A-196 The Texas Postconcussion Symptom Inventory (TPSI): An Exploratory Factor Analysis and Reliability Analysis

Abstract

Abstract Objective Self-report scales are commonly used to evaluate non-specific symptoms following concussion. While several scales have been developed, few were created using a systematic process and most contain several ambiguous items that may be misinterpreted. To address this, a new theoretically-based, multidimensional measure was designed to assess Cognitive, Neuropsychiatric, and Somatic symptoms associated with concussion. This study used sophisticated psychometric techniques to develop the Texas Postconcussion Symptom Inventory (TPSI) and establish initial reliability and validity. Method Because concussion symptoms are non-specific, a pool of 76 potential items was developed and administered to a diverse clinical sample (N = 350) that included patients with concussion, epilepsy, and dementia. Polychoric correlations were utilized to remove items based on poor fit/multicollinearity and an exploratory factor analysis (EFA) with an Oblimin rotation was used to determine factor structure. Results A three-factor model best fit the data, and represented Cognitive, Neuropsychiatric, and Somatic domains as designed. Ten items were discarded, resulting in a total of 66 items. The model explained 48.5% of the total variance and contained adequate sampling (Kaiser-Meyer-Olkin measure =0.92) and sufficient item correlations (Bartlett’s Test of Sphericity, p < 0.05) for EFA. All three factor structures displayed high internal consistency (Cronbach’s α > 0.88). Conclusions The TPSI is a brief, multidimensional measure with evidence of strong internal consistency and reliability as well as distinct Cognitive, Neuropsychiatric, and Somatic symptoms associated with concussion. Future research will investigate its convergent and divergent validity in concussion as compared to existing popular symptom measures.