Screening Tools for Autism in Culturally and Linguistically Diverse Paediatric Populations: A Systematic Review
BMC Pediatrics
Huda, E., Hawker, P., et al. (2024).
BMC Pediatrics, 24(1), 610.
<div>This systematic review explores the psychometric properties of autism screening tools in culturally and linguistically diverse children (i.e., individuals younger than 18 years old). </div>
Not stated
From database inception to October 2022
<div>Published, peer-reviewed studies</div>
51
<div><strong>The Modified Checklist for Autism in Toddlers (M-CHAT)</strong> had good overall reliability and/or validity when used with Spanish, Turkish, Chinese, Korean, French, and Arabic-speaking populations. Reported psychometrics included: </div>
<div>
<ul>
<li>sensitivity = 67%-93%;</li>
<li>specificity = 63%-99%;</li>
<li>positive predictive value (PPV) = 14%-88%;</li>
<li>negative predictive value (NPV) = 77%-100%; and </li>
<li>internal consistency (α) = 0.76-0.9.</li>
</ul>
<div>Reduced reliability and/or validity were reported for th M-CHAT when used with Sinhalese, Kurdish, and Persian-speaking populations. Reported psychometrics included: </div>
<div>
<ul>
<li>sensitivity = 25%-90%;</li>
<li>specificity = 70%-80%;</li>
<li>positive predictive value (PPV) = 5%-13%;</li>
<li>negative predictive value (NPV) = 85%; and</li>
<li>internal consistency (α) = 0.61.</li>
</ul>
</div>
</div>
<div><strong>The M-CHAT, Revised with Follow-Up (M-CHAT-R/F) </strong>had good overall reliability and/or validity when used with Turkish, Spanish, Chinese, Portuguese, Serbian, Malian, Northern Soho, and Albanian-speaking populations. Reported psychometrics included:</div>
<div>
<ul>
<li>sensitivity = 50%-100%;</li>
<li>specificity = 56%-100%; </li>
<li>PPV = 12%-100%;</li>
<li>NPV = 87%-100%; and </li>
<li>α = 0.57-0.96.</li>
</ul>
</div>
<div><strong>The Social Responsiveness Scale (SRS) </strong>had fair to good reliability and good to moderate validity when used with Chinese, Korean, Vietnamese, Japanese, Spanish, and Finnish-speaking populations. Reported psychometrics included:</div>
<div>
<ul>
<li>sensitivity = 64%-100%;</li>
<li>specificity = 67%-98%; </li>
<li>PPV = 39% (one Chinese study);</li>
<li>NPV = 99% (one Chinese study); and </li>
<li>α = 0.81-0.97.</li>
</ul>
</div>
<div><strong>The Autism Spectrum Rating Scale (ASRS) </strong>had good overall validity but variable reliability when used with Chinese-speaking populations in China. Reported psychometrics included:</div>
<div>
<ul>
<li>sensitivity = 89%-94%;</li>
<li>specificity = 77%-85%; </li>
<li>PPV = 39%;</li>
<li>NPV = 99%; and </li>
<li>α = 0.59-0.93.</li>
</ul>
</div>
<div><strong>The Social Communication Questionnaire (SCQ) </strong>had good to high reliability and validity when used with Arabic, Greek, Malian, and Chinese-speaking populations. Reported psychometrics included:</div>
<div>
<ul>
<li>sensitivity = 71%-96%;</li>
<li>specificity = 72%-98%; </li>
<li>PPV = 73% (one Malian study);</li>
<li>NPV = 70% (one Malian study); and </li>
<li>α = 0.92.</li>
</ul>
</div>
<div><strong>The Quantitative Checklist for Autism in Toddlers (Q-CHAT) </strong>had acceptable reliability and validity when used with Italian and Serbian-speaking populations. Reported psychometrics included:</div>
<div>
<ul>
<li>sensitivity = 83% (Italian) and 96% (Serbian);</li>
<li>specificity = 78% (Italian) and 82% (Serbian); </li>
<li>PPV = 39% (one Serbian study);</li>
<li>NPV = 99% (one Serbian study); and </li>
<li>α = 0.68-87.</li>
</ul>
</div>
<div>One study found that the <strong>Childhood Autism Spectrum Test (CAST) </strong>had good reliability and acceptable validity when used with Spanish-speaking populations in Spain. Reported psychometrics included:</div>
<div>
<ul>
<li>sensitivity = 84%;</li>
<li>specificity = 93%; </li>
<li>PPV = 63%; and</li>
<li>α = 0.83.</li>
</ul>
<div>Two studies reported good reliability for Chinese-speaking individuals in China (weighted kappa = 0.53, theta = 0.89).</div>
</div>
<div>One study found that the <strong>Autism Spectrum Quotient-Child (AQ-C) </strong>had good reliability and validity when used with Chinese-speaking populations in China. Reported psychometrics included:</div>
<div>
<ul>
<li>sensitivity = 94%;</li>
<li>specificity = 88%; and</li>
<li>α = 0.77.</li>
</ul>
</div>
<div>Across two studies, the <strong>Indian Autism Screening Questionnaire/ Indian Scale for Assessment of Autism (IASQ/ISAA)</strong> had good reliability and high validity when used with Hindi-speaking individuals in India. Reported psychometrics included:</div>
<div>
<ul>
<li>sensitivity = 97%;</li>
<li>specificity = 81%;</li>
<li>PPV = 63%; </li>
<li>NPV = 99%; and</li>
<li> α = 0.93.</li>
</ul>
</div>
<div>One study found that the <strong>Autism Behaviour Checklist (ABC) </strong>had moderate sensitivity but poor specificity when used with Mandarin Chinese-speaking populations in China. Reported psychometrics included:</div>
<div>
<ul>
<li>sensitivity = 80%;</li>
<li>specificity = 59%; </li>
<li>PPV = 93%; and</li>
<li>NPV = 32%.</li>
</ul>
</div>
<div>One study found that the <strong>Autism Spectrum Screening Questionnaire (ASSQ) </strong>had fair validity when used with Finish-speaking populations in Finland. Reported psychometrics included:</div>
<div>
<ul>
<li>sensitivity = 89%;</li>
<li>specificity = 82%; </li>
<li>PPV = 62%; and</li>
<li>NPV = 96%.</li>
</ul>
</div>
<div>One study found that the <strong>Hiva</strong> screening tool had high validity and poor validity when used with Kurdish and Persian-speaking populations in Iran. Reported psychometrics included:</div>
<div>
<ul>
<li>sensitivity = 100%;</li>
<li>specificity = 97%; </li>
<li>PPV = 38%; and</li>
<li>α = 0.58.</li>
</ul>
</div>
<div>One study found that the <strong>Toddler Autism Screening Questionnaire (TASQ) </strong>had high sensitivity (100%) and specificity (97%) when used with Taiwanese-speaking populations in Taiwan.</div>
