Assessing the psychometric properties of the Symptom Checklist Five-Items (SCL-5) instrument in a clinical setting for treating intimate partner violence

Abstract

Research suggests that the Symptom Checklist Five Items (SCL-5) instrument is a reliable and valid scale for assessing mental health in a range of populations. Yet, to date, the SCL-5 has not been validated in a population of perpetrators of intimate partner violence (IPV) and their partners. Hence, the current study aimed to examine the psychometric properties of the SCL-5 within a clinical IPV context characterized by elevated mental health symptoms. This validation study is part of a randomised controlled trial evaluating a group intervention for men voluntarily seeking treatment for IPV in Norway. The study was conducted in accordance with the guidelines Consensus-based standards for the selection of health measurement instruments (COSMIN). We evaluated floor and ceiling effects, internal consistency and estimated the area under the curve to determine optimal diagnostic sensitivity and specificity. Additionally, we explored scale responsiveness using data from a subsample of participants. The highest percentage of respondents with the lowest possible score (value 1) across the 5 items was 8.7%, indicating no floor effect assuming the 15% criterion. The highest percentage with the highest possible score (value 4) across the 5 items was 1.9%, indicating no ceiling effect either. Cronbach’s alpha was 0.86 for the total sample ( N  = 161), 0.83 for perpetrators only ( n  = 118) and 0.89 for partners only ( n  = 44). The area under the curve was 94.6% (95% CI 0.913-0.978) for the total sample, 93.5% (95% CI 0.893-0.977) for the perpetrators and 96.9% (95% CI 0.922–1.017) for the victims. Regarding responsiveness, we set a criterion of a statistically significant reduction in mean scores over time as a test of the adequate responsiveness of the SCL-5 from the baseline to the one-year follow-up point. As predicted, it fell from 1.98 to 1.74 and the mean change was significant ( t −4.021, CI −0.360 to −0.121; p  < 0.001). The mean total SCL-5 score decreased in 54% of cases (n  = 43). The results provide evidence that the SCL-5 has adequate psychometric properties in a Norwegian sample of perpetrators and victims of IPV. Not applicable.