What’s “atomic knowledge,” and how can it make pharmaceutical labeling easy?

by Jared Kimble

The opportunity to transform the way pharmaceutical companies do business through digital technologies can be felt across the entire life cycle.

One such area is labeling, which has long struggled with issues related to document ownership and control, maintaining accurate status, having a global view of the processes, reviews and approvals, and having a big-picture perspective to improve efficiency.

For example, labeling documents are typically created in English and as they are distributed to affiliates around the world, these documents need to be translated into the local language and adapted to comply with local laws and regulations. Each region may also have its own processes with regard to review and approval. Furthermore, the information required for the label may be held in multiple data sources, such as a database, data sheets, or even in various trial documents.

That creates several problems for the business, such as not having central control over those processes and not knowing where each region is in the approval process. When problems arise, the challenge becomes first, how to identify those issues, and second, how to ensure that anything which shows up regionally gets back into the master document or is distributed to other regions that are working on labeling for that product.

This inability to gain a big picture perspective or to clearly visualize changes and when they are made is a source of frustration for the life sciences industry. When a labeling change management process takes place, for example, how do you know how those changes will affect regional labels?

Moving the dial with direct access to knowledge

One of the reasons companies have struggled with lack of oversight can be attributed to the way document knowledge management has traditionally been handled. During the label-creation process, a user may perform a keyword search for relevant information. Document-oriented search engines will return a series of documents, articles, graphics, etc. that associate with that keyword. These documents will have some relevance to what they’re searching for, but users are forced to then sift through and make sense of a lot of information to gather context-sensitive insights.

Having an atomic knowledge base shifts the dial away from document-oriented searching to knowledge-oriented searching.

Atomic knowledge is the process of extraction and “atomization” of the paragraphs, lists, tables, images, etc. from content and assembling these atoms into a highly specialized neural network. Each atom within the network is “normalized” into a standard format to simplify processing and consistent formatting.

By using what is described as advanced context analysis algorithms, the knowledge gathered can be analyzed and graphed into “interatomic” relationships.

Once in the system, cognitive technology, including artificial intelligence and machine learning, can be used to pull together relevant responses to a search. As users dig deeper into that search, they can further refine the criteria and get a response — be that a definition, a table, a reference or something else — and get answers to questions rather than just a group of related documents.

Once a master label has been created, derivations can be created at a local level, with each region using the same knowledge-oriented searches from the atomic knowledge management system to find relevant information, including compliance and procedural requirements for the region. They can again use the same system to carry out component-based authoring specific to the region.

The benefits of atomic knowledge

Many of the problems traditionally faced in labeling — lack of centralized control over content, no oversight into changes across regions, inability to identify problem areas, and process inefficiencies — are addressed with an atomic knowledge approach.

The benefits to the business include speeding up and enhancing process productivity by accelerating access to knowledge. There’s also greater business agility through the ability to update and change processes as needed and to ensure that these are understood and adopted across all affiliates. Central labeling also provides greater understanding of where each affiliate is in the process and can keep track of the status of the review and approval process in each region and ensure that any changes are quickly updated with the local health authorities to meet compliance requirements.

Component-based authoring is a significantly beneficial byproduct of atomic knowledge management. The atoms of data are made available to use in the creation of documents, tables, charts, etc.  When the data in the atom is changed, documents that reference those atoms are also updated to ensure that the data is synchronized and kept consistent everywhere the information is used.

Another real user benefit of atomic knowledge management platforms is that they don’t require any change in tools and applications. Users are still working with the same solutions they’re familiar with, and still using the same processes to carry out searches. What changes is the end results, in other words, the speed, accuracy and efficiency involved in searching for information relevant to labeling and pulling it together. That means there’s no steep learning curve, only better oversight and control.

Jared Kimble head shotJared Kimble has over 14 years of experience in the life sciences industry. His expertise ranges from software design and development to solution architecture where he is currently the offering lead for Life Sciences Regulatory Transformation Services. He leads the management and development effort for many key projects and was involved in several on-site engagements. Before joining DXC, Jared worked as a software engineer developing applications dedicated to providing financial exchange services for banks and financial institutions.


  1. Hi Jared. Interesting blog. Do you have any examples of ‘atomic knowledge management platforms’ that you can share?

Speak Your Mind


This site uses Akismet to reduce spam. Learn how your comment data is processed.