Your CER Is Not Finished at CE Marking: How MDCG 2025-10 Changes the PMS-PMCF-Clinical Evaluation Evidence Loop Under EU MDR
High-Impact Journals
- The Regulatory Mistake: Treating CER, PMS and PMCF as Separate Documents
- What MDCG 2025-10 Clarifies, Without Pretending It Is a New Regulation
- The Evidence Loop: Signal → Assessment → Risk → Clinical Evaluation → Action
- Not Every PMS Signal Requires a Full CER Rewrite
- Literature Surveillance Is Part of Lifecycle Evidence, Not Just a CER Search
- Three Literature Searches Teams Commonly Confuse
- The State of the Art Can Change Even When the Device Does Not
- PMCF Should Close a Defined Evidence Gap, Not Generate Data for Its Own Sake
- Why Complaint, Registry, PMCF and Literature Data Cannot Be Treated as Equivalent
- The Broken Evidence Loop: When CER, PSUR, PMCF and Risk Files Disagree
- Build a Single Clinical Evidence Traceability Matrix
- Align Evidence Cut-Off Dates and Reconcile What Cannot Be Aligned
- Worked Example: One Safety Signal Moving Through the MDR Evidence System
- Pubrica SIGNAL → ASSESS → TRACE → UPDATE → VERIFY Framework
- Reference
Interesting topics
Your CER Is Not Finished at CE Marking: How MDCG 2025-10 Changes the PMS-PMCF-Clinical Evaluation Evidence Loop Under EU MDR
A new clinical signal reaches a manufacturer in many ways. It may be a cluster of complaints, a registry report or a single published case series. Under the EU Medical Device Regulation (MDR), that signal is never a single-document event. It can touch the risk management file, the clinical evaluation report (CER), the post-market clinical follow-up (PMCF) plan, the periodic safety update report (PSUR) and, for implantable and Class III devices, the summary of safety and clinical performance (SSCP).
Most notified-body findings in this area are not about missing documents. They are about documents that disagree with each other. This article sets out what happens, document by document, when one signal appears. It also explains how to keep the whole evidence system scientifically consistent, and where MDCG 2025-10 helps.
1. The Regulatory Mistake: Treating CER, PMS and PMCF as Separate Documents
Many organisations still split evidence work by department. Regulatory affairs own the PSUR, clinical affairs own the CER, quality owns complaints and risk management, and an external writer updates the PMCF report once a year. Each document can be competent on its own and still contradicts the others.
The MDR does not allow that separation. Article 61(11) requires the clinical evaluation and its documentation to be updated throughout the device lifecycle, using clinical data from the PMS plan and the PMCF plan [1]. Article 83(3) lists what PMS data must be used for. That includes updating the benefit-risk determination and risk management, updating the clinical evaluation, and updating the SSCP [1].
In other words, the regulation describes one evidence system with several outputs. When those outputs are written on different cut-off dates, from different data extracts, by people who do not read each other’s drafts, contradictions follow. The CER says a risk is “acceptable and well characterised”. The PSUR shows a rising trend. The risk file still uses last year’s occurrence estimate. An assessor needs only a few minutes to find the gap.
2. What MDCG 2025-10 Clarifies, Without Pretending It Is a New Regulation
The Medical Device Coordination Group published MDCG 2025-10 on 19 December 2025. It is guidance on post-market surveillance under both the MDR and the IVDR [2]. It is important to be precise about its status. It adds no legal obligation. It interprets obligations already in Article 83 and Annex III of the MDR and shows how manufacturers are expected to meet them in practice [2].
Three clarifications matter most for evidence management.
- PMS is proactive. Waiting for complaints is not enough. Manufacturers are expected to look for information actively through literature screening, user feedback, registries where suitable, and PMCF [2].
- PMS output must feed other processes. The guidance describes how PMS findings flow into risk management, clinical evaluation and SSCP updates. PMS is a source for those processes, not a parallel report [2].
- Data quality comes before action. The guidance cautions against over-reacting to unverifiable information, such as social media posts, before its quality has been judged [2].
