Background: Information regarding in vitro fertilization (IVF) as a pregnancy risk factor (yes/ no) is stored in each birth record of the Estonian Medical Birth Registry (EMBR). This study aimed to assess the validity of registration of newborns' IVF status in the EMBR.
Methods: To identify the newborns conceived by IVF, the birth records in the EMBR were compared to individual records on the embryo transfer procedures in the Estonian Health Insurance Fund (EHIF) database as a reference. Maternal age was restricted to 40 years, the age limit for IVF treatment covered by the EHIF. The embryo transfer procedures, that dated up to eight weeks before pregnancy, were additionally checked in the infertility treatment clinics. The validity of IVF status was measured by sensitivity, specificity, positive and negative predictive values (PPV and NPV). Relative risk (RR) of unrecorded IVF status among IVF mothers by socio-demographic characteristics and birth plurality was estimated using modified Poisson regression models.
Results: There were 3198 newborns identified as conceived by IVF in the EMBR in 2005-2014. Eight of them were incorrectly entered as born after IVF. The record linkage with the EHIF database revealed 1014 newborns with unrecorded IVF status in the EMBR. A total of 4204 newborns were verified as conceived by IVF, 24.1% of them were not categorized as born after IVF. The sensitivity of the IVF status registration was poor (75.9%), specificity (100.0%), PPV (99.8%) and NPV (99.3%) were high. The misclassifications were significantly more common among mothers of younger age or non-Estonians or with singleton birth.
Conclusion: Information based on mother's self-report or her antenatal chart does not accurately identify the newborn's IVF status. The lack of a specialized country-wide assisted reproductive technology register in Estonia requires routine record linkage of the EMBR, EHIF and the infertility treatment clinics' databases to obtain adequate information regarding IVF status in the EMBR. Electronic record linkages between databases would help considerably to improve the validity of data to be used in medical decision making, in research and for statistical purposes.
Keywords: Birth registry; Data quality; Embryo transfer; In vitro fertilization; Record linkage.