Precision of gestational age assessment in the neonate

The precision of the Ballard scale for assessing gestational age (GA) was evaluated in a consecutive sample of 227 preterm and/or low-birth-weight neonates. Each newborn was rated independently by two neonatologists and the difference in GA estimation between them was computed. The estimated precision was not high, the 95% tolerance interval estimate being as large as 7.4 weeks. The precision of the neurologic and physical parts of the scale was poorer than that of the complete scale (95% of differences less than 10.5 and 9.2 weeks respectively), and more influenced by the type of delivery. These findings are not unexpected from statistical theory, and cast doubts on the use of only the physical part of the Ballard scale in assessing GA, since greater accuracy could be accompanied by reduced precision.