Improving the Quality and Findability of Open Medical Information on the Web: Medical Wikidata and the FindMeEvidence Search Engine



Matthias Samwald* Matthias Samwald*, Medical University of Vienna, Vienna, Austria
Veronika Stefanov, Vienna University of Technology, Vienna, Austria
Georg Petz, Vienna University of Technology, Vienna, Austria
Claus-Dieter Volko, Medical University of Vienna, Vienna, Austria
Allan Hanbury, Vienna University of Technology, Vienna, Austria


Track: Research
Presentation Topic: Search, Collaborative Filtering and Recommender Technologies
Presentation Type: Rapid-Fire Presentation
Submission Type: Single Presentation

Building: Sol Principe
Room: C - Almudaina
Date: 2014-10-09 02:00 PM – 02:45 PM
Last modified: 2014-09-03
qrcode

If you are the presenter of this abstract (or if you cite this abstract in a talk or on a poster), please show the QR code in your slide or poster (QR code contains this URL).

Abstract


Background: The World Wide Web has become an important source of information for medical practitioners, and there is evidence that freely available, well-designed medical search engines could have a positive impact on the quality of health care. Furthermore, several independent surveys in the recent past have shown that Google and Wikipedia are heavily used by doctors to find medical information.

Objective: This work has two major aims: First, to improve the content available in Wikipedia by adding drug safety information to the recently added data-backend of the system, Wikidata. Second, to complement currently available search engines for medical content by developing FindMeEvidence, an open-source, mobile-friendly medical search engine. FindMeEvidence combines ease of use with a focus on information quality when accessing both ‘traditional’, scientific documents as well as collaboratively edited information sources such as Wikipedia.

Methods: We added essential drug safety information (such as critical drug-drug interactions) to Wikidata and formulated strategies for automatically integrating the Wikidata content into different language versions of Wikipedia. We developed a prototype of the FindMeEvidence search engine based on selected content from PubMed, Medscape, Merck Manuals, Guideline.gov and other relevant web resources including Wikipedia. Several measures were taken to improve the quality of search results and to minimize the time needed to find appropriate answers, such as giving users the ability to restrict content sources to focused subsets, displaying key findings from clinical studies directly in the search result view, or flagging cues of untrustworthy content or recent vandalism on Wikipedia pages.
We conducted a preliminary comparative evaluation of FindMeEvidence. A list of 36 medical queries was created, and the information needs of each query were written down. Then, each query was submitted to FindMeEvidence as well as to the TRIP Database search engine. The search engines were scored based on their capabilities of quickly returning relevant information for these queries. For each result list, only the first five top-ranked results were analysed. If any of these five results contained information that met the predefined information needs of the query, one point was added to the score of the search engine; if none of the 5 results contained such information, no point was added.

Results: In the preliminary evaluation, FindMeEvidence proved to be competitive with TRIP Database, an established search engine for evidence-based medicine. A public interface of the current system is available at http://FindMeEvidence.org/, the underlying source code is openly available.

Conclusions: We highlighted automated mechanisms for improving medical content on Wikikpedia, one of the most popular medical information sources on the web for medical doctors and patients alike. FindMeEvidence demonstrated its potential to become a useful addition to the digital toolset available to medical professionals. Since FindMeEvidence is fully open-source, its codebase can act as a starting point for the creation of highly-efficient, customized search engines that can be deployed locally and can be tailored to local needs.




Medicine 2.0® is happy to support and promote other conferences and workshops in this area. Contact us to produce, disseminate and promote your conference or workshop under this label and in this event series. In addition, we are always looking for hosts of future World Congresses. Medicine 2.0® is a registered trademark of JMIR Publications Inc., the leading academic ehealth publisher.
Creative Commons License
This work is licensed under a Creative Commons Attribution 3.0 License.