The Data Operating System: Changing The Digital Trajectory Of Healthcare

In 1989, John Reed, the CEO of Citibank and the early pioneer for ATMs, said, “I can see a future in which the data and information that is exchanged in our transactions are worth more than the transactions themselves.” We are at an interesting digital nexus in healthcare. Few of us would argue against the notion that data and digital health will play a bigger and bigger role in the future. But, are we on the right track to deliver on that future? It required $30B in federal incentive money to subsidize the uptake of Electronic Health Records (EHRs). You could argue that the federal incentives stimulated the first major step towards the digitization of health, but few physicians would celebrate its value in comparison to its expense. As the healthcare market consolidates through mergers and acquisitions (M&A), patching disparate EHRs and other information systems together becomes even more important, and challenging. An organization is not integrated until its data is integrated, but costly forklift replacements of these transaction information systems and consolidating them with a single EHR solution is not a viable financial solution. At Health Catalyst, we’re not satisfied with the current trajectory of digital healthcare. Health information exchanges (HIEs) have struggled technically and economically. Interoperability is lagging and remains one of the top concerns of clinicians and executives today.  EHRs are now widely available, but physician burnout is the highest on record-- they now spend over 50% of their day interacting with a computer, not patients. Only 8% of the data required for precision medicine and population health currently resides in today’s EHRs. Clinical and financial decision support at the point of care is almost nonexistent, restricted to a few pioneering organizations who can afford the engineering and informatics staff to implement and maintain it.  The traditional design of batch oriented healthcare data warehouses based on monolithic, early binding data models, focused almost exclusively on conference room analytics, cannot keep up with the rapid reporting and data analysis changes in the industry, or the pace of decision making that requires the right data, at the right time, delivered to the right person. According to the Center for Medical Interoperability, “Health information is trapped.” Yet, while healthcare lags behind in the digital world, the technology and the engineering patterns in data and software have never been better, thanks to the contributions to open source technology by the five largest listed companies in the world--Amazon, Apple, Google, Facebook, and Microsoft. The time is right for a major change in healthcare. At Health Catalyst, we are disrupting ourselves and the industry by developing a FHIR-based healthcare Data Operating System (DOS) which leverages the technology and engineering patterns of Lambda architectures in Silicon Valley, enabling real-time data exchange and interoperability along with decision support and analytics in the same environment, at a fraction of the cost invested in HIEs, EHRs, and traditional data warehouses. Within the DOS, Health Catalyst will open source the APIs within a layer called the Fabric, encouraging 3rd party software vendors to leverage the DOS infrastructure for our clients, much like typical, modern Operating Systems—Android, Windows and iOS—but at the data level. By using the FHIR framework, the Health Catalyst DOS can push the output of machine learning models and other analytic algorithms into the workflow of existing EHRs, turning them from clinical data entry terminals into clinical decision support tools. The Fabric can lay over the top of existing data warehouses, as well, leveraging those investments in data aggregation. The DOS concept will become the digital backbone for healthcare and other industries, as society moves from transaction-oriented information technology to intelligence-oriented information technology. For those who might think this is a conceptual dream, Health Catalyst is already well on its way of implementing the DOS with off-the-shelf APIs that integrate data from 154 different healthcare data sources; 212 reusable healthcare data bindings, and 24 healthcare-specific machine learning models… and adding to each of these libraries, everyday.