Is preschool child health surveillance an effective means of detecting key physical abnormalities?

Method This was a prospective study of a 1-year cohort of children born to mothers registered with 25% of general practices in Nottingham. A cohort of 2308 babies was followed up for 15–18 months. All referrals for five tracer physical conditions (undescended testes, congenital heart disease, squints, developmental dysplasia of the hip and congenital hearing loss) were followed up through hospital record systems and copies of child health surveillance reviews from the personal child health records.

Results One thousand, nine hundred and seventy-two (85%) of the initial cohort were followed up. Four hundred and thirteen (21%) children had been referred for one of the tracer conditions and 91% had been referred directly from child health surveillance reviews. Only 49 (12%) children needed treatment or follow-up after assessment. The sensitivity of screening was high, ranging from 72% for congenital heart disease to 100% for developmental dysplasia of the hip and hearing loss. However, false positive rates were also very high, ranging from 60% for undescended testes to 97% for developmental dysplasia of the hip and hearing loss. Thus the positive predictive value of referral for dysplasia of the hip or hearing loss was only 5%, with a negative predictive value of 100%. In contrast, the positive predictive value for undescended testes was 67%, with a false negative rate of 14%. Eighty-four per cent of those who needed treatment or follow-up after assessment had been referred from the child health surveillance programme. The largest proportion of abnormalities was identified from the 6-week review, but most referrals came from the neonatal review and the health visitor distraction test.

Conclusion The majority of children with the five tracer conditions used in this study were identified during child health surveillance examinations.