Part I: If Blockchain is a hammer, my world is made of nails
From the moment I heard of Bitcoin I was fascinated. The concept of a digital currency struck me as the next logical step in humankind’s progress. When I discovered that Bitcoin is enabled by an underlying technology called “blockchain” which has potential applications everywhere- I was all in.
However, shiny, new tech is ultimately measured in providing enduring solutions to big problems. And product competition shreds away idealism, leaving either the steel thread of a great product or a failed solution. Electronic Health Records (EHR) stands out as a central, problematic element in modern health care. In spite of regulatory mandates and the medical community’s calls, EHR seems to resist all rational product efforts.
Looking at the world through my blockchain tinted glasses, EHR seemed the perfect test. So I asked myself, “Does blockchain offer a natural, problem-solution fit for medical records? Is there a solid argument for a EHR based upon blockchain?”
Surprisingly, in spite of my unbridled optimism, I discovered some compelling reasons that blockchain merits serious consideration for Electronic Health Records.
Electronic Health Records: Lingua Franca or Tower of Babel?
In an information based healthcare system, EHRs occupy a central role. Borrowing from the disciplines of ecology and biology, EHRs are analogous to what are known as a “keystone species”. The defining attribute of a keystone species is occupying a unique ecosystem niche that “holds it all together.” Without the keystone species, the entire food web as well as the physical environment collapses.
Similarly, our individual EHRs role in the growing digital ecosystem of modern healthcare are focal points that all stakeholders rely upon. (see Figure 1.) Virtually all medical services utilize EHRs for services delivery, diagnosis, prescription, procedure assignment, billing and payment cycles, population and medical research, and compliance. Without data from EHRs, the system grinds to a halt.
Figure 1. Electronic Health Records Occupy a Key Role in the Health Services Ecosystem
Source: Chris Montaño
However, there is deep dissatisfaction among physicians and hospital executives regarding EHRs. This has reached a point where the AMA is taking an active role in shaping EHRs by issuing a framework of 8 Usability Challenges that it called upon EHR stakeholders to address. (See Table 1.)
AMA Usability Challenges
|Enhance physicians’ ability to provide high-quality patient care.||Poor EHR design gets in the way of face-to-face interaction with patients. EHRs should be designed to enable physician-patient engagement.|
|Support team-based care.||EHR systems instead should be designed to maximize each person’s productivity in accordance with state licensure laws and allow physicians to delegate tasks as appropriate.|
|Promote care coordination.||EHR systems need to automatically track referrals, consultations, orders and labs so physicians easily can follow the patient’s progression throughout their care.|
|Offer product modularity and configurability.||Few EHR systems are built to accommodate physicians’ practice patterns and work flows, which vary depending on size, specialty and setting.|
|Reduce cognitive work load.||Many physicians say that the quality of the clinical narrative in paper charts is more succinct and reflective of the pertinent clinical information. EHRs need to support medical decision-making with concise, context-sensitive real-time data.|
|Promote interoperability and data exchange.||The EHR should be a coherent longitudinal patient record that is built from various sources and can be accessed in real time.|
|Facilitate digital patient engagement.||Most EHR systems are not designed to support digital patient engagement.|
|Expedite user input into product design and post-implementation feedback.||The meaningful use program requires physicians to use certified EHR technology, but many of these products have performed poorly in real-world practice settings.|
Table 1. Physicians “Must Haves” for EHRs (Edited)
Source: AMA , Chris Montaño
Another looming issue is patient safety. As our medical system has become more complicated, patient safety is a growing problem that cannot be ignored. A recent study published by Johns Hopkins in the BMJ finds that medical error has now become the 3rd leading cause of death (see Figure 2.) When originally conceived, EHRs were seen as a means of preventing errors. Unfortunately, the lack of interoperability of the fragmented EHR market combined with a confusing coding system implemented differently by vendors has devolved into a rising concern. The Center for Disease control recently issued a report calling on Laboratory Professionals to wade into the EHR interoperability issues in order to address patient safety concerns.
Figure 2. Medical Errors are now the 3rd Largest Cause of Death
What’s Ailing EHRs?
The vision of an interoperable, HIPPA compliant, elegantly permissioned, highly usable and complete health record remains aspirational. Data are ultimately models of reality. And the truth reflected by data within EHRs is that modern healthcare is complicated and messy at the services, economic and regulatory levels. Health data captured in an EHR is often subjective, not collected or recorded in any uniform manner, and reflecting idiosyncratic workflows of individual caregivers. Health care visits are episodic, non-uniform and baseline information is often recorded during health crises rather than periodically. Thus far, no broad effort to contain the “wildness” of unstructured health data into a standardized model has been successful.
