This systematic review reports that a range of active safety surveillance methodologies for vaccines and medicines have been conducted since 2013 in most LMICs, showing a growing but somewhat uneven capacity for safety monitoring. The methodologies most commonly utilized were prospective cohort studies that featured many of the characteristics of CEM. Less frequently used methodologies were retrospective cohort studies and pregnancy exposure registries. Various electronic platforms were used for data collection and reporting, with Google Forms, and REDCap being the most common used. Strengths identified include the use of sentinel sites for data collection, frequent multi-sectoral engagement, and large sample sizes.

The articles spanned 96 LMICs, with India, China, and Brazil most represented. Africa accounted for 108 articles across 49 countries with Uganda and Ethiopia most represented. India and China have large populations that translate into larger numbers of persons exposed to vaccines and medicines, making them key sites for safety monitoring. Further, these countries each have a high burden of diseases, necessitating the use of a broad range of vaccines and medicines, including novel medical products. The LMIC PV centers experiences with spontaneous reporting potentially makes them more prepared to complement their PV programs with active safety surveillance. In Brazil, the Agência Nacional de Vigilância Sanitária (ANVISA) promotes active safety surveillance [15, 16]. A WHO report on COVID-19 Vaccine Safety Assessment Activities identified China, India, and Brazil as the leading LMICs engaged in COVID-19 vaccine safety surveillance efforts [17]. However, a systematic review found that the safety surveillance systems in these countries and other LMICs lacked robustness owing to limitations in study methodology, small sample size, inadequate technological infrastructure, a lack of data harmonization, and minimal multi-sectoral collaboration [18]. Countries such as Uganda and Ethiopia have been settings for numerous new vaccine and drug introductions, often supported by donor-funded programs, increasing the need for active safety surveillance [19].

We found that only 15% of the vaccine active safety surveillance articles were assessed using predefined criteria as having the flexibility to incorporate new vaccines. This is due, in part, to a small number of articles that used harmonized data collection forms and tools for active safety surveillance. Those that were assessed as having flexibility to incorporate surveillance of new vaccines met several criteria, such as the use of digital data collection and management systems and engagement in a multi-sectoral and/or multi-country collaboration. The present study highlights gaps in standardization and adaptability, underscoring the need for more structured and flexible surveillance systems in LMICs to enhance vaccine and medicine safety monitoring. We observed that only a small number of the articles employed the Brighton Collaboration case definitions, globally standardized criteria used to identify and classify adverse events [14]. Brighton Collaboration case definitions enable the collection of consistent and comparable safety data across studies and systems. The low adoption of these standardized definitions warrants further investigation, as their use is critical for promoting harmonization, enhancing reproducibility, and ensuring scientifically robust assessments of vaccine safety.

Approximately 34% of articles reported multi-sectoral collaborations, involving stakeholders such as local government, academia, regulatory bodies, immunization programs, and disease control programs. Coordinated efforts across diverse stakeholders enhances safety monitoring systems especially in resource-constrained settings. The conduct of active safety surveillance by NRAs is a key component for achieving and maintaining maturity level 4 under the WHO Global Benchmarking Tool, reflecting a functioning and continuously improving system [7]. Industry supported active safety surveillance in LMICs to a limited extent, often in collaboration with other organizations. For products, such as certain vaccines or antiretroviral agents, industry has funded active safety surveillance studies, often as part of capacity building initiatives and/or postmarketing commitments or requirements. To support equitable access and strengthen local production of medicinal products, greater industry engagement in active safety surveillance in LMICs should be encouraged [20]. Coordinated efforts and harmonized systems are essential to address fragmented data sharing and regulatory challenges, particularly in LMICs, ensuring efficient and equitable access to safety information [21]. However, only 10% of the articles involved multi-country collaborations, signaling missed opportunities for harmonization, cross-learning, and resource sharing across borders. Regional platforms such as the African Vaccine Regulatory Forum (AVAREF), the African Union Smart Safety Surveillance (AU-3S) program, and the Pan American Network on Drug Regulatory Harmonization could serve a larger role in fostering such initiatives [22,23,24].

One third of the included articles reported using digital technologies for data collection, management, and/or reporting of adverse events. In resource-limited settings, the adoption of digital technologies faces challenges such as limited Internet access, a lack of user-friendly mobile technology, and low digital literacy [25]. However, recent advancements have increased digital technology use in LMICs, including electronic data collection and the management, reporting, and use of various electronic health records and administrative databases through record linkage [26,27,28]. A scoping review identified the use of mobile apps, SMS, and electronic systems, which resulted in increased patient-reported adverse drug reactions and improved system integration [29]. We found that few articles used electronic health records and linked them with administrative data. In LMICs, the limited availability of electronic health records restricts the use of record linkage. However, record linkage approaches for PV using routinely generated health records and administrative databases have been shown to be pragmatic and efficient for PV [30, 31].

