Factors influencing the relationship between a newborn assessment of gestational maturity and the gestational age interval.

This study examines the correlation and agreement between the calculated 'gestational interval' from the date of last menstrual period (LMP) and the Ballard newborn assessment of gestational maturity. It then investigates selected maternal socio-demographic, prenatal, intrapartum and infant factors which may account for differences between these two gestational age determination methods. Single live-born deliveries (n = 10,794) were selected from a 1982-1986 medical university perinatal information data system for analysis. The Ballard assessment overestimates the LMP at earlier gestational ages and underestimates in the post-term range. The discrepancy between LMP and Ballard varies across the range of gestational age values and the extent of variation differs by several maternal and infant risk factors. These findings suggest considerable bias may result from the indiscriminate case mixing of LMP and Ballard values for use in gestational age research or in standardising developmental tests. We highlight the deficiencies in using correlation coefficients or mean differences to assess agreement between these measures.