Visual assessment of jaundice

Despite conflicting evidence concerning its reliability, clinicians have long relied on cephalocaudal progression3 to estimate the severity of neonatal jaundice. In one of the largest studies to date, Keren et al report poor correlation between visual assessment by experienced neonatal nurses and transcutaneous or laboratory bilirubin measurement. Correlation was poorer for non-white and near-term infants, two groups at greater than average risk for significant hyperbilirubinaemia. Riskin et al4 reported similar inaccuracy in visual assessment conducted by both neonatal doctors and nurses. Furthermore, while absence of jaundice was found to have good negative predictive value, visual assessment performed poorly in predicting which infants are at risk for significant hyperbilirubinaemia.1

These studies evaluated visual assessment of jaundice in the first 3 days of life. Their findings may not be applicable thereafter, by which time most newborn infants have been discharged. It is unlikely that midwives or other health professionals, working in home conditions of varying space and lighting, will visually assess jaundice more accurately than professionals in the controlled environment of a postnatal ward during a well-conducted research study. While extrapolation is problematic, the findings of Keren et al have significant implications: evaluation of neonatal jaundice should not rely on visual inspection alone but must be supplemented by measurement of serum (or transcutaneous) bilirubin. Furthermore, …