care providers to anticipate other complications and take actions to prevent or treat them as early as possible. Several measures of severity of acute illness (e.g., Scoring for Neonatal Acute Physiology and Clinical Risk Index for Babies) have been developed and are associated with mortality and morbidity rates in preterm infants, although these are not methods for the estimation of gestational age (Gagliardi et al., 2004; Richardson et al., 1993, 1999b). These systems of scoring for the severity of acute illness can be used to compare neonatal intensive care units (NICUs) for quality improvement initiatives and for insight into why complications and outcomes differ among NICUs (Richardson et al., 1999a).

In the 1960s to 1970s, missing or inaccurate gestational age data for many newborns led to a search for postnatal methods of determining gestational age. These methods invariably focused on the degree of infant maturation (Allen, 2005a; Philip et al., 2003). Farr et al. (1966) described the maturation of a number of external physical characteristics in preterm and term infants. Hittner et al. (1977, 1981) proposed a systematic method of grading the disappearance of the pupillary membrane (i.e., the anterior vascular capsule of the lens of a preterm infant’s eye) at 2-week intervals from 27 to 34 weeks of gestation. French investigators developed a method of measuring neurological maturity from observations of the changes in neck, trunk, and extremity flexor tone and posture with gestational age (AmielTison, 1968; Amiel-Tison et al., 2002; Philip et al., 2003; Saint-Anne Dargassies, 1977). A number of postnatal clinical measures of the degree of maturation of external physical characteristics or neurological muscle tone, or both, were developed to estimate gestational age of preterm and fullterm infants (Allen, 2005a; Ballard et al., 1979, 1991; Dubowitz et al., 1970; Parkin et al., 1976).

These clinical postnatal measures are less predictive of gestational age (i.e., pregnancy duration at birth) at the extremes of gestation (i.e., in preterm and postterm infants) and in very sick infants (Alexander et al., 1990; Sanders et al., 1991; Shukla et al., 1987; Spinnato et al., 1984). The most widely used postnatal measures, the New Ballard Score (Ballard et al., 1991) and the Dubowitz gestational age assessment (Dubowitz et al., 1970), overestimate gestational age by 2 or more weeks in 45 to 75 percent of preterm infants with birth weights less than 1,500 grams (Sanders et al., 1991; Shukla et al., 1987; Spinnato et al., 1984). The accuracies of these measures decrease with an increase in gestational age (Alexander et al., 1990; Sanders et al., 1991). Gestational age is more often overestimated in African-American preterm infants than in white preterm infants, even when