Development of an Abbreviated and Statistically Robust Instrument Using Exploratory Factor Analysis: Example from Neurology Research (303584)

Keywords: Survey Instrument, Exploratory Factor Analysis

Research in academic medical centers offer opportunities to collaborate on clinical projects that require novel application of both common and uncommon statistical methods. Lengthy survey instruments can be detrimental to clinical research due to multiple factors; 1) instrument can be time consuming which may add burden to patients who are critically ill and also may impact accuracy of responses and thus sacrifice data quality, 2) scoring algorithm if not refined, may be complicated which may result in obtaining inconsistent scores, 3) some questions may not be meaningful or could be redundant. A redesign of an existing lengthy survey instrument to a simplified, more time-efficient, statistically more robust, yet comprehensive tool to assess and grade symptoms relevant to particular clinical condition can have broad applications in clinical research and practice. This presentation will demonstrate use of exploratory factor analysis as a statistical tool to reduce dimensionality, i.e., reducing the number of items in a survey instrument. Example from neurology research will be used for illustration.