The dynamics within the EHR competitive landscape don’t offer any hope that “EHR kumbaya” with break out anytime in the near future. The market is lucrative, growing, crowded, and hyper-competitive. It is a fierce scramble for market share in an informational franchise on each individual’s complete medical history. If personal information is the new digital currency, EHRs are potentially a “printing press” offering multigenerational wealth to the company that captures this information market. Given these stakes, there is zero economic incentive for vendors to work together on standardization or interoperability.
Lastly, the administrative and bureaucratic dimensions of administering health care among the various stakeholders is complex and in no danger in simplifying anytime soon. Billing codes, local and federal compliance requirements, various industries’ competing interests, medical service specialization, and lack of engagement among stakeholders and the medical IT industry have fostered the current state of confusion.
What about the patients?
Ironically, patients are the stakeholders paying the steepest price and the only ones experiencing existential consequences (see Figure 2.). The frenzy to appropriate EHRs and develop ecosystems with network effects that lock-in patient information behind corporate firewalls have serious implications hardly discussed. Individual patients are paying the price through degraded quality of treatment and having our privacy and dignity associated with our most personal information being trampled upon. Adding economic insult to potential injury or death, we’ve lost control of our digital health records and share none of the profit from intellectual property stemming from our own digital record. Meanwhile, individuals face opaque medical pricing schemas that are impenetrable and appear to be without any accountability.
Faced with systemic complexity, a massive market opportunity, de facto lock in of an arbitrage model for information generation and ferocious competition, it seemed the perfect test for shiny tech. I thought to myself, “If blockchain EHRs can offer any modest product solution for this multi-headed hydra of problems, there just may be a product in there somewhere.”
Part II: Take a Blockchain and Call me in the Morning
Armed with a serious, large and growing problem set, a very large market opportunity and a host of unmet user and stakeholder needs, I sought to assess blockchain’s capabilities for EHRs versus current solutions. My goal was to answer some basic questions about blockchain technology. “Is blockchain mature enough to be considered for product development in an intensely competitive market segment? Or is it a nascent technology needing years of development prior to implementation?”
From a product development perspective, blockchain is not a standalone solution for implementing an EHR. Rather, it is an architectural component that enables a total product solution. A large portion of current EHR problems are outside of the scope of blockchain such as user interface design, workflow customization, usability and a large portion of customer experience. In these cases, blockchain needs to enable solutions to important user needs. Customers don’t use (let alone purchase) any product based upon blockchain (or any other technology.) Customers use products that solve their problems and meet their needs. So I asked the question, does blockchain today enable important solutions to customer needs that significantly exceed current solutions?
Blockchain enables solutions to EHR problems today- and tomorrow
Applying blockchain’s current capabilities to EHR needs, I found a natural fit offering advantages stemming from blockchain’s fundamental properties. (see Table 2. Blockchain Enabled EHR versus Contemporary EHR.) Since it is a distributed, synchronized ledger, blockchain is very difficult to tamper with. This provides strong record integrity but does not satisfy rigorous privacy needs as it’s lacking native encryption. Because it is distributed, it offers high availability and resilience given that there are a sufficient number of geographically distributed nodes. Blockchain has proven that it’s capable managing many parties accessing it simultaneously and is able to reconcile the various events and transactions in an orderly fashion in a reasonable period of time across the entire network of nodes.
Patients are empowered by having ownership, access and control over their health records for the first time. Blockchain enables individuals to give permissions and access to necessary care givers for updates as well as contribute information for population health studies if they so choose. Since patients have control over their health records via blockchain and they are accessed and updated by various care providers, completeness can be achieved for the first time. Regulatory privacy mandates and individual permissions govern copying and reuse of individual health information.
