Validity, reliability, and measurement invariance of an adapted short version of the HIV stigma scale among perinatally HIV infected adolescents at the Kenyan coast

Abstract Background There is a dearth of instruments that have been developed and validated for use with children living with HIV under the age of 17 years in the Kenyan context. We examined the psychometric properties and measurement invariance of a short version of the Berger HIV stigma scale administered to perinatally HIV-infected adolescents in a rural setting on the Kenyan coast. Methods A cross-sectional study was conducted among 201 perinatally HIV-infected adolescents aged 12–17 years between November 2017 and October 2018. A short version of the Berger HIV stigma scale (HSS-40) containing twelve items (HSS-12) covering the four dimensions of stigma was evaluated. The psychometric assessment included exploratory factor analysis, confirmatory factor analysis (CFA), and multi-group CFA. Additionally, scale reliability was evaluated as internal consistency by calculating Cronbach’s alpha. Results Evaluation of the reliability and construct validity of the HSS-12 indicated insufficient reliability on three of the four subscales. Consequently, Exploratory Factor Analysis (EFA) was conducted to identify problematic items and determine ways to enhance the scale’s reliability. Based on the EFA results, two items were dropped. The Swahili version of this new 10-item HIV stigma scale (HSS-10) demonstrated excellent internal consistency with a Cronbach alpha of 0.86 (95% confidence interval (CI) 0.84–0.89). Confirmatory Factor Analysis indicated that a unidimensional model best fitted the data. The HSS-10 presented a good fit (overall Comparative Fit Index = 0.976, Tucker Lewis Index = 0.969, Root Mean Square Error of Approximation = 0.040, Standardised Root Mean Residual = 0.045). Additionally, multi-group CFA indicated measurement invariance across gender and age groups at the strict invariance level as ΔCFI was ≤ 0.01. Conclusion Our findings indicate that the HSS-10 has good psychometric properties and is appropriate for evaluating HIV stigma among perinatally HIV-infected adolescents on the Kenyan coast. Further, study results support the unidimensional model and measurement invariance across gender and age groups of the HSS-10 measure.