Jump to comment:

• Lack of sensitivity of ASQ communication domain remains a problem.
  Philip Wilson, Vicky Gilroy, Jenna Charlton, Robert Rush and Cristina McKean
  Published on: 28 March 2022
• User Beware: For accurate screening, use complete tests
  Jane K Squires
  Published on: 20 February 2022

• Published on: 28 March 2022

  Lack of sensitivity of ASQ communication domain remains a problem.

  • Philip Wilson, Professor of primary care and rural health University of Aberdeen
  • Other Contributors:
    • Vicky Gilroy, Head of Projects and Evaluation
    • Jenna Charlton, Research Fellow
    • Robert Rush, Independent statistical consultant
    • Cristina McKean, Professor of Child Language Development and Disorders

  We thank Professor Squires for her interest in our work and we agree that concern about any neurodevelopmental problem merits early comprehensive assessment of all developmental domains. We would like to reassure her that all the participants in our study received the full ASQ, interpreted by the family’s health visitor who took action according to the overall assessment. We were, however interested specifically in the performance of the ASQ’s communication domain in terms of identifying developmental language disorders. Even when we included children in the ‘Monitoring Zone’ of that domain we found that at least a third of children with significant problems were missed.
  To our surprise, parental concern about their child’s language did not improve the performance of the Sure Start Language Measure (SSLM): parental concern was associated with an increased likelihood of false positivity among the screen-positive children.
  We therefore suggest that if the ASQ is to be used without an additional language measure such as the SSLM on a universal level with 24-30 month old children, consideration should be given at least to lowering the thresholds for monitoring or referral within the communication domain.

  Conflict of Interest:
  None declared.

  • Back to top
• Published on: 20 February 2022

  User Beware: For accurate screening, use complete tests

  • Jane K Squires, Emerita Professor, Researcher University of Oregon

  As a developer and researcher of the Ages & Stages Questionnaires, I read with interest Universal Language Development Screening: Comparative Performance of Two Questionnaires by Wilson et al., published January 6, 2022. I was not able to review this manuscript prior to publication; there are several methodological errors that severely limit the design and consequent outcomes of this study.

  First, the ASQ was developed to be used in total—all 30 items, 5 domains, at each administration point. Domains or areas were not designed to be used individually or independently. The psychometric properties of the ASQ will be robust only if/when the entire test is administered, ideally at periodic intervals over time.

  Second, a research design that uses only the communication domain of the ASQ-3 is flawed. The communication domain contains 3 expressive language items and 3 receptive items. Additionally, because of the overall interdependence of young children’s skills, communication items are embedded throughout the interval in other domains. For example in the intervals targeted by Wilson et al., (i.e., 24, 27, and 30 month ASQ-3) there are a total of 7 items focused on communication skills (e.g., listening, repeating, following directions) at 24 months; 10 items at 27 months, and 12 items at 30 months. Therefore analyzing only the 6 items under the domain heading is not looking at communication as broadly as does the test in its entirety.

  Third, th...

  Show More

  As a developer and researcher of the Ages & Stages Questionnaires, I read with interest Universal Language Development Screening: Comparative Performance of Two Questionnaires by Wilson et al., published January 6, 2022. I was not able to review this manuscript prior to publication; there are several methodological errors that severely limit the design and consequent outcomes of this study.

  First, the ASQ was developed to be used in total—all 30 items, 5 domains, at each administration point. Domains or areas were not designed to be used individually or independently. The psychometric properties of the ASQ will be robust only if/when the entire test is administered, ideally at periodic intervals over time.

  Second, a research design that uses only the communication domain of the ASQ-3 is flawed. The communication domain contains 3 expressive language items and 3 receptive items. Additionally, because of the overall interdependence of young children’s skills, communication items are embedded throughout the interval in other domains. For example in the intervals targeted by Wilson et al., (i.e., 24, 27, and 30 month ASQ-3) there are a total of 7 items focused on communication skills (e.g., listening, repeating, following directions) at 24 months; 10 items at 27 months, and 12 items at 30 months. Therefore analyzing only the 6 items under the domain heading is not looking at communication as broadly as does the test in its entirety.

  Third, the number of items in a test matters--more items focused on a domain will yield more accurate results. The 50 items of the Sure Start Language Measure should provide a more reliable measure of the child’s overall language than 6 ASQ items. In assessment, more is better--results will tip in favor of the longer, more complete inventory of skills in that area.

  Fourth, the ASQ is meant to be used in conjunction with the Overall questions asking about parent concerns in each in interval. Parent concerns are equally weighed with developmental scores and referral decisions are based BOTH on ASQ scores as well as any parent concerns. Parents are often aware their child is not talking quite like the neighbor’s child; asking about concerns allows parents to voice their worries in addition to answering the scored questionnaire items. Asking about concerns enhances the sensitivity of screening outcomes and referral decisions. The Overall section asks 2 additional questions about the quality of a child’s speech in the 24 month and 27 month intervals, and 3 additional questions in the 30 month interval.

  Our hopes, as ASQ developers, is that the screening tool can be used by UK pediatric professionals to improve outcomes for young children and can be used in an efficient way that strengthens pediatric and early childhood procedures such as home visiting and early childhood education. When ASQ scores are in the referral range (i.e. below the cut score in any one domain) and/or parents have concerns, the child should be immediately referred to Early Intervention/Early Childhood Education or other community providers for more in-depth assessment. A monitoring zone result should trigger a healthcare provider to look more carefully at the other ASQ domains as well as Overall questions. If other domains have monitoring zone scores along with parents indicate concerns, the child should again receive a referral. On-going surveillance in combination with periodic screening using the ASQ or other standardized test as defined by the AAP are necessary for early identification and optimal outcomes for children and families.

  In sum, Wilson et al are investigating a topic of critical importance. We hope that their work increases thorough general developmental assessment as well as more specific language evaluation with those children in need. We also hope that the ASQ is used as developed and in total to make screening decisions in the future.

  Show Less

  Conflict of Interest:
  None declared.

  • Back to top