--- title: "Where Pharmacovigilance Meets Materiovigilance: AI Safety in Device and Combination Product Trials" description: "Discover how AI supports safety surveillance in medical device and combination product trials, and what sponsors must consider for governance and compliance." url: https://qbdgroup.com/en/blog/ai-safety-device-combination-product-trials type: "Blog post" language: en published: 2026-07-08 author: "Anam Ahmad" category: "Pharmacovigilance" publisher: "QbD Group" citation: "QbD Group, \"Where Pharmacovigilance Meets Materiovigilance: AI Safety in Device and Combination Product Trials\", https://qbdgroup.com/en/blog/ai-safety-device-combination-product-trials" --- # Where Pharmacovigilance Meets Materiovigilance: AI Safety in Device and Combination Product Trials > Discover how AI supports safety surveillance in medical device and combination product trials, and what sponsors must consider for governance and compliance. Most conversations about artificial intelligence in clinical trial safety focus on medicinal products. Yet some of the most demanding and transformative applications are emerging elsewhere, at the intersection of medical devices, combination products, digital health technologies, and traditional pharmacovigilance. As clinical research becomes increasingly connected, decentralised, and data-driven, safety monitoring is evolving beyond periodic case review. Sponsors and CROs are now managing continuous streams of data generated by wearable devices, remote monitoring systems, Software as a Medical Device (SaMD), digital therapeutics, and AI-enabled diagnostics. The challenge is straightforward. **The volume of safety-relevant data has grown beyond the capacity of traditional manual review processes.**

In This Blog Post

- Why modern device trials create new safety surveillance challenges - Where AI can add value in medical device and combination product studies - How pharmacovigilance and materiovigilance are beginning to converge - What governance and validation expectations still apply - How sponsors and CROs can prepare for increasingly integrated safety ecosystems
## The Data Challenge in Modern Device Trials Medical device and combination product studies increasingly generate continuous streams of patient data. Examples include: - Wearable physiological sensors - Software as a Medical Device (SaMD) platforms - Digital therapeutics - Remote telemetry systems - AI-enabled diagnostic technologies Unlike traditional clinical datasets, these technologies often produce information continuously rather than at predefined study visits. A safety team can review a batch of case report forms. It cannot manually review every second of a continuous telemetry stream. As data volumes increase, identifying clinically meaningful signals becomes increasingly difficult. The challenge is no longer collecting information. **The challenge is finding the signal within the noise.** ## Where AI Can Support Device and Combination Product Safety This is where AI begins to demonstrate its practical value. Many of the challenges associated with device safety surveillance align naturally with the strengths of machine learning and advanced analytics. Potential applications include: - Device deficiency trend analysis - Anomaly detection across telemetry data - Cybersecurity signal monitoring - Real-time patient deterioration alerts - Integration of multi-source physiological monitoring data Each of these activities involves analysing large volumes of high-velocity data that exceed the capabilities of manual review. Real-time deterioration alerts are particularly interesting because they point toward a future where safety surveillance becomes increasingly proactive rather than retrospective. Instead of reviewing events after they occur, organisations may be able to identify emerging risks while intervention is still possible. ## Where Pharmacovigilance and Materiovigilance Begin to Converge One of the most important developments in modern safety management is the growing overlap between disciplines that were historically managed separately. The boundaries between: - Pharmacovigilance - Materiovigilance - Digital health surveillance - Cybersecurity risk management are becoming increasingly blurred. Combination products and connected health technologies often require organisations to evaluate safety signals that span multiple regulatory domains simultaneously. For sponsors and CROs, this is no longer a theoretical discussion. **It is becoming an operational reality.** The teams, processes, and systems responsible for medicinal product safety and device vigilance can no longer function in complete isolation. As digital and AI-enabled technologies continue to evolve, safety management models are likely to become increasingly integrated across medicinal product and medical device environments. ## The Governance Principles Remain the Same Although the technology is changing, the underlying governance principles remain remarkably consistent. AI-enabled surveillance systems still require: - Continuous performance monitoring - Periodic revalidation - Documented intended use - Explainability - Inspection readiness These controls become even more important as protocols evolve, reporting patterns change, and new safety profiles emerge. Model performance that was appropriate during one phase of development may not remain reliable indefinitely. Governance therefore cannot be treated as a one-time validation exercise. **It must remain a lifecycle activity.** ## Data Protection and Compliance Cannot Be an Afterthought Medical device and digital health studies frequently process highly sensitive patient information. Safety datasets may include: - Identifiable patient narratives - Physiological monitoring data - Telemetry outputs - Free-text clinical documentation As a result, organisations must carefully address: - GDPR compliance requirements - Pseudonymisation strategies - Vendor hosting arrangements - Cross-border data transfers - Cybersecurity controls The increasing use of AI does not reduce these obligations. In many cases, it increases scrutiny regarding how patient data is collected, processed, stored, and monitored. ## Validation Should Reflect Risk Not all AI applications carry the same regulatory or patient safety impact. Validation strategies should therefore remain proportionate to risk. For example: - An AI tool that drafts safety narratives presents one level of risk - An anomaly-detection system monitoring real-time patient deterioration presents another The controls, oversight mechanisms, and validation activities applied to each system should reflect the potential impact on patient safety and regulatory compliance. Risk-based validation remains one of the most important principles for organisations adopting AI within regulated environments. ## The Future of Safety Surveillance Is Cross-Functional The direction of travel is becoming increasingly clear. As trials become more decentralised, digitally enabled, and device-integrated, safety data volumes will continue to grow. AI will become a more central component of safety surveillance strategies. The organisations most likely to succeed will not be those that simply deploy new technology. They will be the ones that bring together expertise across: - Pharmacovigilance - Quality assurance - Regulatory affairs - Clinical operations - Medical device compliance - Digital health AI adoption in these environments is not primarily a technology initiative. **It is a cross-functional quality and governance initiative.** Translating that into practice means: - Defining intended use before selecting a solution - Embedding validation requirements into vendor agreements - Establishing governance frameworks early - Training teams on automation bias as well as system functionality This multidisciplinary approach is increasingly essential as safety domains continue to converge. ## Building AI Governance Across Converging Safety Domains As medical devices, digital health technologies, and medicinal products become more interconnected, safety surveillance models must evolve accordingly. --- Source: https://qbdgroup.com/en/blog/ai-safety-device-combination-product-trials — © QbD Group. Quote freely with attribution and a link back.