Local literature monitoring is an important source of medicinal product safety information. It supports the systematic identification, assessment and, where applicable, reporting of emerging safety information published in national sources. Maintaining a robust local literature review process is essential for regulatory compliance and, ultimately, for patient safety.
There are well-established manual methods for literature monitoring, but they have their limitations. Repetitive, resource-intensive review cycles can be time consuming and may introduce inconsistencies due to human variability. To address these challenges, Medfiles transitioned from a manual local literature review process to a validated AI-assisted solution in the beginning of 2025, strengthening the consistency and efficiency of local literature monitoring.
Currently, the AI-assisted solution is used in several EU countries and the UK, supporting literature monitoring for around 500 active pharmaceutical ingredients (APIs). In line with our long-term vision for scalable pharmacovigilance support, we intend to further expand this AI-assisted literature surveillance to additional European countries in the future, strengthening our regional pharmacovigilance capabilities.
Our previous approach: manual but resource-intensive
Before adopting an AI-assisted process, Medfiles conducted regular manual reviews of local publications to identify articles relevant to products within the scope of monitoring. This process covered the selected journals and relied on careful screening by several trained pharmacovigilance professionals to ensure compliance and comprehensive case identification.
While this approach was reliable, it required significant expert resources. The work required the involvement of many trained pharmacovigilance professionals in repetitive screening activities that demanded considerable time and attention. Despite well-established procedures and quality controls, highly repetitive manual activities can be vulnerable to variability between reviewers and over time.
Although the existing process successfully met regulatory requirements, its resource-intensive nature limited opportunities to further expand monitoring activities. By introducing an AI-assisted and automated screening process, Medfiles gained the ability to review larger volumes of publications and broaden literature coverage in a scalable manner, while maintaining the same high standards of quality and regulatory compliance.
Why we introduced AI assistance
Our decision to implement an AI-assisted process was driven by the goal of achieving greater efficiency, consistency and scalability in our local literature review activities. By reducing the time spent on repetitive manual screening, we aimed to offer our clients a more cost-effective service while enabling broader literature coverage within the same operational framework. More timely identification and assessment of potential safety information support both patient safety and regulatory compliance. Using technology in the process offers meaningful automation while preserving essential human oversight where expert judgement is required.
We have validated an AI-powered platform for its intended use in our scientific and medical literature monitoring process. Using advanced natural language processing (NLP) and automation, it streamlines the literature review process.
Importantly, AI does not replace human expertise. It supports pharmacovigilance specialists by rapidly identifying potentially relevant articles, highlighting key information and supporting documentation of the review process. Medfiles’ experts retain full responsibility for reviewing the results and making the final assessment.
In practice, AI helps manage high volumes of information and routine processing tasks, allowing experts to focus their expertise where it matters the most, particularly on clinical assessment and signal interpretation. The result is an efficient workflow that combines the strengths of automation with human expertise and oversight.
How we implemented AI-assisted literature monitoring
Implementing the AI-assisted process followed a structured and carefully controlled approach. The transition began with a pilot phase, during which searches using the AI tool were run in parallel with the existing manual process to compare results and evaluate the tool’s performance in terms of accuracy and completeness. This allowed us to compare the results of the AI-assisted and manual processes and confirm that the system performed in accordance with predefined accuracy and completeness criteria before full implementation.
Training was provided to all users of the platform, covering both the technical features of the AI-assisted process and its integration into Medfiles’ existing pharmacovigilance quality management system. Once the platform had successfully completed validation for its intended use and met the predefined acceptance criteria, we transitioned to the AI-assisted workflow, while maintaining human oversight and expert judgement at every critical step.
Throughout the transition, the focus remained on maintaining regulatory compliance and data integrity. Today, AI assistance is integrated into Medfiles’ routine operations, with our experts reviewing and verifying the output before final reporting.
The benefits of AI-assisted local literature monitoring
The adoption of artificial intelligence has brought clear benefits to our pharmacovigilance operations and the services we provide to our clients.
Efficiency and expert focus:
AI-assisted screening has streamlined our literature review process, reducing both the time and resources required. Tasks that previously demanded multiple experts can now be managed by a smaller, dedicated team, allowing pharmacovigilance professionals to focus on activities requiring scientific and clinical judgement. This also improves cost-efficiency for our clients.
Consistency:
AI-assisted screening, combined with a smaller, dedicated user group, has helped reduce variability between reviewers. The standardised workflow supports consistent application of predefined screening criteria across the literature reviewed, providing clients with more consistent and reliable monitoring.
Compliance and traceability:
The AI-assisted process generates documentation supporting audit and inspection readiness, including search histories and decision logs. This strengthens traceability and transparency throughout the literature review process, supporting clients in demonstrating compliance with pharmacovigilance requirements.
Scalability and flexibility:
AI assistance enables us to expand literature coverage without a proportional increase in manual workload or costs. The solution also allows new journals and other literature sources to be incorporated efficiently based on project-specific needs and client requirements, providing clients with flexible and scalable literature monitoring services.
Combining AI and human expertise in pharmacovigilance
AI is no longer a distant concept in pharmacovigilance; it is an operational reality. Medfiles’ experience demonstrates that AI, when thoughtfully integrated into established pharmacovigilance processes and combined with human expertise, can improve efficiency, consistency and scalability in safety monitoring.
But AI is only part of the solution. At Medfiles, we see AI as a partner to our pharmacovigilance experts, supporting their work rather than replacing their expertise. Technology brings speed and scalability; our experts bring scientific judgement, regulatory knowledge and oversight.
As we continue to develop our AI-assisted workflows, this partnership between technology and human expertise will remain at the heart of our approach, helping us deliver efficient, high-quality pharmacovigilance while maintaining the scientific and regulatory standards our clients expect.
How Medfiles supports your local literature monitoring needs
With over 20 years of experience in pharmacovigilance, Medfiles provides local literature monitoring services tailored to our clients’ needs. Our services include systematic literature screening, assessment of relevant safety information and identification of potential individual case safety reports (ICSRs).
Our validated AI-assisted process currently supports literature monitoring for around 500 active pharmaceutical ingredients across several EU countries and the UK, enabling efficient and scalable monitoring while maintaining expert oversight.
Authors:
Saara Mikkola, Head of Operations, Pharmacovigilance, EU QPPV
Mirka-Maaria Erkkilä, Drug Safety and Medical Information Expert
