Short answer

MDCG 2019-16 Rev.1 (July 2020) explains how to meet the EU MDR’s IT security requirements (Annex I, sections 17.2 and 17.4), which apply to all software devices, not only connected ones; the depth of analysis scales with connectivity and risk. Notified Bodies review this in depth. In practice they expect a threat model, security risk management, security testing (for connected SaMD usually an external penetration test), an inventory of third-party components such as an SBOM, and post-market vulnerability monitoring within your QMS. For AI medical devices, Article 15 of the EU AI Act adds AI-specific attacks from 2 August 2028.

Why cybersecurity has moved to the top of NB review checklists

A few years ago, cybersecurity documentation was often an afterthought in SaMD technical files — a single paragraph in the risk management file acknowledging that software could be vulnerable to attack. Today, in our experience, Notified Bodies review it in depth, often with specialist reviewers and structured questions built around MDCG 2019-16. Technical files that treat cybersecurity as a checkbox tend to collect major findings.

The requirements come from several directions. The MDR's general safety and performance requirements in Annex I require software to be developed according to the state of the art, taking into account risk management including information security (Section 17.2); require manufacturers to set out minimum requirements for hardware, IT network characteristics and IT security measures, including protection against unauthorised access (Section 17.4); and require devices to be protected, as far as possible, against unauthorised access that could hamper their intended function (Section 18.8). MDCG 2019-16 Rev.1 (July 2020) explains how to meet these requirements. The EU AI Act adds AI-specific cybersecurity requirements for high-risk AI systems — for AI medical devices from 2 August 2028, not earlier. And your IEC 62304 software life-cycle processes are where security requirements, SOUP management and patching have to live in practice. For connected SaMD, cybersecurity is not optional — it is a first-class certification requirement.

MDCG 2019-16: the framework

MDCG 2019-16 (Rev.1, July 2020) structures cybersecurity for medical devices around pre-market and post-market activities, consistent with how other medical device safety requirements are structured. The practices below are how we see SaMD teams meet those expectations.

Pre-market: security by design

Security by design means that cybersecurity requirements are defined and implemented during development — not added as a compliance exercise before submission. In practice, this means you:

  • Define a security architecture that identifies network interfaces, data flows, trust boundaries, and security controls at the design level. For SaMD, this means documenting how data enters and exits the system, what authentication and access controls are in place, and how data is protected in transit and at rest.
  • Conduct threat modelling using a structured methodology (STRIDE, PASTA, or equivalent) to systematically identify potential attack vectors, the assets they could compromise, and the clinical harm that could result. The threat modelling must be documented and must connect to the security risk management process.
  • Define minimum security requirements — the security features that the device must implement: authentication and authorisation, encryption for data in transit and at rest, audit logging, session management, and vulnerability management. These requirements must appear in the software requirements specification and be traceable to test evidence.

Security risk management

The security risk management process runs parallel to the ISO 14971 safety risk management process and should be documented in the Security Risk Management File. For each identified threat:

  • Estimate the likelihood of successful exploitation (considering attacker capability, attack surface, and existing controls)
  • Evaluate the potential impact on patient safety, device functionality, and data confidentiality/integrity
  • Document security controls that mitigate the threat
  • Assess residual security risk after controls
  • Conclude on acceptability using pre-defined security risk acceptability criteria

Security testing

Security requirements need verification evidence like any other requirement, and the depth of security testing should match the device's connectivity and risk profile. For most connected SaMD, this means:

  • Penetration testing: Conducted by qualified security testers on a representative production system. The test scope should cover the threat model. Test reports must document scope, methodology, findings, severity ratings, and remediation status.
  • Vulnerability scanning: Automated and manual vulnerability scanning of the software codebase and dependencies. OWASP Top 10, CVE database checking for known vulnerabilities in SOUP components.
  • Security test cases in the V&V suite: Specific test cases in your software system testing that verify security requirements are met — authentication works correctly, access controls enforce separation of privilege, injection attacks are blocked.
Want your cybersecurity file checked before the Notified Body sees it? A technical documentation gap analysis reviews your security risk management, testing evidence and post-market process against MDR Annex I and MDCG 2019-16.
See the gap analysis →

Post-market cybersecurity obligations

Cybersecurity does not stop at certification. Your MDR post-market surveillance and vigilance obligations apply to security issues like any other risk, and MDCG 2019-16 covers the post-market phase as well. In practice, post-market cybersecurity management must be integrated into your QMS and post-market surveillance process:

  • Vulnerability monitoring: Active monitoring of vulnerability databases (NVD, vendor security advisories, CERT/CC) for newly discovered vulnerabilities in your SOUP components and underlying platforms. Many SaMD teams do not have a formal process for this — and the result is deployed devices running software with known critical vulnerabilities.
  • SBOM maintenance: The Software Bill of Materials (or equivalent SOUP list) must be kept current and must feed into vulnerability monitoring — including planning for operating system and platform end-of-life. Other regulators point the same way: Singapore HSA's software guidance (GL-04 R4, December 2025) stresses maintaining and updating a cybersecurity plan across the software life cycle, and HSA's Best Practices Guide on Medical Device Cybersecurity (GL-10, Revision 1 dated August 2026) covers cybersecurity specifically.
  • Coordinated vulnerability disclosure: A documented process for receiving and responding to security vulnerability reports from external researchers or customers, with a published security contact and a defined response process. This is good practice that reviewers increasingly look for.
  • Security patch management: A defined process for evaluating, developing, testing, and deploying security patches, integrated with your change control process. Security patches that affect safety-critical functionality require the same change control rigor as clinical functionality changes.

EU AI Act cybersecurity obligations (AI medical devices: from 2 August 2028)

For AI-powered SaMD, the EU AI Act adds cybersecurity requirements that partially overlap with the MDR and MDCG 2019-16. They are part of the high-risk requirements, which — after Regulation (EU) 2026/1744 — apply to AI medical devices (Article 6(1), Annex I) from 2 August 2028; nothing in this area applies to AI medical devices from August 2026. Article 15 requires high-risk AI systems (which include AI SaMD in Class IIa and above) to achieve an appropriate level of accuracy, robustness and cybersecurity, and to be resilient against attempts by unauthorised third parties to alter their use, outputs or performance by exploiting system vulnerabilities.

Article 15(5) names the AI-specific vulnerabilities that technical solutions should address where appropriate: data poisoning (manipulation of training data), model poisoning (manipulation of pre-trained components), adversarial examples or model evasion (inputs designed to make the model err), confidentiality attacks (for example, model inversion to reconstruct training data) and model flaws. MDCG 2019-16 predates the AI Act and does not cover these attacks in detail; since the MDR already requires security according to the state of the art, it makes sense to add them to the security risk assessment for AI SaMD now rather than in 2028. Our guide to the EU AI Act for medical software covers the full timeline.

Connecting cybersecurity to your ISO 14971 risk file: MDCG 2019-16 describes how security risks that could affect safety feed into safety risk management, so they belong in your ISO 14971 risk management file — not just in a separate security document. A cybersecurity vulnerability that could cause the software to produce incorrect diagnoses is a patient safety risk and must appear in both documents.

If you want an independent check of the security sections of your technical file before your Notified Body sees them, a technical documentation gap analysis covers them alongside the rest of the file.

Official sources & references