Certified nurse-midwives (CNMs) and certified midwives (CMs) attend 11% of US vaginal births, and CNM/CM databases include a rich source of information on the birth outcomes of women not generally available through obstetric electronic health records (oEHRs). Although CNM databases are in wide use, studies on their validity are sparse. We examined the validity of a practice-specific CNM birth database compared with an oEHR in a large tertiary-care facility.Methods:
The study population consisted of all 3133 births in a large CNM practice that were entered into a practice-specific electronic birth database from 2001 to 2008. We assessed agreement with the oEHR overall and according to individual maternal demographic characteristics; obstetric, medical, and social risk factors; labor and birth information; and newborn characteristics. We also evaluated whether there were differences in agreement according to early versus later study periods and by years of CNM clinical experience.Results:
Overall agreement between the CNM birth database and the oEHR was 92.4%. Agreement between the CNM birth database and the oEHR was greater than 90% for maternal age, race/ethnicity, route and type of birth, major genital tract trauma, newborn weight, and primary clinician attending the birth. Lower agreement rates for smoking, total weight gain in pregnancy, beginning of pregnancy body mass index, and anesthesia for birth were due, in part, to missing information in the oEHR. Agreement did not vary significantly by early versus late study periods or by years of CNM clinical experience.Discussion:
Findings indicate that a CNM/CM birth database is a valid data source for quality monitoring and outcomes research and may contain more complete information for some variables than the oEHR.