Public Resources for Detecting Mentions of Medical Data in Russian Social Media

This paper covers the results of a comparative analysis of the effectiveness of passive and active data collection methods for the purpose of extracting mentions of adverse drug reactions (ADRs) in Russian. In terms of their effectiveness, two systems of data collection were compared: a data mining system for gathering post and comment text content from social media, and an experimental chatbot conversational survey, integrated into a thematic community and targeted at collecting ADR reports. The study was conducted on VK, a Russian social network, on a community dedicated to the discussion of user experiences with taking drugs for treating mental illnesses. A comparative analysis of the comprehensiveness of data obtained by the passive method and the chatbot was carried out. The results show that an active information collection system allows subsequent information processing to be performed more effectively. Based on the results, areas for further development of conversational surveys for medical research were identified.