While a stable diagnosis of Autism Spectrum Disorders can be made as early as 2 years of age, the diagnosis of mild ASD cases are usually not made until primary school age or much later; and yet it is these milder cases that can benefit the most from early intervention. The present study aimed at pushing the identification of mild ASD children earlier to preschool age. A review of current screening tools revealed that they were not effective in identifying milder ASD variants. One reason perhaps is that the existing tools rely primarily on parental reports. Note that young children with mild ASD often function adequately interacting with an adult who knows them well; they typically face more difficulty in free play with other children. Parents as a result may not be in a good position to detect milder ASD. The present study therefore took a very different approach from existing screening tools by developing a more objective scale based on observation by of peer interaction in preschools. Considering peer interaction deficits are central for ASD, and deficits of these milder individuals might be more obvious in a setting that tax their social skills. A screening tool based on peer interaction observation in preschool, namely the Structured Classroom Observation Scale (SCOS), was thus developed.
Drawing on existing screening tools and experts input, 84 items were compiled and pilot tested. An initial psychometric study of the scale was conducted using a community sample, with 304 preschoolers aged 3 and 4, from four English-language international schools in Hong Kong. The initial 84-item version was trimmed substantially to result in a user-friendly 13-item observation scale with good psychometric properties. The final SCOS includes 3 items depicting self-regulation challenges and 10 items describing difficulties in peer interaction. The initial psychometric study using a community sample indicated substantial interrater reliability (u= .76) and acceptable test-retest reliability (ICC = .72). The average agreement for individual items was less satisfactory (T = .40). Using Latent Class Analysis, the present scale delineated the children into 4 groups: Typical, Shy, Behavioral and High Risk of ASD.
A subsequent validity study (n = 186) comparing the SCOS with ADOS scores showed that the class membership of the children based on SCOS predicted their ADOS results, after controlling for age and gender. Children from the High Risk group were found to have significantly higher Calibrated ADOS Severity scores than the other 3 groups; and their mean ADOS scores (i.e., 8.18) were above the cutoff for ASD on the ADOS. A 9-month follow found more reports of parental concerns in the High Risk group with ADOS scores above the cutoff. Discriminant validity of the SCOS was demonstrated between the scale and Head Start Competence Scale (parent version).
In contrast to the usual portrayal of active but odd stereotypical children with Asperger Syndrome, the High Risk group identified by the SCOS consisted of children with infrequent interfering behaviors. They were relatively passive but not necessarily odd during social interactions, which might perhaps explain why early identification was difficult. The current scale also identified another two groups of children (Shy group and Behavioral group), which will require longitudinal follow up to ascertain educational or intervention implications. / published_or_final_version / Clinical Psychology / Doctoral / Doctor of Psychology
| Identifer | oai:union.ndltd.org:HKU/oai:hub.hku.hk:10722/180818 |
| Date | January 2012 |
| Creators | Au, Hoe-chi, Angel., 區浩慈. |
| Publisher | The University of Hong Kong (Pokfulam, Hong Kong) |
| Source Sets | Hong Kong University Theses |
| Language | English |
| Detected Language | English |
| Type | PG_Thesis |
| Source | http://hub.hku.hk/bib/B48690387 |
| Rights | The author retains all proprietary rights, (such as patent rights) and the right to use in future works., Creative Commons: Attribution 3.0 Hong Kong License |
| Relation | HKU Theses Online (HKUTO) |
Page generated in 0.0121 seconds