Language Sample Analysis in Bilingual Children: A Meta-Analysis of Diagnostic Accuracy

Language sample analysis (LSA) is a common approach in the identification of developmental language disorder (DLD) in bilingual children. Because LSA is often recommended as a less biased assessment method, a better understanding of its diagnostic accuracy is needed to ensure appropriate decision making in clinical applications. This systematic review and meta-analysis synthesized the existing literature on the diagnostic accuracy of LSA in DLD identification for bilingual children.

We conducted a comprehensive literature search of electronic databases along with supplementary strategies and applied abstract and full-text screening procedures to identify eligible studies. We extracted relevant metrics from all studies to derive pooled estimates of sensitivity and specificity. We then conducted moderator analyses to examine the effect of elicitation language (first or second language) and the performance of integrated measures that included LSA with other methods as well as a risk of bias evaluation for each study.

The search yielded nine articles that met the inclusion criteria. Studies included participants ranging from 2 to 7 years of age. Across studies, diagnostic accuracy in DLD identification ranged from poor to good. The pooled sensitivity was 70%, and the pooled specificity was 83%. Several LSA measures exhibited fair sensitivity and specificity, most of which consisted of combinations of multiple LSA metrics. Language of elicitation was not associated with variation in diagnostic accuracy. Integrated measures exhibited the highest diagnostic accuracy.

LSA is a useful approach in the assessment of language for bilingual children, but the variability in diagnostic accuracy suggests that not all metrics may perform with the same degree of precision. Integrated measures, which include LSA along with other assessment methods, likely provide the most representative characterization of language ability. These results indicate that LSA is best used as part of an assessment battery that includes multiple methods of measurement.

https://doi.org/10.23641/asha.31592305.