Goal: To automate the referral of most incoming grant applications to Integrated Review Groups, Scientific Review Groups and NIH Institutes/Centers through electronic extraction of referral requests contained in application cover letters, computerized text analysis, and human referral experts for unusual situations.

Expected Benefits: Automated referral will—

·         Support the Shortening the Review Cycle Initiative by reducing the time required to refer applications from several weeks to several days. Consequently, review meetings will be conducted sooner after receipt dates than is currently feasible.

·         Promote consistency and transparency in the referral process.

·         Reduce the cost of the referral process.

Status: Feasibility studies began in February 2006, and the primary technical approach for computerized text analysis was selected in October 2006. Automated extraction of referral requests from cover letters was instituted in December 2006 and a proof of concept tool displaying extracted cover letter requests and referral suggestions based on computer algorithms was deployed to staff in the Division of Receipt and Referral in February, 2007. A decision support and workflow tool based on these technologies was also provided to all Scientific Review Officers in February 2007. An Operational Test was conducted in October, 2007, in which the reliability, throughput and accuracy of the system was demonstrated in a simulation of the automated referral of new R01s submitted with review assignment requests.

Schedule for Implementation: Realization of the long-term goal of fully automating most referrals is dependent upon further computer development planned for FY 2008 and FY 2009. A key development for FY 2008 will be integration of the software with IMPAC II.