In 2015 Partners Healthcare with its flagship Massachusetts General and Brigham and Women’s Hospitals completed a multi year $1.2B rollout of Epic. A modern EHR (Electronic Health Record) solution, in 2015 Epic was still built predominately with visual basic — which Microsoft discontinued in 2005. Cerner, their chief rival just signed an initial $4.3B contract to implement the Department of Defense’s EHR solution. Cerner is chiefly written in programming language only they use. Over the past several years these are the best rated solutions in the industry for hospitals.
In healthcare you can feel the future and past colliding. On one hand we’ve sequenced the human genome and are working on changing the very code we humans are written in. On the other we see new enterprise technology initiatives that are disturbingly reminiscent of 1990s ERP solutions.
Why has healthcare lagged on cloud adoption?
One of the chief challenges of working in the healthcare market is the fact that it’s not a market in the traditional sense. Consumers don’t pay for what they consume, providers have or are granted monopolies, government is both an active payer and regulator, and there are not standards for what better health is worth. While solving healthcare overall is beyond our scope, the environment technology exists in is critical to understand what drives change; since change often drives the need for technology.
In 2009 the government saw an industry still running primarily offline — only 12% of hospitals and doctors had computerized records. As a result, through Centers for Medicare and Medicaid Services (CMS) the government created incentives and the threat of future penalties to promote the meaningful use of electronic medical records. As of 2015 the government has paid out approximately $30B in incentives to adopt solutions supporting them. Today over 75% of physicians and hospitals have an electronic solution.
On the surface, this is a significant step forward. But this step potentially hindered the giant leap the industry needed to make. In any industry such incentives will promote technology adoption. But the process of adoption will favor the solutions that have existing market share. Due to the rush to take advantage of the government incentives, incumbent’s legacy solutions benefitted most. All while the cloud computing revolution was dramatically shifting the technology landscape in every other industry.
In 2009 there were no cloud EHR providers for hospitals. Of the two major cloud EHR providers today, then athenaHealth didn’t provide such solutions and PracticeFusion focused on outpatient providers, not hospitals. There were not viable cloud solutions the incentives could have accelerated.
In addition to incentives, other factors also influenced healthcare’s decision to adopt more known legacy solutions:
- regulations such as HIPAA adjusted slowly to the reality of a cloud based world.
- industry EHR providers — intentionally or because of technical debt — were slow to allow integration to their solutions.
- legacy vendor’s breadth of functionality was ahead of what early cloud solutions could provide
Healthcare's future demands cloud based innovation
How well these legacy solutions have done in helping the industry is a matter of perspective. Their performance relative to paper-based records has been positive but mixed — some acknowledging their impact and others seeing initial benefit paired with painfully slow advancement. Evaluating them against the apps we use as consumers has led to vitriol and cruel parody.
Yet the true cost to the healthcare industry is more than just less-usable solutions or being one cycle behind in technology adoption. It has left the industry on the wrong side of a massive technology chasm.
For the first fifty years of technology businesses led the way. Organizations would use technology, like a computer, and in a few years it might trickle down to consumers. Yet despite the advances the impact of technology on business seemed limited. It granted temporary advantages rather than dramatically impacting business models. This led to Nicolas Carr’s incendiary 2003 Harvard Business Review article stating “IT Doesn’t Matter.”The piece drew wide controversy and eventually begrudging acknowledgement that IT was only generating commodity-like returns.
The rise of the internet changed both how technology entered the world and its impact on businesses. Consumer technology leaped ahead — stuck with green screens at work, but empowered by Angry Birds, your investment portfolio and the sharing economy in the palm of your hand.
Cloud technology for businesses transformed organizations; bursting past the question of does IT matter to how can technology transform every industry. Apple, Google and Facebook used technology to redefine industries digitally. Amazon, Uber and Airbnb upended historically offline markets. In the internet era of cloud, technology could finally keep pace with business. In fact it would become the only way business could keep pace with change. Healthcare finds itself on the wrong side of this great divide.
Concretely, healthcare’s legacy solution architecture has limited information flow and exchange. Wave after wave of regulations and standards, private alliances, public shaming, and non profit intermediaries have continued to struggle with the seemingly simple task of moving health information easily and around securely. Part of the challenge is the willingness and incentives for legacy incumbents to change; however, even where the desire exists, the technical challenges are significant. The antiquated technology they are built on and the model of their deployment restricts potential solutions.
Outside of healthcare, throughout the late 90s and early 2000s, the software industry — faced with similar legacy silo’d solutions — touted a myriad of integration strategies and standards (Web Services, UDDI, etc.). These approaches sound eerily like current discussions in healthcare technology — with the same poor results. It was only once internet and cloud applications emerged that real progress was made. Above all this happened for one key reason — cloud solutions and integrations between them are delivered from a single place. It is unnecessary for thousands (or millions) of endpoints to each bear responsibility for all integration.
