Development and validation of the Benzodiazepine Hypnotics Withdrawal Symptom Scale (BHWSS) based on item response theory

Abstract

This study aimed to refine and validate the Benzodiazepine Hypnotics Withdrawal Symptom Scale (BHWSS). The 12-item prototype version of the BHWSS was administered to a sample of 346 patients with chronic insomnia (161 males and 185 females, mean age: 52.8 ± 16.6 years) who had been taking hypnotics (benzodiazepines [BZDs] or BZD receptor agonists) for at least 3 months. The item information curve indicated that two of the 12 BHWSS items should be excluded. As a result of analyzing the 10-item version of the BHWSS (revised-BHWSS), the contribution rate in the case of the factor 1 was 0.49, Cronbach's α was 0.90, and the reliability coefficient ω was 0.91. An analysis of the item information curve for the revised-BHWSS indicated that the information amount per item increased from 3.90 for the original 12-item BHWSS to 4.37 for the 10-item revised-BHWSS. The receiver operating characteristic curve indicated that 6.5 points on the revised-BHWSS was the most appropriate cutoff for estimating moderate or severe withdrawal symptoms using the Benzodiazepine Dependence Self-Report Questionnaire as a reference. These results suggest that the 10-item revised-BHWSS has sufficient reliability and validity for evaluating the severity of withdrawal symptoms after discontinuing BZDs.