RTI uses cookies to offer you the best experience online. By clicking “accept” on this website, you opt in and you agree to the use of cookies. If you would like to know more about how RTI uses cookies and how to manage them please view our Privacy Policy here. You can “opt out” or change your mind by visiting: http://optout.aboutads.info/. Click “accept” to agree.
Developing the infrastructure to assess pregnancy outcomes following vaccination
Influenza vaccines and spontaneous abortion as a use case
Kawai, A. T., Reidy, M., Zichittella, L., Stockdale, C., Longley, E., Walraven, C., Liu, C., Shoaibi, A., & Lee, G. M. (2018). Developing the infrastructure to assess pregnancy outcomes following vaccination: Influenza vaccines and spontaneous abortion as a use case. Pharmacoepidemiology and Drug Safety, 27(S2), Article 863. https://doi.org/10.1002/pds.4629
Objectives: To assess the feasibility of assessing pregnancy outcomes following vaccination with Sentinel by (1) validating algorithms for spontaneous abortion (SAB) and gestational age in pregnancies ending in live births and (2) developing a case‐time control (CTC) approach to study influenza vaccines and SAB.
Methods: The study population included pregnant women ages 18‐34 from two claims‐based Data Partners who received influenza vaccine in the 2008‐2009 or 2010‐2011 seasons. Pregnancies ending in a SAB or live birth were identified using algorithms. Because we intended to match controls (live births) to cases (SABs) by pregnancy start in the CTC, an algorithm was used to identify pregnancy start in controls. Algorithms were validated with medical record review.
Using the same cases and controls, a CTC design was implemented to examine risk of SAB following influenza vaccine in (1) the 1‐28 days post‐vaccination or (2) any time after vaccination occurring from −4 through 4, 2 through 5, or 6 through 11 weeks gestation.
Results: Of the 140 potential SAB cases identified in claims data, 97 had charts available and 53 (55%) met case confirmation criteria (ie, documentation of intrauterine pregnancy and pregnancy loss). The positive predictive value (PPV) did not vary by maternal age, code type (procedure vs. diagnosis code), or medical care setting. Of the 185 potential controls identified in claims data, 133 had medical charts available to confirm pregnancy start. The algorithm to assign pregnancy start in pregnancies ending in live births was accurate within ±14 days in 95% of pregnancies. For the exploratory CTC design, we observed an odds ratio of 0.49 (95% CI 0.15 to 1.60) for occurrence of SAB within 1‐28 days of vaccination. The odds ratios associated with vaccination from −4 through 4, 2 through 5, and 6 through 11 weeks gestation were 0.42 (95% CI 0.05, 3.61), 0.84 (0.19, 3.79), and 1.74 (0.46, 6.63), respectively.
Conclusions: Although the PPV of the pregnancy start algorithm was excellent, the moderate PPV of the SAB algorithm suggests that rigorous validation is needed to study pregnancy outcomes with Sentinel. Using a CTC approach, we successfully implemented a use case (influenza vaccines and SAB), further supporting the feasibility of conducting surveillance of pregnancy outcomes.