Testing measurement invariance of the Brief Symptom Inventory across race, sex, and time among adults with cocaine use disorder.

Abstract

The Brief Symptom Inventory (BSI) is a commonly used measure to understand the influence of psychological distress on substance use treatment outcomes. However, few studies have examined whether the BSI functions equivalently across groups to allow for interpretable comparisons among individuals who use substances. The present study evaluated measurement invariance of the full 53-item BSI and the BSI-18 across race, sex, and time among individuals with cocaine use disorder. Data from seven clinical trials evaluating behavioral and/or pharmacological treatments for cocaine use were combined, providing a total sample of 629 participants with Diagnostic and Statistical Manual of Mental Disorders, fourth edition, cocaine abuse/dependence (42.0% Black, 58.0% White; 38.3% female, 61.5% male; Mage = 39.74 years, SD = 9.49). Factor analyses supported an eight-factor model for the BSI and a three-factor model for the BSI-18, largely consistent with recommended scoring guidelines. We did not find support for measurement invariance of either the full BSI or the BSI-18 across race, sex, or time. After model trimming, we were able to identify models for which invariance testing could be conducted. We found evidence for weak factorial invariance of versions of the BSI and BSI-18 by race and versions of the BSI by sex, and strict factorial invariance of versions of the BSI by time and versions of the BSI-18 by sex and time. Findings suggest that caution is warranted when making comparisons across race, sex, or time using the full BSI or BSI-18 among individuals seeking treatment for cocaine use. (PsycInfo Database Record (c) 2026 APA, all rights reserved).