Matching NIH grants to ClinicalTrials.gov protocols

Wednesday, October 26, 2011 — Poster Session III
10:00 a.m. – Noon	Natcher Conference Center	CIT	CLIN/CULT/AGING/DISPREV-16

Authors

• W Lau
• M Ji
• K Collie
• M Bukowski
• M Vos
• L Krueger
• C Johnson

Abstract

As clinical trials are an important area of research funded by the NIH, the ability to match ClinicalTrials.gov protocols with their corresponding funding sources has many potential applications to tracking and evaluation of NIH research investments. Unfortunately, existing linkages between ClinicalTrials.gov and IMPAC II are incomplete and inadequate. In a sampling of NIH-sponsored clinical trials from 2006 to 2010, grant numbers were specified in only 20% to 25% of all cases. At present, to find the corresponding grants for the remaining 75% to 80% of protocols, a time-consuming exercise of manually matching investigators, institutions, titles, and even text would be required. We have devised a system that automatically matches IMPAC II grants to ClinicalTrials.gov protocols. In addition to matching fields directly when possible, the system also uses edit distance to match titles and a similarity score between concepts extracted from text to map grant abstracts with protocol descriptions. In a preliminary experiment, we have been able to find the matching IMPAC II grant for approximately 80% of protocols. We are in process of creating a resource to make these mappings available for NIH staff.

back to top