Blockchain, Data Obstruction, and the Promise of Information Sharing for a Value-Based Health Care World
The current system of health care delivery and payment has tremendous inefficiencies that relate to an inability to share data among multiple parties, and from that sharing, glean useful and actionable information that yields higher quality and lower cost care. This “data obstruction syndrome” manifests itself in degraded decision-making in real-time clinical care. It also limits the space where prospective value-based agreements can develop between business entities that have pathway or guideline components. Blockchain technology is in its infancy today in health care, but it will expand rapidly in the future. Use of blockchain can facilitate data flows among multiple parties who do not need to agree on all data management policies, but who can agree to participate in a private network based on verifiable trust. This will push the current business environment, which presently promotes the false choice of data security over multi-party data sharing, to move more aggressively on value-based models of care and payment in a richer data environment.
The US health care system’s multiple maladies—fragmentation of care, regulatory/business complexity, and cost burdens with misalignment of incentives—cannot be cured with a single treatment. What is needed is a rigorous campaign of process improvement across multiple areas. Lessons from specific health care systems include shifting services to outpatient care, placing greater emphasis on primary and preventive care, facilitating case management for longer term and chronic care, adopting information technology and a system-wide electronic medical record (EMR), use of performance measurement, and management of prescription drug costs.1 While some businesses in health care are taking these steps today, we must ask ourselves about our present state of affairs and how we got here in the first place.
Data Obstruction Syndrome
The phrase “data obstruction syndrome” refers to the inability of multiple parties to share real-time clinical information in the health care environment in an easy and useful manner. This is not truly an interoperability issue. Health Level-7, Fast Healthcare Interoperability Resources, and other mechanisms of data normalization currently support data exchange between disparate electronic health records (EHRs) and electronic systems.2 The bigger problem resides in the mechanisms for discovering and authorizing access to personal health information (PHI) across organizations while complying with legal, regulatory, and contractual requirements to protect the security, privacy, and commercial rights of various stakeholders.
Data can flow mostly unimpeded when patients receive 100% of their care within one health care system that also maintains a compliant and interoperable set of electronic records for care/payment capture. This “100% enterprise-centric” model is uncommon in the United States today, but it can be seen in systems like Kaiser Permanente3 and Geisinger Clinic.4 More commonly, however, patients seek care from multiple systems of care that are not a single business entity. While the health insurance payer may be able to provide a unified view of data that reflects all care charged and billed, this administrative database is an insufficient fix for what ails health care today. It reaches the limits of utility when it comes to understanding the overall context of care: provider choices, patient behavior and preferences, as well as many of the outcomes of interest. Also, miscoding and incomplete coding of services is rampant, and the claims data often has lags of up to 3 months.
Multiple providers on disparate systems that have not established business-to-business (B2B) agreements to share PHI will not be able to exchange data smoothly enough to deliver coordinated, cost-effective care. The number of B2B agreements needed to support effective data generation and exchange for most patients in the United States today would be too far beyond the capabilities of any single provider or delivery system entity to justify investment. Thus, data obstruction syndrome is born.
Big Data and the Need for Data Agreements
This untenable need to have business agreements among an exponentially growing number of entities is staunching the flow of useful clinical information. Why the reluctance to give and receive PHI? Because holding information carries security risk, thus holding a large amount of other people’s information creates even more risk. In practical terms, we have seen many examples of hacking of financial and health care databases, even those thought to be completely secure. The large risk component acts as a deterrent for business entities, thereby limiting the number of data agreements and the scope of data capture.
The regulatory compliance, contracting, and security risks on the data capture side has created tremendous barriers to easy information sharing. Ironically, this has also created a huge business opportunity and windfall for enterprises in the data mart business. By creating a series of B2B agreements for data sharing and/or using existing data from in-house activities of related businesses (eg, payer, pharmacy benefit manager, group purchasing organization, practice management, electronic data interface companies), the data is processed and then put up for sale. Data monetization in this way leads to the belief that every data source is a “gold mine.” A large amount of our health system’s usable data is locked up in silos due to the owners’ belief that they can monetize the data. Each entity is selling their slices of data to others in the health care system that need it to function optimally and efficiency. However, that data is always flawed, or “lobotomized”—incomplete, de-identified at the patient level, and of uncertain provenance. This is particularly true where the purchaser is not a part of treatment, payment, or operations (and thus not a covered entity). So, the data is expensive and not very complete or easily usable—woefully inadequate to the needs of personalized medicine and meaningful, value-based contracting.
In the era of the Affordable Care Act, what breadth and depth of data—including data on access and affordability, quality, and patients receiving personalized treatments—is truly meaningful when it comes to optimally understanding and managing patients?5 The scale of data required to execute higher order value-based agreements in today’s environment, including personalized medicine therapy, is even higher. Critics of value-based contracts (VBCs) sometimes mention this data acquisition shortfall, but they never acknowledge how central this issue is for success or failure of VBCs.6 The issue goes beyond, as Steve Williamson puts it in Lancet Oncology, “measuring refund-triggering events and submitting requests for refunds that would require additional resources and personnel that may more than offset marginal savings.”7 Without the appropriate level of data to be translated into actionable information, the majority of true clinical insights in VBCs are mostly hidden. This would include full direct/indirect savings, patient reported outcomes, acuity risk adjustment, and the impact of new therapies on holistic value—all unable to be determined. Data context is lost because it is unknown. The value engine in novel contracting has been disabled before even being fully turned on.