VirtualMindTrial: An Intelligent Questionnaire System for Clinical Trail Recruitment

Abstract

The recruitment of human subjects for clinical trials research is a critically important step in the discovery of new cures for diseases. Volunteers are subjected to an elaborate questionnaire process in current recruitment methodologies. Although the questionnaire process is extremely important in clinical trial recruitment, it is inefficient due to redundancy and lack of a systematic approach. Ideally, questionnaire generation and implementation must be guided by intelligent heuristics that minimize redundancy and inconsistency. In this thesis, an intelligent approach to questionnaire flow called VirtualMindTrial is proposed. Given a set of textual inclusion/exclusion clinical trial eligibility criteria and data available from diverse sources such as Microsoft HealthVault, VirtualMindTrial is able to 1) filter known criteria, 2) add associative criteria based on selected criteria, 3) form a neighborhood of patients who satisfy similar criteria, and 4) generate a dynamic questionnaire flow for screening patients. The questionnaire has been implemented using a visual 3-D environment to help volunteer subjects experience a realistic screening process. Experimental results demonstrate the effectiveness of our system in terms of dynamic questionnaire flow generation and in enhancing the user experience with virtual worlds. A visual prototype system has been developed as part of the thesis to illustrate the enhanced efficiency and quality of screening patients with psychiatric disorders for clinical research.