Registry Data for Implants: Why Most of It Doesn’t Count

Written by HATEM RABEH, MD, MSc Ing

Your Clinical Evaluation Expert And Partner

in
S

Registry data looks compelling on paper. Large patient populations. Long follow-up periods. Real-world conditions. For implantable devices, registries seem like the natural answer to demonstrating long-term safety and performance under MDR.

But here is what I see in clinical evaluation reports: manufacturers cite registry data without establishing whether that data meets the threshold for clinical evidence. They reference large datasets as if volume compensates for methodological weakness. It does not.

The question is not whether registry data can be used. The question is what makes it admissible under MDR requirements, and that standard is higher than most teams assume.

What MDR Actually Requires

Under MDR Article 61 and Annex XIV, clinical evidence must be sufficient to demonstrate conformity with the safety and performance requirements. For implantable devices classified as Class III or implantable Class IIb, this means demonstrating both clinical safety and clinical performance through data that is relevant, methodologically sound, and appraised.

MDCG 2020-6 reinforces this. Clinical data must be generated or collected under conditions that allow for proper assessment. Registry data can qualify, but only if it meets specific conditions.

Here is where most submissions falter. They treat registry participation as a checkbox. They reference registry outputs without examining the design of the registry itself, the quality of data capture, or whether the endpoints align with the intended purpose of their device.

Why Registry Data Gets Rejected

I have reviewed CERs where manufacturers cite national registries with tens of thousands of entries. The data is publicly available. The implant is listed. Follow-up exists. And yet, the data is dismissed during review.

The rejection is not arbitrary. It happens because the registry was not designed to answer the specific clinical questions relevant to the device. Most national registries track general outcomes: implant survival, revision rates, major complications. They do not track device-specific performance claims.

If your device claims improved osseointegration, lower subsidence rates, or reduced perioperative complications, the registry must capture those endpoints with sufficient granularity. If it does not, the data is not relevant, regardless of sample size.

Another issue: data completeness. Many registries allow voluntary reporting or partial data entry. Missing follow-up. Incomplete complication reporting. Inconsistent definitions of adverse events. When data quality is uncertain, the evidence is not reliable.

The Problem of Aggregation

Most registries aggregate data across multiple device models, sometimes across different manufacturers. This is useful for surveillance. It is less useful for demonstrating the performance of a specific device variant.

If your implant differs in material composition, surface treatment, or geometry from other devices in the registry, the aggregated data does not isolate the performance of your device. You are left with a population-level observation that may not apply.

Reviewers see this immediately. They ask: how do you know the outcomes apply to your device specifically? If the answer is unclear, the data is considered insufficient.

What Makes Registry Data Admissible

For registry data to qualify as clinical evidence under MDR, several conditions must be met. These are not optional. They are part of the evaluation standard described in MDCG 2020-8.

The Registry Must Have a Defined Protocol

The registry should operate under a documented protocol that specifies objectives, endpoints, data collection methods, and analysis plans. This is not always the case with national registries, which may operate under surveillance mandates rather than research protocols.

If the protocol exists, it must be appraised. You need to assess whether the design supports the conclusions you intend to draw. This means reviewing inclusion criteria, follow-up schedules, endpoint definitions, and data validation procedures.

If the registry protocol is not available or is too general, the data cannot be properly weighted in your clinical evaluation.

Data Quality Must Be Verifiable

MDCG 2020-8 emphasizes that clinical data must be collected under conditions that ensure accuracy and traceability. For registry data, this means understanding how data is entered, validated, and audited.

Some registries have rigorous data quality controls. Others rely on self-reported outcomes with minimal verification. If you cannot verify data quality, you cannot establish reliability.

This is a frequent gap. Manufacturers cite registry publications without access to the underlying dataset or quality assurance documentation. The Notified Body asks how data accuracy was ensured. The answer is often missing.

The Data Must Address Your Device Specifically

This is the most overlooked criterion. Registry data is admissible only if it isolates outcomes for your device or for devices sufficiently similar to yours.

If the registry aggregates multiple designs, you need subgroup analysis. If subgroup data is not available, you must justify why the aggregated results apply to your device. This requires demonstrating equivalence in design, materials, and intended use.

Most registries do not provide device-level granularity. This limits their utility for demonstrating conformity unless you can access raw data and perform your own stratified analysis.

Endpoints Must Align With Your Claims

The registry must capture endpoints that correspond to the safety and performance characteristics you claim. If your device claims reduced revision rates, the registry must track revisions with clear definitions and consistent follow-up.

If the registry tracks general survival but not the specific failure modes relevant to your device, the data does not support your claims. You can supplement it, but you cannot rely on it as standalone evidence.

How to Use Registry Data in a CER

When registry data is methodologically sound and relevant, it can be powerful evidence, particularly for long-term performance and rare complications. But it must be integrated carefully.

Appraise the Registry as a Data Source

Before citing registry data, conduct a formal appraisal of the registry itself. Document the protocol, data quality procedures, population characteristics, and endpoint definitions. Assess whether the registry design supports the inference you are making.

This appraisal should be documented in your CER. It demonstrates that you understand the limitations and strengths of the data source.

Clarify What the Data Does and Does Not Show

Registry data rarely provides complete answers. It may demonstrate long-term implant survival but not address early complications. It may show population trends but not device-specific performance.

Be explicit about what the registry data supports and where gaps remain. If the data does not address a specific claim, acknowledge that and identify supplementary evidence.

Combine Registry Data With Other Evidence

Registry data is strongest when combined with clinical investigations, literature, and PMCF. It provides real-world context. Clinical investigations provide controlled conditions. Literature provides comparative benchmarks. Together, they form a complete evidence base.

Do not treat registry data as sufficient on its own unless the registry was specifically designed to evaluate your device under conditions that meet MDR evidence standards.

When Registry Data Is Not Enough

Even when registry data is admissible, it may not be sufficient. For high-risk implantables, demonstrating conformity often requires direct clinical investigation, particularly when the device introduces novel materials, mechanisms, or claims.

Registry data is observational. It reflects what happens in practice, but it does not isolate causality or control for confounding variables. If your device claims a specific performance advantage, registry data alone may not establish that claim.

This is why PMCF must often go beyond registry participation. You may need device-specific follow-up studies, targeted surveillance, or post-market clinical investigations to address gaps that registries cannot fill.

Practical Implications

If you plan to rely on registry data, start by identifying which registries are relevant and whether they meet the criteria for admissible evidence. Request access to registry protocols and data quality documentation early. Do not assume that public registries automatically qualify.

If the registry does not provide device-specific data, consider whether you can negotiate access to raw data for subgroup analysis. If not, recognize the limitation and plan supplementary evidence collection.

In your CER, appraise the registry explicitly. Describe its design, strengths, and limitations. Show that you understand what the data supports and what it does not. This transparency strengthens the overall evaluation.

Registry data is valuable, but it is not a shortcut. It is admissible when it is relevant, methodologically sound, and properly appraised. Anything less, and it becomes a weak point in your clinical evaluation.

We will continue this series by examining the specific PMCF strategies that work for implantables when registry data is insufficient or unavailable.

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