Caregivers receive the benefits of interoperability as well as completeness of records with a blockchain empowered EHR. As patients see various specialists or care providers, each can update the same record blockchain EHR. These are 2 needs repeatedly cited as impacting patient safety. While physicians consistently complain about lack of usability, I think many of the issues of poor usability have little to do with a blockchain ledger. They are a mix of regulatory complexity, insurer bureaucracy, poor user experience design, medical specialization, cargiver workflows variances and complicated procedures and the various ways that health information is collected and reported.
|Patients||Access||Partially met||Met need|
|Ownership/Control||Unmet need||Met need|
|Completeness||Unmet need||Met need|
|Privacy||Partially met||Met need|
|Availability||Met need||Met need|
|Integrated messaging||Out of scope||Future capability|
|Automated monitoring||Out of scope||Future capability|
|Care-Givers||Usability||Partially met||Out of scope|
|Workflow||Unmet need||Future capability|
|Interoperability||Unmet need||Met need|
|Completeness||Unmet need||Met need|
|Patient Safety||Partially met||Future capability|
|Availability||Met need||Met need|
|Billing/Settlement||Out of scope||Future capability|
|Regulators||Compliance||Met need||Met need|
|Privacy||Partially met||Met need|
|Security||Partially met||Met need|
|Payors||Verification of services||Partially met||Met need|
|Verification of coverage||Unmet need||Future capability|
|Billing/Settlement||Out of scope||Future capability|
|Medical Research||Access to population health data||Partially met||Met need|
Table 2. Blockchain Enabled EHR versus Current EHR
Source: Chris Montaño
Blockchain’s tamper resistance combined with its complete historical record makes it an ideal system of record. Regulators can use this for auditing and verification of compliance with HIPPA on privacy and access. Privacy needs would have to be met with encryption for individual records during storage and prior to transmission among the various parties. (Details on encryption can be found in the proposed architecture section.) Also, access to patient health information for use in government population health studies are enabled by patients granting use of their information if they choose.
Payors and insurers benefit from being able to audit delivery of services prior to settlement. While fraud from misrepresentation of services delivered is outside of scope, blockchain would make fraud due to record tampering obsolete. Cycle time for data record review and verification would decrease as well as accuracy.
Blockchain EHR Roadmap Offers Compelling Opportunities
While many stakeholder needs are not met today, continued development of blockchain technology offers capabilities that current architectures will be hard pressed to deliver.
- Building automated, intelligent transactions into the ledger
- Global regulatory syncronization
- End of revenue extraction from information exclusivity
- End of wholesale medical record theft
The ability to build automated, intelligent transactions into the ledger (a.k.a. “smartcontracts”) offers potential solutions to some of healthcare’s most troubling problems. For example, the opaque, labyrinth of billing and reimbursement cycles. Patients struggle with pre-authorization that payors claim is non-binding. And physicians outsource billing and reimbursement to clearing houses because the pain of getting paid by insurers is so great they are willing to pay others to do it. For example, insurers and payors give pre-authorization for one coded service only to discover later that another code has been entered.
Building intelligent transactions into a blockchain enabled EHR could create an authorization, services, billing, verification and reimbursement cycle where each party’s needs can be met. There could be automated pre-authorization as insurers place policy coverages onto the blockchain. Insurers can receive real time, automated, verification of services matched against insurance coverages and settlement and payments could be handled in a compressed time frame measured in minutes rather than weeks or months.
There are currently 2 methods for implementing blockchain technology:
- “Free Range” Blockchain. This is how blockchain is used for the digital currency, Bitcoin. It is a fully distributed, and an entirely open source code implementation in which anyone can participate. Originally viewed as self-funding and requiring no centralized management, there is currently a foundation that oversees development of Bitcoin. The incentives to establish a full node of the distributed ledger and pay for the costs of running it are derived from the “mining” Bitcoins and additional transaction fees for those that conduct business or send Bitcoins or various alternate coins.
- Permissioned blockchain platform implementation in a private network. There are numerous product innovation labs, experimental groups and efforts to commercialize blockchain. The private sector has focused on blockchain solely as an enabling technology and isn’t seeking aspirational elements such as decentralized governance or pseudonymity. From a commercial perspective, ambiguous identity could pose legal and regulatory risk. The solution is to have a blockchain implementation where all participants in the network have strong identities suitable for contractual and regulatory requirements. The cost of creating and maintaining the blockchain distributed network is borne by the organization that creates the permissioned blockchain.
Either “free-range” or permissioned blockchain EHR are possible. However, a “free-range” EHR is better positioned to meet needs of interoperability and individual ownership. Challenges of an open source EHR include resources for development and maintenance of the client solutions. Also, implementation and consulting services would be a grassroots effort until a large base of installations could attract professional services support.