Its scope also has limits worth stating plainly. MDCG 2025-10 does not explain how to write a PSUR or PMS report; that remains the job of MDCG 2022-21 [6]. The CER and PMCF templates remain in MDCG 2020-13, 2020-7 and 2020-8 [3][4][5]. The practical value of the new guidance is that it describes how the pieces connect.
3. The Evidence Loop: Signal → Assessment → Risk → Clinical Evaluation → Action
Every post-market signal should travel the same path.
Signal → Assessment → Risk → Clinical evaluation → Action → back to PMS planning
At the signal stage, a data point or pattern suggests something may differ from what the technical documentation assumes. At assessment, the team confirms the data are real, judges whether they are clinically relevant, and decides whether they fall within or outside known and quantified risks. At risk, it updates the hazard analysis, occurrence estimates and risk controls under ISO 14971 [8], and it re-examines the benefit-risk determination. At clinical evaluation, it decides whether the CER’s conclusions on safety, performance and benefit-risk still hold. At action, it decides whether a change is needed, such as CAPA, a labelling update, a PMCF amendment or field safety corrective action. The final step, often forgotten, is feedback: the PMS plan and the PMCF plan are revised so the next cycle watches for the right thing.
The loop only works if every step leaves a record that the next step can reference. That record is what an assessor follows when they test whether your system is real.
4. Not Every PMS Signal Requires a Full CER Rewrite
Teams often react in one or two extreme ways. Some treat every signal as grounds for a new CER. Others note it in the PSUR and change nothing else. Both are wrong. The response should be proportionate to what the signal does to the CER’s conclusions.
Signal outcome after assessment | Typical CER consequence | Other documents affected |
Not confirmed, or within known frequency and severity | Document assessment; no CER change until scheduled update | PSUR or PMS report records the analysis |
Confirmed, known risk, frequency changed but acceptable | Targeted CER update of safety data, with the benefit-risk conclusion reconfirmed | Risk file (occurrence), PSUR, SSCP if values quoted |
New risk, or known risk now more severe | CER reassessment of safety conclusions and benefit-risk | Risk file, IFU/labelling, PMCF plan, SSCP, possibly vigilance |
Performance or benefit claim no longer supported | Substantial CER revision; claims may need narrowing | Labelling, IFU, SSCP, PMCF objectives, possibly intended purpose |
A targeted update is fully legitimate when it is traceable. The CER must show which section changed, which data triggered it, and why the overall conclusion still stands. What is not legitimate is a CER whose conclusions no longer match the data the manufacturer already holds.
5. Literature Surveillance Is Part of Lifecycle Evidence, Not Just a CER Search
MDCG 2025-10 lists literature screening as a proactive PMS source [2]. Those changes how literature works should be organised. The search run only when the CER is due is a snapshot. Surveillance is continuous and has its own protocol, frequency and screening criteria, set in the PMS plan.
In practice, the two activities should share a single search strategy but have different purposes. The CER search is appraised in full and drives conclusions. The surveillance search is screened on a schedule, with defined triggers for escalating a paper into signal assessment. When a relevant publication appears between CER cycles, the PMS system should catch it. Otherwise, the notified body’s reviewer may find it first.
6. Three Literature Searches Teams Commonly Confuse
Clinical evaluation relies on three distinct kinds of literature evidence, and a combined search usually serves none of them well [9].
Search | Question it answers | Typical sources | Where it lands |
Device-specific | What do published data show about our device? | Databases, registries, trial registers, own study publications | CER safety and performance data; PMCF evaluation |
Equivalent or similar device | What do data on comparable devices show, and are they usable? | Same databases, with equivalence criteria applied [10] | Equivalence justification, or benchmarks for similar devices |
State of the art | What is current accepted practice and outcome in this clinical field? | Guidelines, systematic reviews, registries, alternative therapies | Acceptance criteria, benefit-risk benchmarks, CER context |
Mixing them causes real problems. A state-of-the-art paper is not evidence about your device, and a device-specific series cannot define what “acceptable” looks like in the field. Each search needs its own question, eligibility criteria and appraisal method. Each also needs its own date range, recorded in the CER.
