Text Mining and copy right laws : a case for change in the medical research field

Background

Mid 2011, the research team at the Department of Ambulatory Care and Community Medicine, University of Lausanne, decided to start a project on a new research topic : Shared Decision Making (SDM). The objective was to identify publication trends about SDM in 15 major internal medicine journals over the last 15 years. It was decided to use a "text mining" approach to systematically  review all the articles published in these main journals and automatically search for the different occurrences of SDM. The research team turned to the medical library for help to collect the electronic publication files.

Methods

The software applications used in text mining allow to search through large sets of unstructured texts. The results are then clustered to extract trends, facts and build new knowledge. In order to work consistently, all the text sources should be aggregated on one single local platform. However electronic scientific publications are currently stored as licensed materials on publishers' sites. Bulk download of thousands of articles are not commonly permitted by licences. The library teamed up with researchers in order to get all the permissions to compile the files for research purpose.

Results

Contacts with publishers and exchange of information over the reseach project were particularly cumbersome and time consuming. After 6 months, only 5 out of 15 publishers had agreed to grant an licence extension that gave the right to systematically download the articles for research purpose. Permission was usually granted under one main condition : all downloaded content for text based analysis should be destroyed when research is complete.

Conclusion

Due to the rapidly expanding body of electronic biomedical literature, text mining should become an essential process for research in the medical field. To allow this new research method to expand, copyright law and licences for electronic access have to be amended and new competences have to emerge in libraries and research centers.