Pregnancy loss <28 weeks gestation in maternal influenza and COVID-19 vaccination studies: a review of data sources, 2009-2024

OBJECTIVES: To address gaps in maternal vaccination and pregnancy loss research, large, complex datasets are needed. We aimed to identify and evaluate data sources and data collection methods currently used to capture pregnancy losses <28 weeks following maternal influenza and COVID-19 vaccination research. STUDY DESIGN: Narrative Review. METHODS: PubMed, CINAHL, Scopus, and Web of Science were used to identify studies that investigated pregnancy loss <28 gestational weeks following influenza and COVID-19 vaccinations from January 1st, 2009, to March 19th, 2024. Within the resulting studies, the data source(s) used to capture exposure and outcome data were identified and categorised. The capacity to capture and measure exposures, outcomes, and missing data within categories was investigated. RESULTS: 28 articles met the inclusion criteria, representing 1,113,878 participants. Most articles (n = 19) used multiple data sources within the one study, often obtaining exposure and outcome data from separate data sources. Categories of data sources included: registries, adverse reporting systems, medical records, and survey or interview methods. CONCLUSION: Current data collection practices and existing data sources are adversely impacting data quality, and the ability to combine large datasets necessary for analysing early pregnancy loss risk factors. This also hinders our ability to evaluate the safety of early maternal vaccination and subsequent miscarriage. Establishing pregnancy loss registries using standardised data collection and coding practices, consistent terminology, and accurate exposure and outcome timing is critical. In the absence of registries, we propose an alternative source to capture both pregnancy loss and maternal vaccination data.