PMCF Studies: When Surveys Are Not Enough

Written by HATEM RABEH, MD, MSc Ing

Your Clinical Evaluation Expert And Partner

in
S

This happens more often than it should. Surveys feel productive. They are relatively easy to design, inexpensive to execute, and they generate quantitative data that looks convincing in a report. The problem is that surveys rarely capture what the MDR requires from post-market clinical follow-up.

The confusion comes from a fundamental misunderstanding of what PMCF is supposed to achieve.

What the MDR Actually Requires from PMCF

Article 61 of the MDR and Annex XIV Part B are clear. PMCF must confirm the safety and performance of the device throughout its entire lifetime. It must identify emerging risks. It must verify that the residual risk remains acceptable. And it must ensure that the clinical benefit continues to outweigh the risks under normal conditions of use.

This is not about satisfaction. It is not about user preference. It is about clinical outcomes, adverse events, and risk management.

MDCG 2020-7 reinforces this. The guidance states that PMCF must generate clinical data that feeds back into the clinical evaluation and the risk management file. The data must be capable of detecting safety signals, performance degradation, or shifts in the benefit-risk profile.

Surveys do not typically generate this type of data.

Why Surveys Fail as Standalone PMCF

Surveys have a role. They can capture user feedback, device handling issues, or subjective outcomes like pain scores. But they rarely meet the core PMCF objectives when used alone.

First, surveys rely on self-reported data. Patients may not recognize adverse events. Clinicians may not report incidents through a survey that they would report through other channels. The data is filtered through the respondent’s perception, which is often incomplete.

Second, surveys lack clinical depth. A questionnaire asking a surgeon if they are satisfied with a device does not tell you whether the device caused tissue damage, whether it failed during a procedure, or whether the patient experienced a delayed complication six months later.

Third, surveys do not integrate with clinical outcome measures. If your device claims to improve healing time, you need objective measurements of healing, not a question asking the patient if they feel better.

And finally, surveys have selection bias. The patients who respond are not representative of the entire user population. The ones who experienced serious adverse events may not respond at all. The data you collect does not reflect the true safety profile.

What Reviewers Look For in a PMCF Study

When I review a PMCF plan, I look for three things.

First, does the study design address the residual risks identified in the risk management file? If your device has a risk of thrombosis, your PMCF must include a method to detect thrombotic events. If it has a risk of device migration, you need imaging follow-up, not a questionnaire.

Second, does the study generate objective clinical data? This means measurable outcomes. Healing rates. Complication rates. Device survival rates. Reintervention rates. Adverse event data linked to specific device use. This data must be verifiable and traceable.

Third, does the study have a mechanism to detect safety signals? Passive data collection is not enough. You need active surveillance. This could be a prospective registry with structured follow-up. It could be systematic review of electronic health records. It could be collaboration with a clinical site that tracks outcomes longitudinally.

If your PMCF plan does not address all three of these, it will be challenged during review.

When Are Surveys Acceptable?

Surveys are acceptable when they are part of a broader PMCF strategy, not the only component.

For example, a survey can supplement a registry. You collect clinical outcome data through structured follow-up, and you add a patient-reported outcome measure via survey. Together, they provide a complete picture.

Surveys are also acceptable for usability or human factors objectives, especially for home-use devices. You want to know if patients can operate the device correctly, if they understand the instructions, if they encounter practical issues. A survey can capture this.

But the survey cannot be your entire PMCF activity. It cannot be the sole source of evidence that confirms safety and performance.

The question is not whether surveys are useful. The question is whether they generate the evidence the MDR requires. Most of the time, the answer is no.

What Should a Robust PMCF Study Include?

A robust PMCF study design starts with the clinical evaluation and the risk management file. You identify the residual risks. You identify the clinical claims that need ongoing confirmation. You identify the data gaps that could not be closed pre-market.

Then you design a study that generates objective data to address those gaps.