With respect to its use during public health emergencies, active safety surveillance was instrumental in monitoring vaccine safety during COVID-19 vaccination [32, 33]. Previously, Olsson et al. identified implementation challenges related to the limited infrastructure, shortage of trained personnel and insufficient funding that were also observed during the recent COVID-19 pandemic [34]. A published landscape review on active safety surveillance of COVID-19 vaccines identified similar study methodologies to our study. However, most of these surveillance efforts were conducted in high-income countries, with only 13 articles focusing on LMICs [33]. This disparity highlights the need for more active safety surveillance in LMICs to ensure comprehensive global vaccine safety monitoring and equitable policy development. The publication rate of articles increased from 2020 onward, likely owing to the growing focus on COVID-19 vaccine safety surveillance. A similar trend was observed by others who reported a sharp increase in COVID-19-related publications [35].

Commonly studied vaccines were COVID-19, Haemophilus influenzae type b, diphtheria, measles-mumps-rubella, pertussis, hepatitis B, polio, influenza, and tetanus. Commonly studied medicines were antiretroviral agents, antituberculosis drugs, antimalarial drugs, and antibiotics. Related to medicines, most of the active safety surveillance conducted were for global health priority diseases, emerging infectious diseases, routine immunization products, routinely used hospital products, and novel or new-to-market medicines or vaccines. Safety surveillance conducted in support of a novel or new-to-market product introduction provides critical opportunities to improve PV systems in LMICs [19].

Funding for active safety surveillance of vaccines and medicines follows different patterns across LMIC regions. In Africa, funding primarily comes from non-governmental organizations/non-profit organizations, WHO, Gates Foundation, US Government, European Union, and other foundations. Similarly, in South America, most funding consisted of government support, including the US Government, and foundation funding. In other LMICs, funding mostly was from academia, in-country governments, and the pharmaceutical industry. Funding appeared to influence the continuity and representativeness of surveillance programs with most large-scale multi-country collaborations donor funded. In contrast, surveillance funded by in-country governments or led by academia tended to be relatively small scale, with limited representation of different regions and population groups. Greater industry support for active safety surveillance generates safety evidence for input into benefit/risk assessments and contributes to building infrastructure and capacity for PV in LMICs [33]. A sustainable funding strategy for PV systems is critical for resource-limited countries to ensure robust and scalable programs. A funding pattern and model for PV systems was suggested in articles proposing a framework for sustainable funding [36], while emphasizing the importance of sufficient funding, changes of priorities by development partners can lead to near-cessation of external support for PV activities. Sustainable improvements will require continued investment in infrastructure, harmonization of regulatory standards, and regional collaboration, and increased obligations and partnerships with marketing authorization holders to support post-approval safety monitoring.

Across the active safety surveillance articles assessed, several strengths emerged consistently across study designs. Cohort event monitoring and prospective cohort studies frequently demonstrated strong follow-ups and the ability to recruit large sample sizes, which are critical for longitudinal data collection. These designs also benefited from collaborations with public health programs, systematic monitoring, and the use of real-world data to support decision making in routine healthcare settings. Pregnancy exposure registries and retrospective cohorts shared similar strengths, with some articles incorporating analytical rigor, multiple outcome tracking, and standardization of data collection. However, limitations were equally notable, including loss to follow-up, incomplete data, and limited engagement of healthcare providers and patients. Other methodological issues, including recall bias, limited or single-site study settings, and small sample sizes, were also frequently observed. These challenges, particularly in articles with retrospective or single-center designs, may hinder the generalizability and sustainability of PV efforts in LMICs. The findings underscore the need for integrating technology-enabled solutions, multi-center collaborations, and standardized protocols to enhance the effectiveness and scalability of PV systems. Similar findings were reported in previous articles, which highlighted common barriers to PV implementation, including a lack of reporting tools, low motivation, and inadequate communication mechanisms. Additional challenges reported by others included insufficient financial and human resources, weak coordination among national PV stakeholders, and poor sustainability of interventions owing to the absence of long-term planning and stable funding [29, 37]. For improving these challenges, there is a need to strengthen PV through phased infrastructure development, regulatory harmonization, community and workforce engagement, system integration, and use of harmonized digital health tools for data collection, coding, and analysis, thereby enabling reliable, comparable, and actionable safety data globally [38].

While this study was a systematic review, a formal quality assessment of the included protocols was not conducted. This decision was intentional, as our objective was not to evaluate research outcomes but to perform a landscape assessment of active vaccine and medicine safety surveillance system-related articles. To ensure methodological rigor, we implemented a structured data extraction process supported by peer validation with subject matter experts. Another limitation is that our systematic review identified but excluded abstracts that were not in English, thereby missing articles in other languages. However, we conducted searches both with and without language restrictions to English and found no significant difference in the number of articles retrieved (data not shown). Identifying those active safety surveillance articles that are presently operating may have been overstated because the definition used is that they were reported to be ongoing in the articles. Because of limited information on various attributes of the studied surveillance systems, our evaluation of their flexibility to incorporate new vaccines and the extent of multi-sectoral collaboration may not fully reflect real-world conditions. However, we utilized a structured matrix to assess these criteria as accurately as possible. This evaluation was based on factors such as ongoing systematic collection of adverse events of special interest and adverse drug reactions, use of standard case definitions, stakeholder involvement in surveillance development and implementation, authorship, funding sources, adherence to standard protocols, sample size, use of electronic technologies, multi-country collaboration, and data harmonization.