As a researcher who studies electronic health records (EHRs), I’ve lost count of the times I’ve been asked “Why can’t the systems talk to each other?” or, in more technical terms, “Why don’t we have interoperability?” The substantial increase in electronic health record adoption across the nation has not led to health data that can easily follow a patient across care settings. Still today, essential pieces of information are often missing or cumbersome to access. Patients are frustrated, and clinicians can’t make informed decisions. When our banks talk to each other seamlessly and online ads show us things we’ve already been shopping for, it is hard to understand why hospitals and doctors’ offices still depend on their fax machines.
A big part of the reason is that interoperability of health information is hard. If it were easy, we would have it, or at least have more of it, by now. Though it’s a technological issue, it’s not just a technological issue. As we have seen in other industries, interoperability requires all parties to adopt certain governance and trust principles, and to create business agreements and highly detailed guides for implementing standards. The unique confidentiality issues surrounding health data also require the involvement of lawmakers and regulators. Tackling these issues requires multi-stakeholder coordinated action, and that action will only occur if strong incentives promote it.
Though billions in monetary incentives fueled EHR adoption itself, they only weakly targeted interoperability. I have come to believe that we would be substantially farther along if several key stakeholders had publicly acknowledged this reality and had made a few critical decisions differently. While it’s too late for “do-overs,” understanding initial missteps can guide us to a better path. Here is how those key stakeholders, intentionally or not, have slowed interoperability.
The 2009 Health Information Technology for Economic and Clinical Health (HITECH) legislation contained two basic components: a certification program to make sure that EHRs had certain common capabilities, and the “Meaningful Use” program, divided into three progressively more complex stages, that gave providers incentive payments for using EHRs. The legislation specified “health information exchange” (HIE) as one of the required capabilities of certified EHR systems. However, the Centers for Medicare and Medicaid Services (CMS) and the Office of the National Coordinator for Health IT (ONC) had substantial latitude to decide how to define health information exchange as well as how to include it in the Meaningful Use program and the accompanying EHR certification criteria.
CMS and ONC decided to defer the initial HIE criterion — electronically transmitting a summary-of-care record following a patient transition — to Stage 2 of the Meaningful Use program. While many of the capabilities required for providers to meet this criterion would have been challenging to develop on the Stage 1 Meaningful Use timeline, deferring HIE to Stage 2 allowed EHR systems to be designed and adopted in ways that did not take HIE into account, and there were no market forces to fill the void. When providers and vendors started working toward Stage 2, which included the HIE criterion, they had to change established systems and workflows, creating a heavier lift than if HIE had been included from the start.
While the stimulus program needed to get money out the door quickly, it might have been worth prioritizing HIE over other capabilities in Stage 1 or even delaying Stage 1 Meaningful Use to include HIE from the start. At a minimum, this strategy would have revealed interoperability challenges earlier and given us more time to address them.
A larger problem is that HITECH’s overall design pushed interoperability in a fairly limited way, rather than creating market demand for robust interoperability. If interoperability were a “stay-in-business” issue for either vendors or their customers, we would already have it, but overall, the opposite is true. Vendors can keep clients they might otherwise lose if they make it difficult to move data to another vendor’s system. Providers can keep patients they might otherwise lose if they make it cumbersome and expensive to transfer a medical record to a new provider.
As a result, the weak regulatory incentives pushing interoperability (in the form of a single, fairly limited HIE Meaningful Use criterion), even in combination with additional federal and state policy efforts supporting HIE progress, could not offset market incentives slowing it. Without strong incentives that would have created market demand for robust interoperability from the start, we now must retrofit interoperability, rather than having it be a core attribute of our health IT ecosystem. And, if there had been stronger incentives from the start, we would not now need to address information blocking: the knowing and intentional interference with interoperability by vendors or providers.
Another criticism levied at policymakers is that they should have selected or certified only a small number of certain EHR systems, because narrowing the field of certified systems would have at least limited the scope of interoperability problems. A few even advocated that Congress mandate a single system such as the VA’s VISTA. While such a mandate would have helped solve the interoperability issue, it would have violated the traditional U.S. commitment to market-based approaches. Moreover, the United Kingdom failed very visibly with a heavily centralized approach, and in a health care system the size of the United States, an attempt to legislate IT choices in a similar manner could backfire catastrophically.
Most observers assign EHR vendors the majority of blame for the lack of interoperability, but I believe this share is overstated. As noted above, by avoiding or simply not prioritizing interoperability, they are acting exactly in line with their incentives and maximizing profit.
Normally, the United States glorifies companies that behave this way. When neither policymakers nor providers were demanding interoperability, vendors risked harming their bottom lines by prioritizing it, and they cannot be blamed for acting in their economic best interest in wholly legal ways.
Nevertheless, senior leaders at EHR vendors should have been more willing to come forward and explain why their economic best interest was at odds with existing regulations. Instead, they often claim to have robust interoperability solutions when they do not, and similarly claim that interoperability is a top priority when it is not. This gap between rhetoric and reality makes it harder for providers and policymakers to demand greater interoperability.
As noted above, providers may not have a strong business case to prioritize interoperability. However, providers have professional norms and mission statements that should motivate them to pursue interoperability (or at least not actively interfere with it) to benefit their patients. As a result of these conflicting motivations, some provider organizations let competitive pressures drive interoperability decisions (including not demanding robust interoperability from their vendors), while others have chosen to pursue interoperability because it is best for their patients, even if this decision incurs competitive disadvantage. More providers may tip toward the former simply because today’s interoperability solutions are complex and costly. It is hard to justify investing in a complicated, expensive capability that also poses a strategic risk — a double whammy.
The emergence and rapid growth of Epic’s Care Everywhere platform (which connects providers using Epic) suggests that even in highly competitive markets, providers may easily tip the other way when the cost and complexity of interoperability are reduced. Therefore, any efforts that successfully reduce cost and complexity are highly valuable, though not a substitute for stronger incentives for providers (and vendors) to engage in interoperability.
As with vendors, we cannot fault providers for behaving in ways that are aligned with their incentives, but we can argue that their patient care mission requires, at a minimum, more public disclosure about the business rationales behind their interoperability decisions.
The point of the blame game is not to punish the players. It is to understand the dynamics at play and plot a path forward. Of the stakeholders, only policymakers have a clear, strong interest in promoting interoperability. Therefore, it is up to them to ensure that robust, cross-vendor interoperability is a stay-in-business issue for EHR vendors and providers. Once the business case for interoperability unambiguously outweighs the business case against it, both vendors and providers can pursue it without undermining their best interests.
This article originally appeared in NEJM Catalyst on February 22, 2017.