, creating reference architecture for off-chain data lakes since healthcare data itself may not make sense to put on-chain, and resolving system boundary challenges in the context of cybersecurity so information sharing can safely take place between traditional data silos.

With that, IEN’s blockchain event came to a close. Two days of panels and discussions did a great job engendering accessible conversations on both high-level and granular topics related to blockchain technology. The collective sentiment evoked by presenters was one of opportunity tempered by reality of existing technology and regulatory limitations. We look forward to hearing more about how healthcare technology companies will begin implementing blockchain-driven solutions in the coming years.

Can’t wait to learn more about companies using blockchain technology in healthcare today? Check back for upcoming interviews with the event’s two vendor sponsors: Luxoft and ULedger.

Previous coverage: Healthcare on the Blockchain, Day 1: Tech Primer, Use Cases, & Privacy

Links: Healthcare on the Blockchain and Government Blockchain Leadership Forum…