I have developed a hypothetical architecture for a blockchain EHR (See Figure 3. Hypothetical Blockchain-EHR Upload and Figure 4. Hypothetical Blockchain-EHR Update by Caregiver). The components of the system include:
- A Client application on the patient’s personal computer
- Client application(s) on the stakeholder system (doctors, hospitals, insurance, etc.,…)
- A Distributed Storage Network (DSN)
- A blockchain to store information necessary to manage individual patient records
The patient client acts as our personal control center for our health records. It allows us to perform a number of important tasks including viewing our health care records, managing permissions for caregivers and other stakeholders to access, view, or update our EHRs. Importantly, it manages the uploading of our health records in a secure manner to a Distributed Storage Network (DSN).
Future capabilities of the patient client might include dazzling features such as managing automated monitoring and updating of EHRs via personal health devices such as sports monitors, implanted insulin administration and dosing, pacemakers and other various devices that are being developed. An integrated messaging and alarming capability could be incorporated to manage communication among the broad array of caregivers and ensure a complete medical record. If the privacy and security needs could be met, there could be an entire ecosystem of healthcare applications that could link to the system focused not only on therapeutics but healthy lifestyles and wellness outcomes.
Figure 3. Hypothetical Blockchain-EHR Upload
Source: Chris Montaño
Since blockchain is not presently suited for large data storage, the expected large amount of EHR data suggests we need an alternate storage system. There are a number of efforts underway to use blockchain technology within a “distributed storage network.” Included in these efforts are Sia, Maidsafe, Storj, and Ethereum-based efforts. I expect a host of DSN efforts in a rainbow of designs and flavors to percolate forth. The DSN I used in my architecture is based upon the Storj DSN which is currently in beta testing. It is a geographically distributed group of storage nodes available for use to anyone. However, in the case of a permissioned network, the DSN nodes would belong to the company that operates the DSN. The clients in the blockchain-EHR parse a patient EHR into a group of smaller blocks of uniform size. These parsed blocks are then each encrypted and the cyphertext blocks are then individually hashed and sent out onto the DNS. These hashed shards of an individual record are geographically distributed, each are stored at multiple locations such that no location has the information necessary to assemble a complete group of hashed shards of an EHR. The client interacts with the blockchain to record locations of the hashed shards, and is able to audit for tampering using hashes. The use of blockchain in the DSN means that we can use its multi-signature capability to manage access to the records for various stakeholders.
Security and privacy compliance in the blockchain-EHR far exceeds that of any of today’s EHR systems. The client ensures that no health record information is transmitted in cleartext. And since each parsed block is hashed, auditing for tampering can be integrated into the system with minimal effort. The DSN, contains no clear text of any type and since the blocks are identical in size and hashed files of encrypted information, there is no way to infer content or identity by file size. Since the hashed shards are distributed such that they are replicated across the network with no node containing a complete set of shards of an individual EHR, then even if a node was compromised, the bounty would be a massive set of hashed, encrypted shards of incomplete EHRs- each with their own private key. Replication ensures disaster proofing as well as high availability of EHRs.
Figure 4. Hypothetical Blockchain-EHR Update by Caregiver
Source: Chris Montaño
Updates to EHRs are initiated by caregiver client machines making an access request to a EHR owner. If the owner chooses to grant access, multisignature capability in the blockchain verifies and grants access. The caregiver client then receives the necessary information to retrieve the shards which it validates with hashes, then decrypts and reassembles the EHR for viewing and updating. Once the update is completed, the caregiver client uses the same upload process as the patient client and messages the patient that their record has been updated.
Overall, I was more than a little surprised that there was a natural fit and compelling potential roadmap for blockchain enabled EHRs. The one caveat being that the distributed storage networks are early in development and in beta rather than deployed for general availability. If this component proves problematic, then storage of the burgeoning health care data set presents a major design challenge.
The current EHR market confirms the conclusion of natural problem-solution fit. There are several efforts comprised mostly of partnerships between entrepreneurial startups and large IT specialists. And given the potential and extraordinary market size, there are efforts currently underway with many more to follow.
|Philips Blockchain Labs||Philips Health IT
|Multiple announcements with several partners. None which are definitive other than there is a Blockchain Lab|
|MedRec||MIT student effort||Latest publication|
|IBM||IBM||Launched blockchain services targeted at healthcare and other applications.|
|Estonian e-Health Foundation||Estonian e-Government Infrastructure
|EHRs at about $10 per record to participate.|
Table 3. Incomplete list of EHR Labs and Efforts using Blockchain
Source: Chris Monaño
Author’s note: There were no Merkle trees altered in the research, writing, and editing of this article.