7. The State of the Art Can Change Even When the Device Does Not
A device can go unchanged for five years and still have a weaker CER at the end of that period. The benchmark it is judged against has moved.
State of the art shifts when clinical guidelines are revised, when alternative therapy shows better outcomes, or when registries publish new revision or complication benchmarks. When that happens, the CER’s acceptance criteria may no longer be current. A complication rate that was “comparable to the state of the art” may now sit above it. No complaint or vigilance report will flag that change. Only deliberate state-of-the-art surveillance will.
This is why the PMS plan should name the state-of-the-art sources to be monitored and how often. It is also why every CER update should re-read its acceptance criteria against the current benchmark rather than copying them forward.
8. PMCF Should Close a Defined Evidence Gap, Not Generate Data for Its Own Sake
MDR Annex XIV Part B describes PMCF as a continuous process that updates the clinical evaluation [1]. It aims to confirm safety and performance over the device’s expected lifetime, to identify emerging risks and to keep the benefit-risk profile acceptable. The logic is simple. PMCF starts where the CER’s uncertainty sits.
A defensible PMCF plan, built on the MDCG 2020-7 template [3], links each activity to a named residual uncertainty in the CER. Each activity carries a measurable objective, an acceptance criterion and a pre-defined action if that criterion is not met. “Collect user feedback on device performance” is not an objective. “Confirm that the two-year revision rate stays below the registry benchmark” is.
The PMCF evaluation report [4] must then report against those objectives and state whether each gap is closed, narrowed or still open. When a new signal appears, the first question is whether the current PMCF activities can answer it. If they cannot, the PMCF plan changes.
9. Why Complaint, Registry, PMCF and Literature Data Cannot Be Treated as Equivalent
A common weakness in CERs and PSURs is pooling data sources as if they measured the same thing. They do not.
Data source | Strength | Main limitation | Appropriate use |
Complaints and vigilance | Fast, device-specific, captures serious events | Under-reporting; no reliable denominator | Signal detection; trend monitoring |
Registries | Large denominators, long follow-up, real practice | Coding limits; may not separate device variants | Rates and benchmarks; revision and survival |
PMCF studies | Defined population, planned endpoints | Small samples; selected centres | Confirming specific objectives |
Published literature | Independent; covers competitors and the state of the art | Publication bias; often not device-specific | Context, benchmarks, emerging signals |
A complaint rate cannot be compared directly with a registry revision rate. They differ in numerator, denominator and detection. The CER should say which source supports each conclusion and why that source is fit for that purpose.
10. The Broken Evidence Loop: When CER, PSUR, PMCF and Risk Files Disagree
Across CER, PMS and PMCF projects, the same disconnects recur. The taxonomy below is Pubrica’s working classification of this disconnects evidence, drawn from recurring review themes rather than a formal statistical survey.
Disconnect | What it looks like | Why it matters |
Frequency mismatch | Risk file occurrence differs from the rate in the PSUR or CER | Undermines the benefit-risk conclusion |
Stale conclusion | CER conclusion predates a PSUR trend it does not mention | Shows PMS data not feeding clinical evaluation [1] |
Orphan PMCF | PMCF objectives not linked to any CER uncertainty | PMCF cannot demonstrate purpose [3] |
Frozen benchmark | Acceptance criteria copied from an old state-of-the-art review | Claims judged against outdated standards |
SSCP drift | SSCP quotes figures no longer in the CER | Public-facing document becomes inaccurate [7] |
Date collision | Documents use different, unexplained data cut-off dates | Reviewer cannot reconcile numbers |
Label lag | New risk in the risk file not reflected in the IFU | Risk control incomplete under ISO 14971 [8] |
Most of these are process failures, not scientific failures. They can be prevented with one shared tool.
