An evaluation of developmental examination as a method of detecting neurological, visual, and auditory handicaps in infancy.

Abstract

A recent editorial (British Medical Journal, 1971) discussed the value of universal neurological screening of the newborn in the light of criteria accepted as generally applicable to all screening procedures: firstly, that a reliable screening test should not give many false negative or positive results; secondly, that early detection is worth while because early treatment has proved advant ageous; and finally, that the use of medical and other resources is a reasonable priority in terms of total supply of these resources and the demands made on them. (It was estimated that a 15-minute neonatal neurological examination on every new born would take up 470 hours per year of all practising paediatricians and senior paediatric registrars.) The editorial noted that none of the requirements had been fulfilled for neonatal neurological screen ing but went on to say that 'It is now widely accepted that all children should have periodic developmental screening tests beginning at, say, 6 weeks, 6 months, and 10 months'. Clear and practicable schemes have been devised for doing these tests (Sheridan, 1968; Egan, Ulingworth, and MacKeith, 1969). We can find no published evidence showing how effective developmental screening (or even full developmental examination) is in detecting abnor malities. The screening schemes in current use (Frankenberg and Dodds, 1967; Sheridan, 1968; Egan et al, 1969) presume a degree of effectiveness based on the admittedly reasonable assumption that a proportion of infants with neurological, auditory, and visual defects will present as developmental delay (Gesell and Amatruda, 1941). Frankenberg and Dodds claim to have undertaken both reli ability and validity studies on the Denver scale. Estimates of the former were based on the results of two tests separated by an interval of one week and undertaken by the same examiner on 20 children whose ages ranged from 2J months to 5 J years (test retest reliability) and the percentage agreement between different examiners of the same subject (examiner reliability); estimates of the validity were based on correlations between results of examination with the Denver scale and examination with the Yale developmental schedule (itself a developmental examination based on Gesell, with the additional items from various IQ tests). Im portant as these parameters are they clearly give no indication of the ability of developmental screening to detect clinical handicap in infancy. Assessment of reliability (or effectiveness) in this context can be determined only from concordance studies between the results of developmental screening and definitive clinical examination of the same popu lation. Similarly, there is no published evidence to support the criterion of advantageous early therapy in respect of infant developmental screening be cause of the considerable ethical and practical difficulties of undertaking controlled trials of early treatment. In this situation one must therefore be guided by the opinion of experts and accept that early treatment is of value. Finally, it would seem plausible to argue that the reasonableness of developmental screening as a priority can only be approached when evidence of the reliability and efficacy of the procedure has been measured. Thus, in view of the current absence of evidence to support any of the above criteria for de velopmental screening, we felt that the one aspect which needed to be looked at urgently was that of the reliability and effectiveness of developmental examination. Such a study forms the basis of this paper. A full developmental examination procedure is time consuming and requires considerable training and expertise for its application and interpretation. These qualifications make it unsuitable for routine use on large populations of apparently healthy children. An attempt has been made to overcome this problem by devising simplified developmental tests of items abstracted (somewhat arbitrarily) from the battery of observations in the full diagnostic 94