Psychometric properties of the spinal cord injury-quality of life (SCI-QOL) Resilience item bank in a sample with spinal cord injury and chronic pain.

Abstract

To describe the psychometric properties (e.g., data distribution characteristics, convergent/discriminant validity, internal consistency reliability, and test administration characteristics) of the spinal cord injury quality of life measurement system (SCI-QOL) Resilience item bank delivered as a computer adaptive test (CAT) in a sample of individuals with chronic pain and spinal cord injury (SCI). Descriptive statistics were calculated to investigate variable data distribution characteristics. Correlation analyses were conducted for convergent and discriminant validity. Item response theory-derived reliability was calculated for the SCI-QOL Resilience CAT. One hundred thirty-three adults with SCI (N = 133; 73.5% male, 26.5% female) were enrolled. Sample mean T score on the SCI-QOL Resilience measure was 48.40, SD = 8.60 (min = 29.4; max = 70.0). The CAT administered between 4 (most common, 41.4% of cases) and 12 (9% of cases) items with the Mean#items = 5.73, SD = 2.45. The SCI-QOL Resilience CAT scores were normally distributed, with very low ceiling (0%) and floor (3%) effects. The SCI-QOL Resilience CAT had a reliability of 0.89, and the mean length of time for respondents to complete the SCI-QOL Resilience CAT was 44.34s. SCI-QOL Resilience CAT validity was supported by significant moderate correlations with pain acceptance, depressive symptoms, pain catastrophizing, positive affect and well-being, and pain interference (convergent validity) and small non-significant correlations with age, sex, injury level, pain intensity, mobility level, and years since injury (discriminant validity). The SCI-QOL Resilience CAT demonstrated good convergent and discriminant validity. The CAT administration characteristics were impressive: With few items (low response burden), the scale achieved good reliability.