Effectiveness of screening for congenital dislocation of the hip.

The incidence of congenital dislocation of the hip (CDH) in Birmingham was measured in the period 1942-52 and again in 1950-54; it is measured now in 1974-83. Despite the introduction of neonatal screening for CDH in November 1966, between the times of the earlier and the present measurements there was no indication that CDH requiring prolonged treatment had declined. The performance of the procedure itself was poor. It detected only a third of genuine cases, and the false positives outnumbered the true positives by 10:1. The incidence of 'definite' cases varied markedly from year to year in a manner suggesting diagnostic idiosyncracies rather than genuine fluctuations. There was no evidence in particular maternity units that high levels of early detection led to low levels of late diagnosis. Indeed, maternity units with the worst results--the highest rates of late diagnosis and of prolonged treatment--had the highest rates of early detection and early treatment. Current theories of natural history, aetiology, and treatment should be reappraised.