Interformat Reliability of Web-Based Parent-Rated Questionnaires for Assessing Neurodevelopmental Disorders Among Preschoolers: Cross-sectional Community Study (Preprint)

Abstract

Early detection and intervention for neurodevelopmental disorders are effective. Several types of paper questionnaires have been developed to assess these conditions in early childhood; however, the psychometric equivalence between the web-based and the paper versions of these questionnaires is unknown.

This study examined the interformat reliability of the web-based parent-rated version of the Autism Spectrum Screening Questionnaire (ASSQ), Attention-Deficit/Hyperactivity Disorder Rating Scale (ADHD-RS), Developmental Coordination Disorder Questionnaire 2007 (DCDQ), and Strengths and Difficulties Questionnaire (SDQ) among Japanese preschoolers in a community developmental health check-up setting.

A set of paper-based questionnaires were distributed for voluntary completion to parents of children aged 5 years. The package of the paper format questionnaires included the ASSQ, ADHD-RS, DCDQ, parent-reported SDQ (P-SDQ), and several additional demographic questions. Responses were received from 508 parents of children who agreed to participate in the study. After 3 months, 300 parents, who were among the initial responders, were randomly selected and asked to complete the web-based versions of these questionnaires. A total of 140 parents replied to the web-based format and were included as a final sample in this study.

We obtained the McDonald ω coefficients for both the web-based and paper formats of the ASSQ (web-based: ω=.90; paper: ω=.86), ADHD-RS total and subscales (web-based: ω=.88-.94; paper: ω=.87-.93), DCDQ total and subscales (web-based: ω=.82-.94; paper: ω=.74-.92), and P-SDQ total and subscales (web-based: ω=.55-.81; paper: ω=.52-.80). The intraclass correlation coefficients between the web-based and paper formats were all significant at the 99.9% confidence level: ASSQ (r=0.66, <i>P</i>&lt;.001); ADHD-RS total and subscales (<i>r</i>=0.66-0.74, <i>P</i>&lt;.001); DCDQ total and subscales (<i>r</i>=0.66-0.71, <i>P</i>&lt;.001); P-SDQ Total Difficulties and subscales (<i>r</i>=0.55-0.73, <i>P</i>&lt;.001). There were no significant differences between the web-based and paper formats for total mean score of the ASSQ (<i>P</i>=.76), total (<i>P</i>=.12) and subscale (<i>P</i>=.11-.47) mean scores of DCDQ, and the P-SDQ Total Difficulties mean score (<i>P</i>=.20) and mean subscale scores (<i>P</i>=.28-.79). Although significant differences were found between the web-based and paper formats for mean ADHD-RS scores (total: <i>t</i>₁₃₂=2.83, <i>P</i>=.005; Inattention subscale: <i>t</i>₁₃₃=2.15, <i>P</i>=.03; Hyperactivity/Impulsivity subscale: <i>t</i>₁₃₃=3.21, <i>P</i>=.002), the effect sizes were small (Cohen <i>d</i>=0.18-0.22).

These results suggest that the web-based versions of the ASSQ, ADHD-RS, DCDQ, and P-SDQ were equivalent, with the same level of internal consistency and intrarater reliability as the paper versions, indicating the applicability of the web-based versions of these questionnaires for assessing neurodevelopmental disorders.