Clever for Healthcare

June 29, 2015

Following up from a previous post on the digital opportunities that could transform healthcare, I’ve had quite a few people reach out and ask why there hasn’t been a “Clever”-like solution in healthcare (e.g., the ability to connect legacy, disparate IT systems used by providers together into a universal API)?

I've discussed and explored this topic with the help of many in the industry, both on the provider side as well as those on the IT side. A few key obstacles impede this opportunity from being easily realized:

1. Non-trivial technical challenge: Providers maintain multiple IT systems, often from a variety of vendors, to make their back-end technology work. Getting these systems to speak with each other is no easy task. In the education space, a consolidated set of IT providers to public k-12 institutions allowed Clever to create easily scalable data layers on top of the existing systems. In healthcare, not only are there a wide variety of vendors, but all IT implementations are also highly customized. This high-level of customization makes gaining scale extremely difficult as each integration becomes a very manual process. Even simple items like a gender identification marker can become complicated to match between systems as they may denote these in multiple permutations (e.g., "Female", "F", "Fem", etc.) Companies such as Health Catalyst and Explorys (recently acquired by IBM) have created data warehouse products to do the hard work of tying systems together, but rely heavily on large engineering teams to go into systems manually to develop these large data warehouses. These companies are then building analytical tools on top of their warehouses to sell to providers, but have not opened up APIs to outside developers (like Clever has done in the education sector).

2. Pain points not evenly distributed: Regulation from Obamacare will compel many providers to leverage their data more effectively to manage items such as population health, managed care outcomes, and improved cost management, but many providers have not felt the burning need to invest heavily in these issues as of yet. As hospitals have limited budgets, selling software that has not yet been deemed a core necessity by hospitals may limit the revenue potential of such a venture and may not offset the high-engineering costs of creating custom/manual data integrations in the first place. 

3. Culture: Selling software to healthcare providers is no easy task as traditional sales-cycles last at least eighteen months. Doctors hold the majority of decision making power in hospital administration, and as they are dealing with highly critical issues (i.e. keeping people alive), there is a general sentiment towards being more risk averse, especially when it comes to spending on reliable IT systems. Software with long-term ROIs will not make for an easy sell in this environment. This makes the task of adoption for new, unproven startups very difficult and the reliance of a strong network within hospital administration extremely important. 

4. Long-term vision: In the education space, Clever made its API attractive by being able to provide data for a large percentage of public schools quickly through integrations with a set of common IT vendors. In healthcare analogy, one would have to go provider by provider to create custom integrations for each (as discussed earlier) to gain large market share. Thus, it will take many years for such a venture to reach enough scale with these integrations at which point the API would become attractive enough for other developers to begin to use. Any team looking to pursue this route should be ready for this long journey and also understand the funding requirements needed to be successful on such a path. 

Although these challenges are large, all of these obstacles could be overcome with persistence, an intelligent solution, and a well-connected team. Regulation may also help the cause of individuals looking to tackle this problem as more standardized EHR requirements emerge and the federal government pushes existing providers to create more open APIs (EPIC has already started on this path, to some degree). Additionally, Flatiron Health has shown that canny business models may also help facilitate the adoption of such tools in specific verticals.