Concurrent Validity and Cross-Validation of the Child SCAT3 PCSS-C Sleep Cluster

Abstract

Objective: The Post-Concussion Symptom Scale (PCSS) has been used to evaluate impaired sleep, despite not being validated for this purpose. The objective of this study was to examine the concurrent validity of the Post-Concussion Symptom Scale-Child (PCSS-C) sleep items/cluster with a validated measure of sleep functioning and the effect of sleep problems on baseline performance in children. Design: The study design was retrospective and cross-sectional. Participants: Participants included male children (n = 80), between 10 to 12 years of age who were enrolled in a comprehensive concussion study and completed pre-season testing. Assessment of Risk Factors: Athletes were divided into 2 groups: (1) Sleep Problems (i.e., defined as PCSS-C sleep cluster ⥠1) and (2) Control (i.e., defined as PCSS-C sleep cluster = 0). Main Outcome Measures: The Child Sport Concussion Assessment Tool 3 (Child SCAT3) and the Sleep/Wake Problems Behaviour scale (SWPB). Results: The reliability coefficients of measures by Ordinal alpha were .62 and .70, respectively. Individual PCSS-C sleep items were moderately associated with the validated sleep measure (i.e., SWPB). A large effect (rs = .58, p < .001) was observed between the SWPB total score and the PCSS-C sleep cluster. Sleep difficulties did not affect the Child SCAT3 Standard Assessment of Concussion (SAC-C) or Balance Error Scoring System (BESS-C) scores. In contrast, the effect of sleep difficulties on symptom reporting was significant. Large effect sizes (d = 1) were found between the sleep groups for symptom total, t(78) = -4.72, p < .001 and symptom severity score, t(78) = -5.31, p < .001. Participants in the Sleep Problems group reported both a greater number of symptoms and higher symptom severity score. Supplementary analyses indicated that sleep difficulties were associated with Post-Concussion Symptom clusters of the PCSS-C. Medium effect sizes were observed between the PCSS-C sleep cluster and the PCSS-C cognitive cluster (rs(80) = .34, p =.002) and PCSS-C somatic cluster (rs(80) = .40, p < .001). Conclusions: If the data is not normally distributed, the Ordinal alpha is a more accurate coefficient for testing reliability. Although the PCSS-C of the Child SCAT3 was never designed to thoroughly assess sleep problems, the sleep items on the PCSS-C may be useful when other sleep measures are not available. Clinical Relevance: The subjective measures of the Child SCAT were all significantly affected by self-reported sleep problems. Sleep problems may be an important confound when assessing post-concussion symptoms.