A Czech version of the Overall Anxiety Severity and Impairment Scale (OASIS): standardization and psychometric properties

Abstract

Background The Overall Anxiety Severity and Impairment Scale (OASIS) is a transdiagnostic measure that assesses severity and impairment associated with anxiety disorders. However, its psychometric properties were primarily examined in English-speaking or western countries. Therefore, this study aims to examine its psychometric parameters in the Czech Republic. Methods A large representative sample (n = 1738), a clinical sample (n = 57) and a retest sample (n = 20) were used. In addition to the OASIS, conventional measures of anxiety, depression, personality traits, self-esteem, life satisfaction, and other scales were also administered. Moreover, we examined the latent structure, reliability, validity, and the cut-off score for the Reliable Change Index (RCI) and the Clinically Significant Change Index (CSI). Results Higher anxiety was found in females, religious non-church members, and students. The Confirmatory Factor Analysis supported the adequate fit of the unidimensional solution: x2(4) = 3.20; p < 0.525; CFI = 1.000; TLI = 1.000;RMSEA = 0, SRMR = 0. The measurement equivalence examination indicated that the OASIS measures anxiety invariantly between males and females. The validity of the OASIS was supported by positive associations with neuroticism, depression, perceived stress, guilt, shame, and the established anxiety measures. The internal consistency was excellent (Cronbach’s alpha = 0.96, McDonald’s omega = 0.96). The test-retest reliability was acceptable (r = 0.66). The cut-off for the CSI is 6 and the RCI is 5.32. Conclusions The OASIS represents a reliable and valid instrument for assessing anxiety in adults. Due to its shortness, excellent psychometric properties, and percentile norms, it is especially useful for short and accurate screening of anxiety and mapping therapeutic changes in clinical practice.

Psychometric evaluation and Rasch analyses of the German Overall Anxiety Severity and Impairment Scale (OASIS-D)

“…It has been argued that covariance between items 1 and 2 is theoretically plausible because the responses are partly contingent on each other (i.e., if frequency of anxiety is rated zero, severity should also be rated zero) 25 , 28 , 34 . Following this explanation, a method effect induces a so-called response dependency between these two items 74 .…”

“…The authors have interpreted their findings as initial support for measurement invariance with respect to gender but recommended similar analyses in broader samples (e.g., primary care patients) since they have found a potential (nonsignificant) instance of DIF. Two other studies have investigated the Czech version of the OASIS for measurement invariance by means of multi-group CFAs 34 , 35 . In both studies, evidence was found that men and women responded similarly to the OASIS.…”

“…Recently, it has been adapted to yield a caregiver-report of youth anxiety and interference 29 . The OASIS has been translated and validated into Japanese 30 , Dutch 31 , Spanish 32 , Persian 33 , and Czech 34 , 35 .…”

“…Exploratory and confirmatory factor analyses support a unidimensional factor structure. However, in all but one study 33 an acceptable model fit could only be reached by allowing the error terms of the first two items to be freely estimated (i.e., a single-factor-model with correlated residuals of items 1 and 2) 24 , 25 , 28 , 30 , 32 , 34 – 36 . In two studies, the best model fit was obtained by additionally allowing a correlation between the residuals of items 1 and 3 27 , 30 .…”

“…This assumption is implicitly made whenever test results between different groups of respondents are compared 37 . Using confirmatory factor analyses (CFAs), three studies consistently found measurement invariance of the OASIS with respect to gender 28 , 34 , 35 . To our knowledge, no study has yet analyzed the OASIS for measurement invariance or local item independence within the framework of probabilistic test theory.…”