Integration challenges create friction on the entire system. New mobile applications, consumer technologies, third party comparisons tools and scores of other categories are difficult to create without costly data integration. Integration challenges, directly arising from the legacy architectures, place a tax on all future innovation.
Similarly, we see user experience in healthcare lacking because legacy architectures make it difficult to escape from client server paradigms. Internet and cloud solutions get better / faster / easier for the user nearly every day. In contrast, even minor software updates for hospitals today can take months and require doctors undergo extensive training and planned reduction in productivity.
Many times, technologists at healthcare organizations site security and privacy, claiming their situation makes cloud solutions less feasible. There is no doubt there are important necessary rules in securing healthcare solutions. But this focus misses the point — the current paradigm of deploying and managing local solutions forces every hospital and provider to tackle the exact same problems. Recently attackers have relentlessly targeted hospitals with ransomware. Each hospital defends against these problems most likely without the expertise or resources necessary. Cloud solution providers have not had similar issues because protecting against such attacks is core to their business model.
In addition to missing out on new capabilities, better user experience and enhanced security, cloud solutions in every industry have dramatically lowered total cost of ownership for technology.
Five steps to a cloud future
Integration, usability, security, privacy, rate of innovation — diminished success with all of these are the consequences of healthcare’s continued investment in legacy solutions. But given healthcare systems today are still investing in these solutions, what’s likely to happen in the industry over the next decade ?
Despite all the negative externalities imposed by legacy solutions, they are here to stay — for the intermediate future. The level of investment and change needed to shift again (and likely without government subsidy) will ensure legacy healthcare solutions have a mainframe-like ability to stick around. However, we will rapidly see hospital systems go through their own stages of grief in coping with the consequences.
- Healthcare systems will realize the extent of the limitations of their “updated” EHR / legacy systems. Information silos, slow pace of innovation, poor ability to keep up with changing players and conditions — all of these and more will dawn on business and technology executives in the aftermath of their rollouts.
- Current systems will become tailored to be systems of records vs. systems of engagement. The primary “users” of Legacy EHR systems will be resources who spend most of their working time interacting with it (i.e. those who understand how to navigate its limitations where possible). Systems of engagement will start to live outside of the EHR and other legacy solutions and be what most users see. Behind the scenes and away from the users these newer cloud based solutions will exchange data with legacy technologies.
- Middleware / integration solutions will start to shift to the cloud to support this new app innovation model. By abstracting the point of integration away from EHR / legacy solutions, healthcare participants will be able to build social, mobile and engagement apps for their personnel, patients and partners more effectively. The heavy lifting of integrating with their on-premise environment will be done just “once” to the cloud to create abstraction. While not as rich as having a cloud based EHR, it will partially mitigate the integration challenge. It will allow third party apps and custom developed ones to plug in more effectively and with more reuse across customer environments.
- Data infrastructure will continue to be a struggle and sit both on premise and in the cloud. The sheer volumes of data and legacy limitations will make it more difficult to shift this integration pattern completely to the cloud. Organizations will start to realize traditional data warehousing solutions from large vendors are too expensive and not as well situated towards their needs (storage patterns of billing, clinical data, real time patient monitoring and genetic data are too varied). They will start to mimic the open source direction of internet scale companies and SaaS providers to create big data environments.
- EHRS will move to the cloud. Even with all the steps above to wrap and isolate legacy EHRs and solutions, it will become clear that the only reliable way to eliminate much of this complexity and increase pace of innovation is through cloud based EHR solutions. Cost and speed will both favor cloud based EHRs, with switching costs remaining as the key barrier. Opportunistically progressive, or desperate, hospital systems will start to select solutions with cloud-based architecture. Ironically, those who have best isolated their legacy EHR solutions may be the slowest to adopt cloud solutions. In order to overcomes the limitations of legacy systems they will have established larger IT groups that are focused on low level technical tasks — potentially creating another large group resistant to change. As a result, moving to cloud based EHRs will take years for some and decades for others.
With such radical shifts affecting healthcare technology, what kind of vendors will flourish or flounder? Find out in Part 2.
Narinder Singh co-founded Appirio — the leading independent cloud services provider — in 2006. Prior to that he worked in corporate strategy in the Office of the CEO at SAP, ran R&D at integration leader webMethods, and worked in Accenture’s Center for Strategic Technology. His current work promotes the use of machine learning and artificial intelligence to aid decision-making in healthcare.