Abstract

Developmental screening tests are in widespread use, but few reliable and valid tests are available. The most frequently used screening instrument for detecting young children who are at risk for developmental delays is the Denver Developmental Screening Test (DDST). Although the DDST has excellent test specificity, overreferring few children, results from more than a dozen studies of the DDST's concurrent and predictive validity fail to replicate the original validation and demonstrate a uniformly poor sensitivity, ie, a high proportion of underreferrals. Whether samples are stratified by age, risk, duration of time between predictor and outcome, or type of outcome measure used, these studies demonstrate that the DDST underrefers children at nearly a 2:1 ratio. Several other screening tests with more optimal psychometric properties are presented. It is urged that caution be exercised in using tests for predicting the risk of developmental problems in young children unless the tests have acceptable levels of sensitivity and specificity.