For most devices, this means a prospective observational study or a registry. You follow patients over time. You collect structured data at defined intervals. You track adverse events systematically. You measure clinical outcomes using validated tools.

If your device is used in a clinical setting, you collaborate with sites that have structured follow-up protocols. If your device is used at home, you implement remote monitoring or periodic clinical assessments.

You also implement a complaint handling system that feeds directly into your PMCF database. Complaints are not separate from PMCF. They are part of the post-market surveillance ecosystem. If a complaint reveals a safety signal, it must trigger an investigation and be reflected in your clinical evaluation update.

The PMCF study must be designed so that the data it generates can be analyzed and interpreted in the context of your clinical evaluation. It must be capable of detecting changes in the benefit-risk profile. It must be capable of identifying trends that were not visible in pre-market studies.

The Role of Real-World Data

Real-world data is often mentioned in PMCF discussions. It sounds promising. Electronic health records, insurance claims data, patient registries operated by third parties.

But real-world data is only valuable if it is structured, accessible, and relevant to your device.

I have seen manufacturers plan to use hospital databases for PMCF. The problem is that the data is not device-specific. You cannot isolate outcomes related to your device from outcomes related to the procedure, the patient’s condition, or other devices used concurrently.

Real-world data works when you have a clear data extraction protocol, when you can link device use to patient outcomes, and when you have agreements in place to access and analyze the data. Otherwise, it becomes a vague intention that cannot be executed.

If you plan to use real-world data, you must demonstrate in your PMCF plan exactly how you will extract, validate, and analyze it. You must show that it is feasible. Notified Bodies will challenge any plan that relies on data sources you do not control or cannot access.

The Consequence of Weak PMCF

A weak PMCF plan does not just delay certification. It undermines the entire clinical evaluation.

The clinical evaluation must be updated with PMCF data. If your PMCF generates no meaningful data, you have nothing to update. Your clinical evaluation becomes static. It cannot respond to emerging risks or performance issues.

This creates a compliance gap. The MDR requires ongoing demonstration of safety and performance. If your PMCF does not generate evidence, you cannot demonstrate compliance.

This also creates a liability risk. If an incident occurs post-market and you have no PMCF data to show that you were monitoring safety, regulators will question whether you exercised due diligence.

The PMCF is not optional. It is not a formality. It is a core requirement that supports the validity of your clinical evaluation and the ongoing compliance of your device.

How to Build a Defensible PMCF Plan

Start with the clinical evaluation report. Identify the residual risks. Identify the clinical claims. Identify the assumptions that need confirmation.

Then ask: What data would prove or disprove these assumptions? What outcomes would indicate a safety problem? What metrics would confirm ongoing performance?

Design your PMCF study to collect that data. Use structured follow-up. Use objective measures. Use validated outcome tools where they exist.

Integrate your PMCF with your post-market surveillance system. Make sure adverse events, complaints, and vigilance reports feed into your PMCF database.

Document the rationale for your study design. Explain why your chosen methods are appropriate for your device. Show that the data you collect can answer the questions your clinical evaluation needs answered.

And finally, plan for data analysis and reporting. PMCF is not about collecting data. It is about analyzing data and using it to update your clinical evaluation and risk management. If you have no plan for analysis, you do not have a functional PMCF.

Final Thought

Surveys are easy. They are comfortable. They feel like progress. But they rarely generate the clinical evidence the MDR requires.

A real PMCF study is harder to design and harder to execute. It requires collaboration with clinical sites. It requires structured data collection. It requires ongoing commitment.

But it is the only way to demonstrate that your device remains safe and effective under real-world conditions. It is the only way to keep your clinical evaluation valid. And it is the only way to avoid major objections during review.

If your PMCF plan is built on surveys, it is time to reconsider what you are actually measuring and whether it answers the questions the regulation asks.

Peace,
Hatem
Clinical Evaluation Expert for Medical Devices
Follow me for more insights and practical advice.