Background
The Georgian Birth Registry (GBR) is a digital nationwide birth registry implemented in 2016 that currently covers 99.8% of deliveries in the country. Registration in the GBR is mandatory by law for antenatal care providers and hospitals, and much of the data are transferred from medical records (MRs) to GBRs by a designated person at the medical facility, a so-called single-entry data transfer. We aimed to assess the correspondence of the registration of selected antenatal, intrapartum, and newborn variables between GBR and MRs for a random selection of delivering women.
Methods
We randomly selected 1,050 women who gave birth in 2018. We collected data from the GBR, encompassing around 26 distinct variables associated with pregnancy, childbirth, and the newborn, and mirrored this data extraction process for individual MRs. For dichotomous, ordinal, and date variables, we compared the proportion of registrations with complete agreement between the GBR and the MRs, as well as the proportion of registrations deviating +/- one category or more. Differences in the registrations of newborn birthweight were compared using paired t-tests and visualized with a Bland-Altman plot.
Results
Of the 25 dichotomous, ordinal, and date variables related to pregnancy, childbirth, and newborn birth in the GBR, 20 displayed more than 95% complete agreement with the information in the MRs, and no substantial difference in newborn birthweight registered in the GBR and MRs was observed. The prevalence of maternal morbidity and postpartum hemorrhage (PPH) registered in the MRs was lower than expected, while the proportion of fetuses with transverse lies was higher than expected.
Conclusions
Most antenatal, intrapartum, and newborn information registered in the GBR has satisfactory agreement with the MRs, and error rates are as expected for a single-entry data processing method. The lower-than-expected prevalence of gestational diabetes, preeclampsia, hypertensive disorders, and PPH registered in the MRs, as well as the higher-than-expected prevalence of transverse fetal presentation, warrants in-depth investigation to ensure that the quality of care is satisfactory and to further improve registration in both the MRs and GBR.