Diagnostic validity, accuracy and inter-rater reliability of a phonological assessment for Danish-speaking children
Marit C Clausen1, Annette V Fox-Boyer2
1Department of Language and Communication, University of Southern Denmark, Campusvej 55, 5230 Odense M, Denmark.
Purpose:
The identification and differential diagnosis of children with speech sound disorders (SSD) is an important task of paediatric speech-language pathologists (SLPs). A correct identification requires valid and reliable assessment tools. Therefore, the aim of the present study was to investigate the diagnostic validity and accuracy as well as inter-rater reliability of the phonology test of LogoFova which was developed for Danish-speaking children.
Method:
Investigation was carried out on two Danish-speaking populations: 61 suspected typically developing (TD) children and 61 children with suspected speech sound disorder aged 2-6 years. All children were assessed with the single-word picture-naming test of LogoFoVa. In order to determine diagnostic validity, it was investigated whether the test differentiated TD children from children with SSD as well as subgroups of SSD via a phonetic and phonological pattern analysis. Two different cut-off criteria were applied for the definition of patterns. Sensitivity and specificity as well as likelihood ratios were additionally calculated. To determine inter-rater reliability, transcriptions, pattern analyses and subgrouping were compared across raters.
Results:
Overall, diagnostic validity of the picture-naming test of LogoFoVa was good as differentiation between TD children and children with suspected SSD as well as amongst subgroups of SSD (articulation impairment, phonological delay and atypical speech development) was possible. However, accuracy of differentiation between TD children and children with SSD was affected by the cut-off criterion applied. Inter-rater reliability was found to be almost perfect for transcription and moderate for identification of phonological processes as well as for subgrouping. Again, agreement rates depended on the cut-off criterion chosen for the definition of a phonological pattern.
Conclusion:
LogoFoVa was found to be a reliable and valid clinical tool for the identification and subgrouping of children with SSD in Danish SLP practice if a new cut-off criterion was applied.
