Adverse Event Following Maternal Immunization (AEFMI) Causality Assessment Methodology and Software Development
The GACVS was provided with an in-depth overview of the development and plans for implementation of a novel causality assessment methodology for the evaluation of AEFMI prepared by global experts, WHO and five pilot countries - Argentina, Australia, Canada, Germany and India. This initiative was driven by global requests to develop standardized methods for assessing potential associations between vaccination and subsequent potential vaccine associated adverse events and/or pregnancy outcomes—during a period that is inherently prone to pregnancy-related events, such as miscarriage and preterm birth, which may occur coincidentally and not be causally related to vaccination. Given the ongoing evolution of vaccine recommendations in pregnancy and increasing maternal vaccine uptake worldwide, as well as the prospect for new maternal vaccines on the horizon, the causality assessment methodology and software aims to provide a framework to investigate and distinguish coincidental outcomes from those potentially causally related to immunization in pregnancy.
The newly developed AEFMI methodology builds on existing WHO causality assessment frameworks introducing critical adaptations to account for the uniqueness of maternal, fetal, neonatal and pregnancy-related outcomes. A key innovation is the shift in emphasis toward the case investigation, with approximately 70% of the process focused on thorough field-level data collection, followed by expert review that categorizes the event utilizing existing validated case definitions (Global Alignment of Immunization Safety Assessment in Pregnancy (GAIA) Brighton Collaboration case definitions). The methodology includes a redesigned classification system to incorporate pregnancy outcomes as AEFMI, as well as known adverse events following immunization (AEFI), with addition of new subcategories in the final event-assessment tool to include AEFMI. These will help distinguish between indeterminate and clearly related or unrelated events, based on existing evidence. Supporting this process is a custom-developed software that enables rational systematic data collection to conduct event/outcome -specific investigations, followed by expert notification, case assessment, classification, and automated report generation.
The five participating countries pilot-tested the methodology and software and provided detailed feedback, which was shared with GACVS. They recognized the tool’s value in improving standardization, transparency, and safety monitoring of vaccines administered in pregnancy, particularly in low- and middle-income country (LMIC) settings. Argentina and Canada indicated plans to propose integration of the tool into national vaccine safety systems, while Germany proposed its international application. India highlighted the need to contextualize findings due to high background prevalence of adverse pregnancy outcomes independent of vaccination, suggesting a dual focus on individual case assessment and population-level trend analysis. While the feedback was overwhelmingly positive, some challenges were identified. Australia flagged issues around software usability and security, while some countries underscored the need for guidance and flexibility in the estimation of gestational age, and allowing manual review of data.
GACVS provided feedback on the interpretation of the new classification categories—particularly B1 (temporally related, but insufficient evidence), B2 (conflicting evidence from literature or investigation), and C2 (common pregnancy outcomes proven not to be associated with the vaccine) —noting that these may be interpreted as including events that fall within indeterminate zones. This underscored the need for clear explanations with examples including pre- pregnancy risk factors in the accompanying guidance documents. During the meeting, it was agreed that specific outcomes and vaccines require focused, case-by-case evaluation, with classification guided by existing evidence and established case definitions. Participants stressed the importance of rigorous training for both field investigators and expert causality assessors, and emphasized that accurate diagnosis, including confirmation of pregnancy and pregnancy onset, timing of exposure during gestation, and streamlining workflow of AEFI with AEFMI are crucial for appropriate classification. The limitations of spontaneous adverse event reporting in assessing causal links were identified and will be considered in the causality assessment of AEFMI.
There was consensus on the necessity of coupling AEFMI causality assessment process with effective public communication strategies to explain findings—especially when results are inconclusive—to avoid misinterpretation and maintain public trust. The importance of integrating AEFMI causality assessments with existing surveillance and reporting systems was also noted. Additionally, participants recommended prioritizing specific pregnancy outcomes for further evaluation based on public health significance and available background incidence rates.
Overall, the session affirmed the relevance and robustness of the AEFMI causality assessment methodology and digital tools, while recognizing areas for refinement. GACVS suggested finalizing the guidance manual, developing training materials, improving software features and continuing pilot testing to refine the causality assessment platform. The committee suggested expanding implementation to equip national programs with the tools needed to ensure maternal immunization is monitored safely and effectively, especially as newer vaccines for maternal immunization (e.g. Influenza, Tetanus, diphtheria & acellular pertussis (Tdap), Respiratory Syncytial Virus (RSV) and Group B Streptococcus (GBS)) are introduced globally, particularly in LMIC settings.
Safety monitoring of maternal and newborn health
The Committees were informed of recent work to
strengthen data quality and harmonize data collection for safety monitoring of medicines and vaccines in
pregnant and breastfeeding women and neonates,
including the results of a pilot study of the WHO minimum maternal and newborn health data sets (mMNHDS).
The Committees discussed support for harmonization
of definitions and methods used in monitoring maternal and newborn health and use and improvement of
existing data sources, particularly on means of collection, such as electronic clinical records and pregnancy
exposure registries.
The Committees recommended that the mMNHDS be
enhanced by developing an implementation strategy,
including it in registers, creating training materials for
health-care professionals and data managers, and advocating for its dissemination (core and catalogue data
sets).
Updates were provided on harmonization of data for monitoring maternal, fetal and newborn outcomes and on analysis of pregnancy exposure registries in low– middle-income countries. An overview was provided of work by WHO on assessment of causality for adverse events following maternal immunization (AEFMI).
The WHO PVG provided an overview of a project for harmonizing data for monitoring maternal, fetal and newborn outcomes in a set of standardized indicators and data elements. An interdepartmental team comprising 12 WHO departments developed a minimum maternal and newborn health dataset (WHO mMNHDS), which comprises 15 core indicators to be collected in all WHO regions and countries and 62 complementary “catalogue” indicators. The next stage is to assess collection of the WHO mMNDHS in settings with limited resources. A pilot study is under way in nominated hospitals in 8 countries eligible for funds from the Global Fund for HIV, Malaria and Tuberculosis.
The Committee was also provided with details of a project to improve causality assessments for AEFMIs involving WHO, the International Clinical Epidemiology Network in India, the AEFI Committee, the National Centre for Immunisation Research and Surveillance in Australia, the University of Buenos Aires in Argentina and the Paul-Ehrlich-Institut in Germany. The expected outcomes of the work are collection of data on the mother, fetus and the neonate at the time of AEFMI reporting and investigation. Software has been developed to guide systematic, standard causality assessments of AEFMIs for use by members of AEFI or AEFMI committees, physicians, immunization programme managers and others and as an education and research tool for various stakeholders. The new approach is intended to reduce errors and ensure better understanding of AEFMIs. This work is anticipated to be completed by the end of 2025.