Discoverable and Governable CODefs

By Michael Buck, Navidence SVP of Knowledge Engineering 

In our last post, One Value Set Reference to Rule Them All?, we discussed that in most cases it is neither desirable nor accurate to limit a gold standard value set to one reference. There are however two big problems with leveraging standard CODefs. First, it is hard to discover what existing CODefs are out there in the research literature or standards organizations. Second, even within large research organizations CODefs are not usually governed as a strategic asset, creating an increased likelihood of multiple reference standards being developed.

Navidence addresses the “discover” need at multiple levels.

To start we leverage an industry-leading AI-driven literature review solution from Nested Knowledge to develop and maintain our indication libraries. Sherpa also makes it easy to discover what existing CODefs are already in our libraries and it gives researchers the ability to compare and suggest which value sets are the most useful for a particular need.

Navidence addresses the “govern” need through its robust implementation of lifecycle, workflow and provenance management.

Sherpa already allows us to extensively document each information asset including value sets, elements, statements, libraries, and studies with descriptions, references, and notes. This ensures that we can find out which resource was used to develop each asset, to justify study design decisions and to provide comparisons between assets when designing new content. In 2025, Sherpa will be expanding its features further to allow users to promote vetted content created by an individual user/team to be a reusable asset by entire departments or organization. We will be creating a more extensive version tracking system to allow users to be able to snapshot asset development at various stages which is important for lifecycle governance as well as possible regulatory review. Lastly we will be developing more advanced provenance tracking and governance mechanisms to ensure we can group related content together maximizing the use of original research-sourced value sets as well as derivative value sets with up-to-date codes.

To put this in perspective our actively growing CODEF library used by multiple life science organizations has just in its value set area nearly ~1,000 discrete concepts defined by ~2,500 related value-set groups from 28 different domestic and international code systems.

Even if you are using your own processes today to manage your company’s CODefs remember to ask yourself if your content is both “discoverable” and “governable”.