Objective: to compare expected date of delivery (EDD) and gestational age (GA) obtained by midwives with those calculated using the Confidential Enquiry into Stillbirths and Death in Infancy (CESDI) recommended formula.
Design: retrospective study of obstetric records and prospective study of clinical cases examined by multiple midwives.
Setting: postnatal wards, Maternity Unit, Plymouth, Devon, UK.
Participants: two studies were performed. The first was a retrospective analysis of 115 sets of obstetric records. The second, a prospective study which included five clinical cases and 19 qualified midwives whose experience ranged from six months to 25 years.
Findings: in the retrospective study, 68 (59%) of the 115 obstetric case notes had sufficient information to apply the CESDI formula. The midwives'calculated EDD was interpreted to a GA and 35 (52.5%) agreed to within three days of the GA derived from the CESDI formula EDD. In the prospective study, the midwives' calculation of EDD was in good agreement with the CESDI formula in cases where last menstrual period (LMP) was known and menstrual cycle was 28 days with 17 (90%) of the 19 midwives providing the EDD to within three days of the CESDI formula. In the clinical case where LMP was known but menstrual cycle length was 33 days only two (10%) of the 19 calculated the EDD within three days of the CESDI formula.
Key conclusions: when LMP is known and the cycle length is 28 days, midwives produce results consistent with the CESDI formula. However, when the menstrual cycle differs from 28 days or LMP is unknown, inaccurate or ambiguous and obstetric ultrasound scan (USS) information is used to calculate the EDD, the midwives show varying degrees of departure from the CESDI formula. Further, we found that the midwives tended to produce underestimates when calculating GA from EDD. This bias occurs across the range of gestations covered, including early gestations when such biases may have more important medical implications.
Implications for practice: midwives' calculation of EDD and estimation of GA showed both random and systematic errors; in some cases, the errors were so large that they may have important medical consequences. If the CESDI-recommended formula for calculating EDD were used there would be improved accuracy and reliability of EDD and the calculation of GA.