11. Build a Single Clinical Evidence Traceability Matrix
| ID | Item | Evidence source and cut-off | Risk file | CER | PMCF | PSUR | SSCP | IFU/label | Status |
| C-01 | Primary performance claim | PMCF study, registry; date | — | Performance | Objective 1 | Benefit-risk | Clinical performance | Indications | Current |
| R-04 | Known risk: early loosening | Complaints, registry, literature; date | Hazard 12 | Safety | Objective 2 | Trends | Residual risks | Warnings | Under review |
| U-02 | Long-term survival beyond 5 years | Literature only; date | — | Uncertainties | Objective 3 | PMCF summary | — | — | Open |
12. Align Evidence Cut-Off Dates and Reconcile What Cannot Be Aligned
Documents in the loop run on different cycles. Under Article 86, PSURs are updated at least annually for Class IIb and III devices and at least every two years for Class IIa [1]. For Class III and implantable devices, Article 61(11) requires the PMCF evaluation report, and the SSCP where indicated, to be updated at least annually [1]. CERs are often updated on their own schedule.
Perfect alignment is rarely possible, so the goal is reconciliation. Each document should state its data cut-off date for every data source. Where two documents use different cut-offs, the latter one should say what has changed since the earlier one. A short reconciliation note serves this purpose, stating for example: “Since the CER cut-off, two additional events were reported; neither changes the conclusion because…”. It turns an apparent contradiction into evidence of a working system.
13. Worked Example: One Safety Signal Moving Through the MDR Evidence System
The following example is illustrative and uses hypothetical figures.
A German manufacturer markets a Class III cementless knee implant. Its CER concluded that aseptic loosening occurred at a rate consistent with state of the art. This conclusion was based on registry benchmarks and a PMCF cohort.
- Quarterly PMS review shows three loosening complaints in six months, against a historical average of one. In the same quarter, the literature surveillance picks up a published single-centre series reporting early loosening with a similar design.
- The team verifies the complaints and requests registry data from the German Arthroplasty Registry (EPRD). It finds that the implant’s two-year revision rate for loosening is now slightly above the registry benchmark for its class. The finding is confirmed and clinically relevant. It concerns a known risk, but at a higher frequency than assumed.
- The risk file’s occurrence estimate for this hazard is raised. Risk controls are reviewed, and a surgical-technique factor (bone preparation) is identified. The benefit-risk determination is re-run and remains favourable, but with a narrower margin.
- Clinical evaluation. A targeted CER update revises the safety data and the state-of-the-art comparison, adds the new publication, and reconfirms the benefit-risk conclusion with its reasoning. A full rewrite is not needed.
- The surgical technique guide and IFU are updated, and surgeon training is added as CAPA. The PMCF plan aims to monitor loosening at two years against the registry benchmark, with a defined escalation threshold.
- The next PSUR describes the trend, the analysis and the actions [6]. The SSCP’s residual-risk figures are updated to match the CER [7]. The traceability matrix row for this risk is updated, with the cut-off date and document versions.
Each document now tells the same story. Each also cites the same data, with the date of each extract.
14. Pubrica SIGNAL → ASSESS → TRACE → UPDATE → VERIFY Framework
The framework turns the evidence into five checkpoints that a manufacturer can audit.
Stage | Key question | Output | Check before moving on |
SIGNAL | Has new information been captured from every planned PMS source, including literature and the state of the art? | Logged signal with source, date and data quality note | Source is verifiable; not acted on prematurely |
ASSESS | Is it real, clinically relevant, and within or outside known risks? | Signal assessment record with decision rationale | Clinical input documented; decision proportionate |
TRACE | Which claims, risks and documents do it touch? | Updated traceability matrix rows | Every affected document identified |
UPDATE | What must change, and how much? | Targeted or full updates to risk file, CER, PMCF plan, PSUR, SSCP, IFU | Cut-off dates stated; reconciliation note added |
VERIFY | Do all documents now agree? | Cross-document consistency check signed off | Numbers, conclusions and dates match across the set |
Does every CER conclusion cite data no older than the latest PSUR? Does every PMCF objective map to a named uncertainty in the CER? Do risk file occurrence estimates match the rates reported in the PSUR? Were the state-of-the-art acceptance criteria re-checked at the last CER update? Does the SSCP quote only figures that appear in the current CER? Does every document state its data cut-off date? Is there a written record for every signal assessed since the last CER, including those closed with no action?
