Skip to main content

Table 3 Gestational age methods, accuracy and limitations

From: Born Too Soon: The global epidemiology of 15 million preterm births

Method

Accuracy

Details

Availability/feasibility

Limitations

Early ultrasound scan

+/- 5 days if first trimester

+/- 7 days after first

trimester

Estimation of fetal crown-rump length +/- biparietal diameter/femur length between gestational age 6 - 18 weeks

Ultrasound not always available in low-income settings and rarely done in first trimester

May be less accurate if fetal malformation, or maternal obesity

Fundal Height

~ +/- 3 weeks

Distance from symphysis pubis to fundus measured with a tape measure

Feasible and low cost

In some studies similar accuracy to LMP Potential use with other

variables to estimate GA when no other information available

Last menstrual period

~ +/- 14 days

Women's recall of the date of the first day of her last menstrual period

Most widely used

Lower accuracy in settings with low literacy. Affected by variation in ovulation and also by breastfeeding. Digit preference

Birthweight as a surrogate of gestational age

More sensitive/specific at lower gestational age e.g.

<1500 g most babies are preterm

 

Birthweight measured for around half of the world's births

Requires scales and skill. Digit preference

Newborn examination

~ +/- 13 days for Dubowitz, higher range for all others

Validated scores using external

+/or neurological examination of the newborn

e.g. Parkin, Finnstrom, Ballard and Dubowitz scores

Mainly specialist use so far. More accurate with

neurological criteria which require considerable skill. Potential wider use for simpler scoring systems

Accuracy dependant on complexity of score and skill of examiner. Training and ongoing quality control required to maintain accuracy

Best obstetric estimate

Around +/- 10 days (between ultrasound and newborn examination)

Uses an algorithm to estimate gestational age based on best information available

Commonly used in high-income settings

Various algorithms in use, not standardized

  1. Adapted from Parker, Lawn and Stanton (unpublished Master's thesis)