Siloed Systems or Siloed Data?

Categorical funding, insufficient resources, and lack of agency vision keep public health systems isolated and unintegrated – a phenomenon often referred to as “siloed” systems. In a 2007 white paper on the Evolution of Public Health Information Systems I argued that public health agencies needed to think about enterprise-wide systems integration, and several models from distributed, to cooperative, to more centralized were offered for consideration. More recently the action has shifted to interoperable systems (see both the ONC Nationwide Interoperability Roadmap, and my reaction to it in The Interoperability of Things). Interoperability focuses more on the implementation of standards that permit data to flow between systems rather than on combining disparate systems into larger objects to facilitate data sharing.

While this distinction between system integration and system interoperability does have implications for system architects, at the end of the day users are concerned about data. It is more important to focus on siloed data rather than siloed systems as a way to break the impasse that often arises when an agency tries to improve its operational or analytical capabilities. Agencies should consider many different strategies – system redesign, Commercial and Government off-the-shelf (COTS/GOTS) solutions, Master Data Management (MDM) strategies, service-oriented architecture (SOA), Open Source components – but always focus on the impact on the user and ultimately on the user’s experience. And of course as time goes on agencies will want to exchange data with systems outside of public health more and more which requires even more flexibility.