Psychometric and Clinical Evaluation of Schizophrenia Remission Criteria in Swedish Outpatients with Psychotic Disorders

Abstract

Abstract Background: Psychotic disorders including schizophrenia have debilitating effects on health and functioning. In recent years, symptomatic remission has emerged as an increasingly viable treatment goal. The Remission in Schizophrenia Working Group criteria (RSWG-cr) are based on 8 items from the Positive and Negative Syndrome Scale (PANSS-8). The criteria, frequently used in clinical and research settings, lack Swedish validation. Objectives: To evaluate the psychometric properties of PANSS-8, and to examine the clinical validity of RSWG-cr in a Swedish outpatient population. Methods: Cross-sectional register data was collected between 2016–2019 from outpatient psychosis clinics in Gothenburg, Sweden. Firstly, confirmative and exploratory factor analyses were performed on PANSS-8 data (n=1744) to assess the scale and subscales’ psychometric properties. Internal reliability was evaluated using Cronbach’s alpha. Second, 649 patients with sociodemographic and clinical data registered within the same 7-day period were classified according to RSWG-cr. Remitted and non-remitted patients were compared in terms of clinical and sociodemographic characteristics. Depending on the nature of data, parametric (e.g., t-tests) and non-parametric (e.g., Mann-Whitney U test) analyses were used. Lastly, odds ratios (OR) were estimated using binary logistic regression to assess the independent contribution of each variable to remission status. Results: PANSS-8 showed good internal reliability (alpha = 0.85). A three-dimensional model reflecting psychoticism, disorganization, and negative symptoms showed good fit in confirmatory factor analysis. Of 649 patients, 356 (55%) were in remission according to RSWG-cr. Remitted patients were significantly more likely to live independently, have their own income, be non-smokers, not currently on antipsychotic medication, and have a recent health interview and physical examination registered (all p < 0.01). Binary logistic regression found increased likelihood of remission among patients living independently (OR = 1.98), having their own income (OR = 1.89), obese patients (OR = 1.61), and patients with a recent physical examination (OR = 1.56), and decreased likelihood among patients on antipsychotic medication (OR = 0.41). Conclusions: PANSS-8 as a whole was internally reliable, and remission according to RSWG-cr was associated with significant clinical variables in a naturalistic Swedish outpatient sample. Future longitudinal studies should be conducted to better assess the direction of these relationships.