eRA currently is building a prototype for using knowledge management (KM) technology to locate qualified peer reviewers for research proposals submitted to the NIH. KM refers to the family of text-mining tools for examining vast quantities of data to identify patterns and establish relationships. Given the exponential rate at which the world’s information is growing (estimated at 10¹⁸ new bytes yearly), KM holds great promise for optimizing the knowledge assets of an organization.

Dr. Richard Morris, eRA advocate for knowledge discovery, recently presented KM concepts at the Third Annual eRA Symposium on April 30 (see article above). He also outlined the process by which eRA will apply KM techniques to identify candidate reviewers. Basically, eRA will “fingerprint” (profile) each incoming research proposal using the National Library of Medicine Medical Subject Headings (MeSH^®) Thesaurus. MeSH is a database of hierarchical (e.g., “ankle” is subordinate to “anatomy”) and cross-referenced topics (e.g., “for vitamin C, see ascorbic acid”), which permits searching at various levels of specificity. The fingerprint for each research plan will consist of the most appropriate MeSH terms.

The other sources of input are databases of reviewer biosketches. These can be both internal and external to NIH. Once the biosketches are fingerprinted, locating subject-matter experts consists of comparing proposal profiles with expert profiles to produce the best matches. Thus, KM will enable NIH to transform raw text (research plans and biosketches) into useful information.

eRA also can use the same KM techniques to ascertain similar research plans (from proposed and already funded research) and to create opportunities for scientific collaboration by identifying experts with common interests. These and other KM applications are expected to improve NIH performance and efficiency by facilitating data sharing, verifying facts, informing decisions, shortening grant cycle times, and reducing costs. NIH also hopes to realize qualitative gains.

The pre-production phase of the KM reviewer identification module will continue through August. If you are interested in becoming an early adopter, contact Richard Morris.