Assisted curation of clinical trials using text mining tools

Abstract

The amount of biomedical literature is growing rapidly, having nearly doubled the last decade. As the literature grows, it is becoming increasingly difficult for curators to keep up, as manual curation is a time-consuming process. Past research has indicated that computer-assisted curation can speed up this process considerably. Our project aims to create a data mining approach using the ClinicalTrials.gov database as our source of data. Using external dictionaries and resources, new data will be added to the published trials by expanding on the pre-existing data tags and applying various text mining tools. As end product, we will develop a web-based curating tool that allow users to post more advanced queries for clinical trials than the currently existing interface at clinicaltrials.gov, as well as curate and add/edit manual annotations was developed. The system is available at http://invitro.titan.uio.no/clinicaltrials