Background The accuracy of gestational age (GA) calculators has been demonstrated to be poor but they remain commonly used. The authors surveyed the use of GA calculators in our institution and determined interobserver and interdevice variability of different models, including new electronic methods. The authors devised a protocol for validating device accuracy.

Methods All clinicians in obstetrics were asked their most recent method of calculating GA and whether they felt this was accurate to within 1 day. 10 clinicians then assessed 16 available devices: 14 manual and 2 electronic. Five dates arbitrarily spread across the year were used as the last menstrual period. Individuals had to calculate the estimated date of delivery (EDD) of the five dates for each device. Analysis of variance (ANOVA) was used to statistically analyse variability.

Results 97 staff responded. 73% last used a manual device. 72% believed their method was accurate. There were significant differences (ANOVA, p<0.0001) between device-calculated and control EDD for all manual devices except one, with differences of up to 4 days; electronic devices were consistently accurate with perfect agreement at all gestations and observers. The degree of variability altered throughout the year and was consistent between observers. Interobserver variability was insignificant.

Conclusion Clinicians should be aware of the inaccuracy of manual GA calculators. Electronic devices are perfect and recommended. Manual devices should be validated before use by comparing device-calculated EDD with a 280 day control at five points throughout the year, as errors are scale dependent and related to time of year.