Diagnostic Accuracy and Clinical Potential for the Strengths and Weaknesses of Attention-Deficit/Hyperactivity Symptoms and Normal Behaviors Scale (SWAN)

Abstract

Background: In theory, the continuum of attention-deficit hyperactivity disorder (ADHD) is normally distributed within the general population. Unlike many existing ADHD rating scales, the Strengths and Weaknesses of ADHD–Symptoms and Normal Behavior (SWAN) scale is theoretically designed to capture positive and negative variations in symptoms along this continuum. Nonetheless, despite its use in several research studies, the SWAN has received little attention to its clinical utility. Most studies that found the SWAN to have high validity and diagnostic accuracy for ADHD screening in children have been conducted in community/nonclinical samples. It is still unclear in clinical populations whether the SWAN captures a similar underlying construct relative to commonly used ADHD rating scales. There is inadequate information about the SWAN’s diagnostic validity compared with other validated ADHD scales to identify ADHD in clinic-referred children. Aim: The aim of the current study was to estimate the SWAN’s convergent validity with the a common ADHD rating subscale, the Hyperactivity–Inattention (HI) of the Strengths and Difficulties Questionnaire (SDQ/HI). A second aim of the study was to compare the concurrent validity of the SWAN with that of the SDQ/HI, as measured by correlations with scores of a measure of ADHD-related functional impairment, the Impairment Rating Scale (IRS). A third aim was to examine the diagnostic validity of the SWAN to differentiate between children with and without an ADHD diagnosis. A fourth aim was to compare the diagnostic validity of the SWAN and SDQ/HI to differentiate between children with and without an ADHD diagnosis. Methods: This study used a cross-sectional cohort design to assess and compare the performance of two ADHD rating scales for identifying cases of ADHD. The study utilized archived data of 357 children who visited a mental health outpatient clinic at a children’s hospital, specializing in ADHD and related disorders, in the northwest United States. One parent/caregiver completed the three rating scales, an average of 3 days before their child’s first visit to the clinic for a comprehensive evaluation for ADHD. Bivariate correlation analyses were used to examine the convergent and concurrent validity among the current study rating scales. Receiver operating characteristic (ROC) curve analyses were used to determine the SWAN and SDQ/HI cutoff scores and compare their performance in identifying cases with ADHD from cases with other disorders in this sample. Results: Regarding convergent validity, the SWAN total and SDQ/HI correlation was large and significant (r = .68). Likewise, the correlation between the SDQ/HI and the SWAN Hyperactivity/Impulsivity subscale (SWAN HI) was large and significant (r = .65). The correlation between the SDQ/HI and the SWAN Attention Deficit subscale (SWAN AD) was significant but smaller (r = .49). For concurrent validity, the correlation coefficients of the SWAN total and the SWAN subscales were larger than the correlation coefficient values with the IRS items compared to the SDQ/HI. Results from the ROC curves and Area Under the Curve (AUC) of the SWAN total, the SWAN AD and SWAN HI (AUCs =.7, .69 and .66, respectively) showed good to useful abilities in differentiating between children with and without an ADHD diagnosis. The ROC curves analyses that compared the SWAN total and the SWAN HI subscale as well as the SWAN AD subscale with the SDQ/HI subscale when differentiating between the ADHD and non-ADHD groups indicated that the SWAN total had the largest AUC among all scales. However, the results from the pairwise comparison of the ROC curves with the critical value for significance for the AUC set at (p =.05) showed no significant differences among the AUCs of the SWAN total, the SWAN AD subscale, and the SDQ/HI. The discriminant abilities of the SWAN AD and the SWAN HI were slightly lower than the SWAN total. The difference between the SWAN total and the SWAN HI AUC was significant (p = .023). Conclusion: The SWAN total and subscales showed good to useful abilities for classifying ADHD cases from non-ADHD clinic-referred children with other psychiatric disorders. While the SWAN demonstrated low specificity in the current study, these results might be specific to the current sample from an ADHD tertiary clinic, wherein the majority of referred cases were complex with significant psychiatric comorbidity, diverse medical and treatment history. Findings from comparing the diagnostic validity of the SWAN to SDQ/HI yielded no significant difference in their performance in identifying ADHD cases in the current sample. Nonetheless, the SWAN has advantages over SDQ/HI for describing aspects of variations (strengths) in attention and activity abilities.