The linked study by de-Wahl Granelli and colleagues (doi:10.1136/bmj.a3037) assesses the contribution of neonatal pulse oximetry saturation screening to early detection of life threatening congenital heart disease.1 Neonatal screening started in the 1960s with the Guthrie bacterial inhibition test for detection of phenylketonuria.2 Screening for this rare condition (about 1/12 000 live births) is almost universal in the developed world, and it illustrates the criteria for such a programme. Phenylketonuria can be reliably identified when asymptomatic, with a low false positive rate, and when treated early death or disability are prevented. Newer screening tests using techniques other than blood sampling should follow the same principles.

Critical congenital heart disease—cardiac malformations that are either ductal dependent or need surgery in the first month of life—occurs in about 170 per 100 000 live births. At first sight, universal pulse oximetry screening for this disease seems to contravene the above principles because clinical examination of the cardiovascular system is a part of the routine care of the newborn. However, clinical examination cannot reliably detect critical congenital heart disease. Cyanosis …