Post-market surveillance 2.0: Greater efficiency with GenAI

The world of medical devices is not just about developing and selling innovative solutions. Once a product reaches the market, the real challenge begins: monitoring its safety and performance. The post-market surveillance (PMS) process is a central component of the product life cycle and, with the introduction of the MDR (Medical Device Regulation), is an indispensable element of continuous market observation. Manufacturers must not only document, but also understand, evaluate and respond proactively – on a regular basis and based on data.

However, the traditional PMS process is resource-intensive and time-consuming, and regulatory requirements have increased significantly for manufacturers. PMS requirements vary depending on the risk class of the product. But what if generative artificial intelligence (GenAI) could transform and optimise these processes?

Post-market surveillance affects everyone

The MDR requires all manufacturers to develop, implement, document, apply, maintain and continuously update a post-market surveillance system for every medical device. A distinction is made between reactive and proactive approaches. Reactive PMS is based on passive data collection and involves reporting adverse events that have already occurred. This includes event-driven investigations, complaint handling, recalls and warnings.

Proactive measures focus on the continuous collection and analysis of data in order to identify and minimise potential safety risks at an early stage. These measures have become particularly important since the introduction of the MDR. They include regular risk assessments and ongoing product monitoring.

A PMS plan specifies in advance which data is to be collected and how it is to be evaluated.

However, as part of practical PMS work, a large number of other documents must be created which, depending on the risk class of the product, are essential for the successful implementation of the plan. These include reports on literature searches, vigilance reports, plans and reports on post-market clinical follow-up (PMCF) and the Periodic Safety Update Report (PSUR).

These documents are not only used for regulatory reporting, but also for systematic risk assessment and the identification of safety concerns. They build on the information and processes defined in the PMS plan and are crucial for identifying potential risks at an early stage and taking appropriate measures.

Pain points in the PMS process

Literature review: Time-consuming but essential

Literature review is an essential part of proactive medical device monitoring. It helps identify potential risks and is used in important reports such as PMCF and PSUR to assess the long-term safety and effectiveness of a product.

• PMCF report: This report specifically documents the results of clinical follow-up observations and uses the literature to collect data on the long-term safety and efficacy of a product and address any safety concerns raised in the scientific literature.
• PSUR: The Periodic Safety Update Report summarises the results of the entire PMS, including PMCF, and also uses the literature to collect all safety-relevant information and ensure that new scientific findings are incorporated into the safety assessment of the product.

Literature research is a time-consuming and complex task, as it involves a large number of scientific studies and articles. The process begins with a systematic search based on a defined literature search plan. This is followed by a detailed full-text analysis of the relevant studies. The final report summarises the most important publications and helps to identify potential risks. The MDR stipulates that at least two scientific databases must be used for the research, which makes the process even more time-consuming.

This is where Generative Artificial Intelligence (GenAI) offers significant relief. GenAI automates the literature research process so that relevant studies can be identified, analysed and quality-checked more quickly. This improves efficiency and ensures that no important information is overlooked.

Possible applications of GenAI:

• Automated screening of abstracts according to defined criteria
• Generation of tables with relevant study data
• Suggestions for summarising clinical evidence

GenAI takes over the research and structuring work, saving specialists valuable time. The decision-making authority remains with qualified personnel.

Data integration: From chaos to clear insight

Post-market surveillance activities involve processing large amounts of data from various sources. This includes complaints, vigilance reports and user feedback. This data is often unstructured and stored in different tools, which makes analysis difficult. The challenge is to integrate these data sources and convert them into valuable, actionable insights.

Possible applications of GenAI:

• Natural language processing (NLP) for classifying free-text feedback
• Clustering similar events for early detection of trends
• Data enrichment and keyword assignment for later analysis
• Support in prioritising risks

GenAI can be used to convert unstructured data into valuable information and prepare it optimally for audits or evaluations. This improves the efficiency and accuracy of data analysis.

Reporting: More efficient and more accurate

The creation of reports such as PSUR, PMCF and PMS reports often involves a great deal of effort. These reports must not only meet regulatory requirements, but also demonstrate that a product is being properly monitored for safety. GenAI can speed up this process and increase accuracy by automatically analysing relevant data and converting it into structured reports. This reduces sources of error and increases efficiency, as GenAI quickly extracts the necessary information and integrates it into the report.

Possible applications of GenAI:

• Automated creation of text modules for reports
• Structuring of arguments and suggestions for tables and visualisations
• Identification of missing or inconsistent information

GenAI helps to make reporting more efficient and ensure consistent and complete documentation.

PMS 2.0 – Efficient, compliant and future-proof

The MDR has not only expanded the requirements for post-market surveillance, but also made it a strategic process. In an increasingly data-driven world, it is almost impossible to meet these requirements without modern technologies. The MDR requires valid, traceable and auditable processes. This also applies to the use of GenAI. Validated workflows offer transparency and auditability and are therefore a key prerequisite for integration into PMS processes. The final evaluation, selection and interpretation of the data must be carried out by qualified personnel (‘human-in-the-loop’). Those who comply with these conditions can use GenAI as a supporting tool without compromising regulatory requirements.

GenAI offers a promising solution for making PMS more efficient and robust. It can automate literature searches, make confusing data sources analysable and generate high-quality reports. The integration of GenAI into the post-market surveillance process therefore not only improves efficiency but also ensures more accurate, data-based decision-making. The move towards data-driven, GenAI-supported monitoring can significantly improve product safety and performance in the long term.

Now it's your turn: How do you approach PMS with new technologies?

• Are you already using initial tools, or are you planning to get started?
• Where do you see the greatest potential – and the biggest hurdles?

We support you every step of the way!

GenAI can help you not only meet regulatory requirements efficiently, but also integrate PMS as a strategic process into your organisation. Together, we can examine which approaches make sense for you, which implementation steps are important and how you can make your compliance and quality processes future-proof.