RÉSUMÉ
INTRODUCTION: Irritability, a common neuropsychiatric symptom in Huntington's disease (HD), lacks a standardized measurement. The Irritability Scale (IS), tailored for HD, has patient and informant versions, but variable interrater agreement has been reported frequently in previous studies. To enhance the clinical utility of the IS, this study aimed to identify the most reliable components estimating the underlying construct and develop a shortened version for time-limited contexts. METHODS: Participant and informant/observer concordance and the relationship of individual items to the complete IS scale were assessed. The short-form (SF) items were selected based on interrater agreement, exploratory factor analysis (EFA), and Item Response Theory (IRT) analysis results. Pair-wise correlation and covariance models were used to examine how SF predicted total IS score in 106 participants from the STAIR (Safety, Tolerability, and Activity of SRX246 in Irritable Subjects with Huntington's Disease) trial. Item Response Theory (IRT) analysis was used to evaluate the range and function of the selected items. RESULTS: IS interrater agreement was statistically significant (r = 0.33, p = .001). In combination with EFA factors and IRT analyses, five items were identified that showed good reliability and performance in differentiating levels of irritability. CONCLUSION: The proposed 5-item SF IS provided a reliable measure of the full scale and may be less burdensome for use in a clinical setting.