Germany in context. None of this is uniquely German, because the MDR applies across the EU. Germany does, however, combine a large manufacturing base, demanding notified bodies, BfArM as the competent authority, and strong national registries such as the EPRD. Together, these make evidence consistency a practical, day-to-day concern for German manufacturers.
Request for an MDR Clinical Evidence Consistency Review Pubrica’s clinical and regulatory writers review your CER, PMS plan, PMCF plan and report, PSUR, risk file and SSCP as one system. You receive a completed Clinical Evidence Traceability Matrix, a list of evidence disconnects ranked by risk, a cut-off date reconciliation, and clear recommendations on which documents need a targeted update, and which need more.
Get expert support from Pubrica for addressing reviewer comments, preparing point-by-point rebuttals, and strengthening manuscripts for journal resubmission. Prepare your revised manuscript with confidence.
Conclusion
Under the EU MDR, clinical evaluation continues beyond CE marking. PMS, PMCF, CER, risk management, PSUR and SSCP must remain connected throughout the device lifecycle. MDCG 2025-10 reinforces this evidence flow. The key is simple: assess every signal, trace its impact, update affected documents, and verify that all conclusions remain consistent and evidence-based.
References
- European Parliament & Council of the European Union. (2017). Regulation (EU) 2017/745 of the European Parliament and of the Council of 5 April 2017 on medical devices. Official Journal of the European Union, L 117, 1–175. https://eur-lex.europa.eu/eli/reg/2017/745/oj
- Medical Device Coordination Group. (2025). MDCG 2025-10: Guidance on post-market surveillance of medical devices and in vitro diagnostic medical devices. European Commission. https://health.ec.europa.eu/latest-updates/mdcg-2025-10-guidance-post-market-surveillance-medical-devices-and-vitro-diagnostic-medical-devices-2025-12-19_en
- Medical Device Coordination Group. (2020a). MDCG 2020-7: Post-market clinical follow-up (PMCF) plan template. European Commission. https://health.ec.europa.eu/medical-devices-sector/new-regulations/guidance-mdcg-endorsed-documents-and-other-guidance_en
- Medical Device Coordination Group. (2020b). MDCG 2020-8: Post-market clinical follow-up (PMCF) evaluation report template. European Commission. https://health.ec.europa.eu/medical-devices-sector/new-regulations/guidance-mdcg-endorsed-documents-and-other-guidance_en
- Medical Device Coordination Group. (2020c). MDCG 2020-13: Clinical evaluation assessment report template. European Commission. https://health.ec.europa.eu/medical-devices-sector/new-regulations/guidance-mdcg-endorsed-documents-and-other-guidance_en
- Medical Device Coordination Group. (2022a). MDCG 2022-21: Guidance on periodic safety update report (PSUR) according to Regulation (EU) 2017/745. European Commission. https://health.ec.europa.eu/medical-devices-sector/new-regulations/guidance-mdcg-endorsed-documents-and-other-guidance_en
- Medical Device Coordination Group. (2022b). MDCG 2019-9 Rev.1: Summary of safety and clinical performance: A guide for manufacturers and notified bodies. European Commission. https://health.ec.europa.eu/medical-devices-sector/new-regulations/guidance-mdcg-endorsed-documents-and-other-guidance_en
- International Organization for Standardization. (2019). Medical devices: Application of risk management to medical devices (ISO Standard No. 14971:2019). https://www.iso.org/standard/72704.html
- European Commission. (2016). MEDDEV 2.7/1 revision 4: Clinical evaluation: A guide for manufacturers and notified bodies under Directives 93/42/EEC and 90/385/EEC. https://health.ec.europa.eu/medical-devices-sector/new-regulations/guidance-mdcg-endorsed-documents-and-other-guidance_en
- Medical Device Coordination Group. (2020d). MDCG 2020-5: Clinical evaluation – Equivalence: A guide for manufacturers and notified bodies. European Commission. https://health.ec.europa.eu/medical-devices-sector/new-regulations/guidance-mdcg-endorsed-documents-and-other-guidance_en