Accuracy of the Childhood Autism Rating Scale: A Systematic Review and Meta‐Analysis
Developmental Medicine & Child Neurology
Moon, S. J., Hwang, J. S., et al. (2019).
Developmental Medicine & Child Neurology, 61(9), 1030-1038.
This systematic review and meta-analysis investigates the accuracy of the Childhood Autism Rating Scale.
Not stated
From database inception to August 15, 2018
Not stated
24 studies included in the qualitative analysis; 9 in the meta-analysis
<p>Ten studies assessed for internal consistency, with Cronbach’s alpha in the range of 0.82 to 0.95. Meta-analysis of 9 studies showed considerable heterogeneity (I²=76.2%). Six studies of multidisciplinary teams showed reduced heterogeneity (I²=52.2%), and the pooled value was 0.90, 95% CI=[0.87– 0.92].</p>
<p>Five studies assessed for test-retest reliability: three assessed for Pearson correlation (r=0.91–0.99, all p<0.01), one for coefficient Kappa (k=0.99), and one did not specify.</p>
<p>Eighteen studies assessed for criterion validity. Meta-analysis was performed on nine samples from nine studies that used the combination of CARS with the DSM-IV as a reference standard with the cut-off value of 30. Heterogeneity could not be resolved and pooled samples could not be obtained. For two studies that were "low risk of bias", sensitivity was 0.86 and 0.71 and specificity was 0.79 and 0.75. Seven studies were "at risk of bias" and sensitivity was in the range of 0.41 to 1.00, with specificity in the range of 0.62 to 1.00.</p>
<p>Six studies provided construct validity by factor analysis. Total variance ranged from 41.7% to 70.8%. Four out of six studies had cross-loading or dropping items or failure to meet the factor loading criteria.</p>
