To become better at two key capabilities–saving lives and saving money–American healthcare providers will need sophisticated analytic tools to guide the transformation. To use those tools will require that providers build the foundation that all good analytics need: good data. Many people are talking about “big data” in healthcare but we already have lots of data. We just need it to be accessible and connected.
A recent Forrester report noted that most providers are still new to the idea that they have large amounts of structured data in clinical systems. According to Forrester, until recently the mission of the healthcare CIO was ancillary to a provider’s core mission because IT often fell under the CFO's domain since it focused largely on business systems.
While the HITECH act has pushed provider clinical systems into the digital era, huge issues remain. A lack of standardization and interoperability and the proliferation of proprietary electronic medical record (EMR) systems and clinical applications have created data silos that make it difficult to aggregate the data needed to answer important questions. And sophisticated analytics rely on comprehensive data from multiple sources.
To be sure, hospitals can and should be using analytics now to answer the questions that can improve care and make the system more efficient. But real gains will come when we can use population-level data to prevent disease and guide individual treatment. The best way to truly tame healthcare costs is to create an environment where the most effective treatment is used first and only as necessary.
Lay the foundation now
So what do we need to build the foundation? Ultimately, we need EMRs built to share data seamlessly. We need all clinical applications to use a standard terminology and operate on platforms that are interoperable. The situation now is comparable to the early days of personal computers, when a proliferation of proprietary word-processing applications meant that users often couldn’t open a file sent by someone using a different application or operating system. Now, everyone can share text-based files, although sharing healthcare data should probably not be based only on one vendor’s standard.
Until that transformation happens with EMRs and other clinical applications, we need to create interfaces that allow us to aggregate data from a multitude of sources, including images and structured and unstructured data. And we need to be able to store it in vendor-neutral archives that facilitate collaboration and data sharing or to create virtual, federated environments that facilitate the same thing. That will allow us to aggregate data for the predictive and prescriptive analytics that can help us identify the actions that will help people stay healthy.
Governance and standard terminology needed
At the hospital level, creating a data governance structure and a standard set of terminology is important. Clinical applications should be modified wherever possible to use a standard terminology to make interoperability easier. A master patient index, which ensures that each patient has a unique identifier, must be created to allow patient records to be aggregated in one complete file. And a data management strategy should be created that guides the organization toward creation of a unified data warehouse.
Data management and storage should also be on the list of important foundational issues. With the deluge of healthcare data that is headed your way, you need to have a clear process for managing how data is stored. This should include automated tools to ensure that current data is quickly accessible and less urgent data is sent to more cost-effective storage. Fortunately, advances in cloud technology and flash storage can help you create a logical, affordable data storage system that is scalable for future growth.
These are not exciting tasks, and the short-term results won’t necessarily wow your clinicians, but they are necessary to enable the exciting analytic projects that will transform the U.S. healthcare landscape.
As with any building project, someone has to focus on the foundation while someone else focuses on the overall system. These tasks are both crucial to your future analytics ability, and they require different skill sets. That’s why it makes sense to have one team focused on (and accountable for) jump-starting your analytics program while another team focuses on (and is accountable for) building a foundation that will last. That will ensure that both jobs get done quickly and thoughtfully.