Psychometric evaluation of the lithuanian version of the brief negative symptoms scale – a process in progress

Abstract

Introduction. The Brief Negative Symptoms Scale (BNSS) is a semi-structured negative symptom assessment interview comprising 13 items and six subscales [1]. BNSS is gaining increasing acceptance as the first choice tool for negative symptom assessment [2]. In our pursuit to enhance the quality of mental health care in Lithuania, our study group aimed to translate and validate the Lithuanian version of BNSS. Methods. We performed a double translation from English into Lithuanian. The back-translation version of BNSS was corrected according to the comments of the authors. Two experienced professors of psychiatry performed clinical validation and approved the final version. Four clinicians conducted psychometric validation in a sample of patients diagnosed with Schizophrenia, schizotypal, and delusional disorders according to ICD-10. The psychometric tools used were BNNS, Positive and Negative Symptoms Scale (PANSS), Montgomery Asperger Depression Rating Scale (MADRS), and Calgary Depression Scale for Schizophrenia (CDSS). We calculated the reliability of the 13 items and six subscales of BNSS. Convergent and discriminant validity were calculated by applying BNSS in clinical practice with other psychometric tools for negative symptoms assessment and tools for positive, depressive symptoms, and cognitive deficits. We calculated the convergent and discriminant validities using Pearson and Spearman correlations. Results. We have included 84 patients from an inpatient ward diagnosed with schizophrenia so far. Great internal consistency was observed for the 13 items (alpha = 0,944) and the six subscales (alpha = 0,874) of BNSS. Good convergent validity is illustrated by strong Pearson correlations with the PANSS negative subscale (r= 0.74, p<0,001) and the PANSS Marder negative factor (r= 0,726, p<0,001). Adequate discriminant validity is shown by a weak Pearsons correlation with PANSS positive subscore (r= 0,238, p=0,028), non-significant correlation with PANSS Marder positive factor (r= 0,207, p = 0,057), Marder depression-anxiety factor (r= 0,35, p = 0,749) and by a non-significant Spearman's correlation with CDSS total score (r= 0,163, p=0,138) and weak correlation with MADRS total score (r= 0,23, p=0,034). Discussion. We have reported data from an ongoing validation study. We plan to include up to 200 patients; therefore, the results might change. The weak positive Pearson's correlation between the BNSS total score and PANSS positive subscore might be present because of known content validity flaws of the original PANSS positive subscore [3]. This is supported by the non-significant correlation between the BNSS total score and the PANSS Marder positive factor, which has reduced content validity flaws compared to the original PANSS positive subscore. The weak positive correlation between BNSS and MADRS total scores is in line with findings from other researchers [4]. Also, the sample size is too small to perform factor analysis, which would enhance the psychometric validation process. Conclusions. The Lithuanian version of BNSS is proving to be a psychometrically sound tool for the assessment of negative symptoms. Further data collection is needed to solidify the validation process.