Modeling exposures of medications used episodically during pregnancy: Triptans as a motivating example

Purpose To assess the validity of dispensed prescription to classify exposure to medications used episodically during pregnancy, and to explore individual trajectories of episodic medication use across pregnancy, using triptans for migraine as the motivating example. Methods We compared self‐reported triptan use during pregnancy in The Norwegian Mother, Father and Child Cohort Study (MoBa) to dispensed prescriptions in The Norwegian Prescription Database and calculated Cohen's kappa coefficient (κ), sensitivity, specificity and predictive values using MoBa as reference standard. We used group‐based trajectory modeling to estimate exposure trajectories in MoBa according to probability of triptan use across pregnancy. Results We identified 6051 pregnancies where mothers filled at least one triptan prescription or reported migraine or triptan use in the 6 months before or during pregnancy. Sensitivity of prescribed triptans during pregnancy was low (39.1%), but specificity was quite high (95.4%). Agreement between the two data sources was fair (κ 0.36). We identified three trajectory groups in MoBa including constant‐high, decreasing‐medium and decreasing‐low probability of triptan use across pregnancy. Conclusions Using dispensed prescriptions rather than self‐report to classify exposure to triptans during pregnancy is likely to result in substantial under‐estimation of exposure. In this study, traditional definitions of ever‐exposed vs never‐exposed failed to capture variations in drug utilization during pregnancy.