You have heard the mantra about searching for “value” in healthcare. There have been dozens, hundreds, or thousands of federal policies — depending on how you parse the Federal Register — and entire organizations such as AHRQ, MedPAC and the CMS Center for Clinical Standards and Quality dedicated to this one goal. Why has finding value been so elusive?
To a large extent, our struggle to find value dates back to 1942 when, after the Roosevelt administration froze wage hikes to forestall inflation during World War II, employers turned to healthcare benefits as a major recruiting tool. The subsequent growth of employer-sponsored insurance limited the impact individual consumers could have on the market. Today, we are seeing advances in computing transform the calculus of consumer choice in almost every industry – telecom, retail, travel, entertainment, banking, publishing – but healthcare still lags. Much of the “value” measurement space represented by tools such as quality measures and prior authorization have been attempts to provide proxies for value to consumers / patients to fill the void in choosing what to pay for in healthcare.
What would it take to bring transformation to healthcare? The near-unanimous passage of the 21st Century Cures Act laid a foundation for a new dynamic in healthcare based on getting patient’s their patient data from providers enabling them, their apps, and their payers to explore entirely new ways of shopping for and getting care. Central to this dynamic are the Cures Act requirements for “application programming interfaces (APIs) without special effort” – i.e. standardized rather than vendor-specific APIs and the requirement that patients actually get that data — meaning no more information blocking by providers and vendors.
Getting Population-level Data
Interoperability rules promulgated by the Centers for Medicare and Medicaid Services and the Office of the National Coordinator for Health Information Technology set the stage for an app-based environment and those rules have received considerable attention. But there’s another area of potential transparency that could be of great benefit to patients. While at ONC, I heard from numerous providers and payers that they had limited data to design and evaluate the care they were providing or paying for.
So, in conjunction with multiple stakeholders, we worked to develop a population API standard. It is built on the same Fast Healthcare Interoperability Resources — commonly known as FHIR — standard that is used to enable apps via the interoperability rules. This is referred to as the Bulk FHIR API. Individual patient data downloads from an electronic health record to an app of the patient’s choosing are covered by the HIPAA Right of Access provisions. Any population of patient downloads are governed by the HIPAA Treatment, Payment, Operations provisions, and require agreement from both the covered entities, and a formal privacy preserving contract to allow that data access.
We are all familiar with apps and can imagine the types of apps patients might be offered in the future. The dynamics of Bulk FHIR for patients play out a bit differently. Our interest is in accountable and effective care, but we haven’t had a mechanism to robustly compute the value of care – no mechanism to look at the care, outcomes, and costs of an entire population of patients and assess performance.
The challenge of computing in detail on a much broader set of care outcomes beyond narrow pre-defined metrics has been one of getting the data and getting it in a uniform manner. We typically look at standards for single patient transactions such as processing a claim or retrieving a medical record. Looking at a population of patients is just as, or potentially more important to patients because that is how we spend resources in healthcare. We can only compare performance if we can obtain comparable measures. That has not been possible. For example, today there are many firms offering to use artificial intelligence algorithms to analyze clinical data. But if each provider’s data set is formatted differently, as they are today, it is hard to draw learnings or comparisons.
EHR Bulk FHIR APIs change this dynamic by allowing simple, consistent access to the US Core Data for Interoperability — think problem lists, medications, chart notes. This provision takes effect on Dec. 31, 2022 and will allow us — providers and payers with a patient duty under HIPAA — to start being smarter about the care we provide and to better use the modern algorithms available.
Harnessing New Technologies
Algorithmic thinking has transformed many industries and healthcare is ripe for rethinking. While individual outcomes of using this data are hard to predict, it seems inevitable that lower cost and more consistent treatments are quite likely. For providers burdened with a vast array of documentation requirements and laborious processes, we can use these tools to replace the artificiality of today’s quality, prior authorization, and piecemeal safety documentation processes and spend a much higher percentage of time on the actual care of patients.
One of the outgrowths of both the individual and population APIs will be the ability to harness the amazing new software stacks available today to bring scalable real-time care services to patients and to do so in a highly protected way with more accountability for privacy than has been possible with today’s mechanisms.
Bulk FHIR appears nerdy but this computationally straightforward approach to marshalling clinical and claims data has a high potential to provide the same magnitude of transformation in healthcare as we’ve seen elsewhere in our lives.