Medication and vaccine safety in post-approval pregnancy research: challenges and innovations

Real-world data offers a unique opportunity to address one of the most persistent evidence gaps in healthcare: the safety of medications and vaccines during pregnancy. Because pregnant people are routinely excluded from clinical trials, critical questions about maternal, fetal, and infant safety often remain unanswered at the time of regulatory approval.^1-5 As a result, the Food and Drug Administration (FDA) and the European Medicines Agency (EMA) require post-approval pregnancy safety studies to evaluate the use of drugs and vaccines during pregnancy.⁵ By combining large-scale healthcare data with advanced epidemiologic methods, researchers can generate the evidence needed to inform regulatory decision-making, clinical practice, and improve outcomes for pregnant people and their infants.

Large, integrated real-world data sources, including healthcare claims and electronic health records (EHRs), provide the scale needed to study pregnancy outcomes that would be difficult or impossible to evaluate in clinical trials. As the research arm of Elevance Health, we have healthcare claims, EHRs, and other healthcare data for approximately 100 million health plan members across the U.S. through the Healthcare Research Integrated Database (HIRD®).⁶ These resources, combined with deep expertise in safety and epidemiology, enable researchers to conduct robust post-approval pregnancy studies that support regulatory commitments and answer complex questions about medication safety.

Despite these advances, pregnancy research presents unique methodological challenges. Most pregnancy database studies identify pregnancies retrospectively from pregnancy outcome, such as a live birth or pregnancy loss.⁷ While useful for congenital malformations, this approach can introduce bias because study eligibility is determined after follow-up has effectively begun. One important example is immortal time bias, which occurs when eligibility, treatment assignment, and the start of follow-up are not properly aligned.^8,9 If left unaddressed, this bias can distort estimates of medication safety during pregnancy.

Additional complexities arise because the greatest risk of pregnancy loss occurs before most pregnancies are clinically recognized, and many major congenital malformations develop early in gestation but may not be diagnosed until birth or months or even years after birth. These timing challenges make it difficult to accurately identify both the population at risk and the outcomes of interest, underscoring the need for sophisticated epidemiologic methods that minimize bias and strengthen causal inference.

Our safety and epidemiology team is helping advance these methods while conducting dozens of pregnancy studies across many therapeutic areas. Our scientists contribute to the broader scientific community by applying innovative approaches to pregnancy identification, study design, and causal inference. Principal Scientist Dr. Stephan Lanes recently participated in an International Society for Pharmacoepidemiology (ISPE)-sponsored webinar, Avoiding Immortal Time in Studies of Medication Use During Pregnancy, while Dr. Maria Van Rompay and Dr. Michelle Delahanty will present new research on pregnancy study methods at the 2026 ISPE Annual Meeting.⁹ Together, these efforts reflect our commitment to generating high-quality, real-world evidence that supports regulators, healthcare providers, and, ultimately, pregnant patients and their families.
 

Related ISPE presentations

Monday, August 31 | 12:00–12:30 p.m. CEST

[A-327] Is disenrollment in pregnancy database studies ignorable?

[A-328] Ultrasound use during pregnancy in the Healthcare Integrated Research Database (HIRD), 2024-2025

Wednesday, September 2 | 8:30–8:45 a.m. CEST

[5A-01] When to Start Follow-up in a Database Pregnancy Study?
 

References

1. Spong CY, Bianchi DW. Improving Public Health Requires Inclusion of Underrepresented Populations in Research. JAMA (Jan 2018): pubmed.ncbi.nlm.nih.gov/29285540/.

2. Lyerly AD: Pharmacokinetic Research in Pregnancy: Ethical Low-Hanging Fruit? J Clin Pharmacol. (Jun 2023): pmc.ncbi.nlm.nih.gov/articles/PMC10275484/.

3. Lyerly AD: 'Experimental pregnancy' revisited. Theor Med Bioeth (Aug 2022): pubmed.ncbi.nlm.nih.gov/35859087/.

4. Sewell CA, Sheehan SM, Gill MS, et al.: Scientific, ethical, and legal considerations for the inclusion of pregnant people in clinical trials. Am J Obstet Gynecol (Dec 2022): pubmed.ncbi.nlm.nih.gov/35934117/.

5. European Medicines Agency. Post-authorisation safety studies (PASS) (Jul 2026): ema.europa.eu/en/human-regulatory-overview/post-authorisation/pharmacovigilance-post-authorisation/post-authorisation-safety-studies-pass. 

6. Barron JJ, Willey VJ, Doherty BT, et al.: The Healthcare Integrated Research Database (HIRD) as a Real-World Data Source for Pharmacoepidemiologic Research. Pharmacoepidemiol Drug Saf (Feb 2025): pubmed.ncbi.nlm.nih.gov/39909722/.

7. Ailes EC, Zhu W, Clark EA, et al: Identification of pregnancies and their outcomes in healthcare claims data, 2008-2019: An algorithm. PLoS One (Apr 2023): pubmed.ncbi.nlm.nih.gov/37093890/.

8. Hernan MA, Sauer BC, Hernandez-Diaz S, Platt R, Shrier I: Specifying a target trial prevents immortal time bias and other self-inflicted injuries in observational analyses. J Clin Epidemiol (Nov 2016): pubmed.ncbi.nlm.nih.gov/27237061/.

9. International Society for Pharmacoepidemiology. Avoiding Immortal Time in Studies of Medication Use During Pregnancy (Jul 2026): webed.pharmacoepi.org/products/avoiding-immortal-time-in-studies-of-medication-use-during-pregnancy.