1. Is the DENVER II a valid developmental screening instrument?

Review criteria requiring a high degree of "sensitivity and specificity" in an instrument pertains to only this one type of validity, or accuracy, and fails to consider other parameters of validity. The authors have made no attempt to measure sensitivity and specificity of the DENVER II, because these terms refer to the accuracy in detecting a single disease. Child Development is a broad term pertaining to a variety of developmental entities such as IQ, social and motor development, self help skills, etc. Furthermore, there is no broad agreement on what degree of deviancy constitutes a "delay" on each of those entities. Other types of validity are "predictive" and "face validity." "Face validity" is a measure of the extent to which the measurement reflects what it purports to measure. A popular example of "face" validity is a human growth chart that, as the DENVER II, has norms based upon a representative population.

While we were the first to evaluate developmental screening tests using sensitivity and specificity [1], it has been found that sensitivity and specificity are of extremely limited value because a test's sensitivity and specificity will vary with prevalence and severity of the problem, age of the child, and the absence of a definitive reference test.[2] Therefore, developmental screening tests' validity is confined to the application of screening to an identical population in terms of types, severity, and prevalence of the problems and the specific age for which it was determined.[3] In other words sensitivity and specificity measures are only appropriate for populations similar to those of the reported studies. Since most of the developmental screening test validation studies have used samples that have a relatively high number of deviant children while those seen in the general population are for the most part non-deviant, such validation studies are of little value.[3] It is not surprising Altman finds "…the disadvantage of sensitivity and specificity is that they do not assess the accuracy of the test in a clinically useful way."[4]

2. Does the DENVER II generate an excessive number of over-referrals?

This question primarily comes from one study that used an antiquated method of scoring the test and failed to retest all children who were suspect to rule out temporary delays.[5],[6] Had these errors been avoided, there would have been most likely fewer failing the retest and, therefore, fewer over-referrals. In addition had the author followed the suggested interpretation guideline comparing her results with those from the same state, the authors of the study would have suspected a potential problem. [7]

Finally, it should be emphasized that the accuracy of a test is not only dependent upon test characteristics but also upon the accuracy with which it is administered and interpreted. To simplify training and assure its accuracy the DENVER II Technical Manual sets forth a training protocol, a test of proficiency, instructions for establishing a community screening program and monitoring the program, as well as examples of screening results in different settings. In addition the Denver office has video training tapes, DVDs, etc. The office also provides training and proficiency workshops in Denver and elsewhere. To date it has trained 1,589 Master Instructors in 49 states and 20 foreign countries.

Should readers of the above have questions, it is suggested that they call Dr. Frankenburg at Denver Developmental Materials (800) 419-4729.