Background: Increasing attention has been given to the long-term effects of assisted reproductive technology (ART). This study assessed the validity and completeness of ART as registered in the Medical Birth Registry of Norway (MBRN) using drug prescription data from the Norwegian Prescription Database (NorPD) as reference.
Methods: In this nationwide registry validation study, we included all pregnancies recorded in the MBRN between 2005 and 2017. We estimated sensitivity, specificity, and positive and negative predictive value (PPV and NPV) of the MBRN, using data from the NorPD as reference. We obtained the total percentage of ART pregnancies that could be identified (completeness) from both registries using the capture-recapture method. We analyzed subgroups by maternal age, gestational length, mode of ART treatment, health region, and mode of registration of ART (ART institution or birth notification form).
Results: Twenty-three thousand seven hundred eighteen of a total 765,789 pregnancies were registered as ART pregnancies through the MBRN and 20,807 as ART pregnancies through the NorPD. The sensitivity of the MBRN was 85.1% (95% confidence interval [CI] = 84.7, 85.6) and the PPV was 74.7% (74.1-75.2). Sensitivity declined with increasing maternal age: 71.5% (69.4-73.7) in the age group 40-44 years, and 40.7% (22.2-59.3) in the ages above 45 years. Completeness when combining data was 96.2% (96.0-96.5).
Conclusions: Our analysis shows that, when identifying women pregnant through ART, NorPD data complemented MBRN data to obtain a more complete count of all women giving birth after ART in Norway.