The useful planning question is which unfinished dependency controls your next milestone. If a clinical evidence gap needs new work, a faster document review will not remove it. If your evidence is ready but the assessment slot remains unconfirmed, adding more writers will not bring that slot forward.
This guide covers MDR certification involving a Notified Body, including first certification and planning for an eligible legacy device. It does not estimate IVDR certification, routine surveillance or the time needed to establish transition eligibility.
What the Commission survey actually measures
The Commission-hosted Notified Body survey dated 2 July 2026 provides a useful benchmark, with data status at 28 February 2026. Its page 30 measures time from a signed written agreement to issuance of a new MDR certificate. Preparation before that agreement sits outside the measured interval.
The chart separates quality management system certification, or QMS, from QMS plus product certification.
| Reported duration | QMS alone | QMS plus product |
|---|---|---|
| Under 6 months | 2% | 0% |
| 6 to 12 months | 30% | 8% |
| 13 to 18 months | 62% | 51% |
| 19 to 24 months | 2% | 31% |
| Over 24 months | 4% | 10% |
These are percentages of reporting Notified Bodies that had issued the relevant certificates. They are not percentages of devices, applications or certificates. The QMS denominator is 47 Notified Bodies, with five excluded because they had not issued that certificate type. For QMS plus product, the denominator is 39, with thirteen excluded for the equivalent reason.
That distinction changes how you use the figures. The 51% in the QMS-plus-product column does not mean your device has a 51% chance of certification within 13 to 18 months. The chart also does not specify whether each body's reported duration represents a mean, median or typical case. It cannot establish a market-wide average or describe the prospects of an unresolved application.
Use the chart to challenge an unsupported assumption. If your plan allows only six months from agreement to combined certification, ask what device-specific and Notified Body-specific evidence supports that expectation. Do not replace the assumption with an automatic 18-month promise either. Your assessment route, evidence and agreed schedule still need their own plan.
Give every date a start and an end
A duration becomes useful when you can identify both events it measures. “Certification in twelve months” leaves too much unanswered if one person means twelve months from today and another means twelve months after an assessment-ready submission.
Split your forecast into distinct lanes. Some lanes can overlap, but each needs an owner and a completion condition.
| Planning lane | Start event to record | Completion evidence | What can hold it up |
|---|---|---|---|
| Evidence preparation | Agreed device scope and gap review | Approved evidence set covering the proposed scope | Missing tests, clinical evidence or unresolved claims |
| Application and contracting | Application lodged under the body's process | Signed agreement and confirmed assessment scope | Incomplete application, designation mismatch or unresolved terms |
| QMS assessment | Agreed audit activity begins | Body confirms the assessment has reached its required decision inputs | Missing operating records, audit findings or unavailable personnel |
| Technical documentation assessment | Agreed review begins | Body confirms the required technical assessment inputs are complete | Inconsistent documents, evidence gaps or questions |
| Manufacturer responses | A request or finding arrives | Submitted response and subsequent body determination | New evidence generation, linked revisions or specialist availability |
| Decision and issuance | Required assessment inputs reach decision review | Applicable certificate issued following a favourable decision | Unresolved issues or remaining decision requirements |
Treat these as planning lanes, not a universal description of every Notified Body's procedure. Ask your body which events it uses and what it needs before each activity can start. In particular, “submitted”, “received” and “accepted as complete” may describe different milestones.
Record the evidence behind each date. A signed agreement, an email confirming an audit slot and an internal target are not equally firm commitments. Showing the source next to the date makes the forecast easier to challenge without turning every meeting into an argument about who remembers what.
Build the plan around evidence dependencies
A document can be finished while the evidence supporting it remains incomplete. Before you forecast submission readiness, identify work that could change the device scope or invalidate another workstream.
For example, an unresolved performance claim may affect the clinical evaluation, risk management, verification work and instructions for use. Planning those documents as four independent writing tasks hides the shared decision beneath them. Resolve or explicitly schedule the claim decision, then identify which conclusions need to change together.
Separate work you can estimate internally from work that depends on another organisation. Your team may control the time needed to reconcile a risk file with a completed test report. It does not necessarily control laboratory availability, access to clinical data or a Notified Body review slot. Give those dependencies a confirmation date rather than quietly treating an estimate as a booking.
The practical output is a short register that connects the next milestone to its blocking evidence. Each row should identify the owner, predecessor, source of the estimate, overlap assumption and next confirmation. Keep the full task list elsewhere. This register should show what can change the forecast.
A worked plan that separates parallel work
The following example is illustrative. The device, durations and scheduling assumptions are invented to demonstrate the method, not to predict an assessment or describe a customer result.
Suppose you are preparing a first MDR application for a Class IIa device, with the classification and proposed assessment route already reviewed. Your gap assessment identifies an incomplete verification report, linked clinical-document revisions and QMS records that still need internal review. You plan in weeks from a defined project start rather than attaching a certificate date to an unconfirmed application.
| Workstream and owner | Illustrative duration | Predecessor and overlap | Evidence or confirmation needed |
|---|---|---|---|
| Complete verification evidence, engineering lead | 8 weeks | Starts at week 0 | Confirmed laboratory schedule and approved report |
| Complete linked technical and clinical revisions, RA and clinical leads | 4 weeks | Starts after the verification report, ends at week 12 | Conclusions reconciled across affected documents |
| Review QMS records and close internal gaps, QA lead | 10 weeks | Runs alongside evidence preparation | Approved readiness review, not just drafted procedures |
| Assemble application and resolve scope questions, RA lead | 6 weeks | Runs alongside preparation, subject to the body's process | Body confirms required application content and scope |
| Complete cross-document readiness review, RA and QA leads | 2 weeks | Starts after the slower evidence or QMS lane | Approved submission baseline, ends at week 14 |
| Confirm assessment scheduling, RA lead and body contact | Not estimated here | Depends on the body's requirements and availability | Written confirmation of what can start and when |
The internal readiness calculation is the longer of the twelve-week evidence lane and the ten-week QMS lane, followed by the two-week integrated review. That gives fourteen weeks. Adding all the rows would produce thirty weeks and incorrectly assume that every activity runs in sequence.
Fourteen weeks is still only an internal readiness assumption. It is not a commitment to sign an agreement then, start assessment then or receive a certificate fourteen weeks later. Application activity may begin earlier if the body permits it, and assessment scheduling may introduce dependencies not yet confirmed.
Now change one assumption. If the verification report arrives at week 12 rather than week 8, the four-week revision period ends at week 16 and the integrated review ends at week 18. The internal readiness forecast moves by four weeks. The effect on assessment depends on whether you can still meet the body's submission and scheduling requirements. Missing a confirmed slot could have a different impact from moving an unbooked target.
This is the point of the exercise. The plan identifies the report as a controlling dependency and tells you which external date to check when that report moves.
Keep responses inside the forecast
Your response work belongs in the calendar even when you cannot predict the findings. A request for a clearer reference and a request that exposes missing evidence can demand very different amounts of work.
For a reference problem, you may need to identify the supporting record, explain the traceability and check that the document set remains consistent. For an evidence problem, you may need new testing or a clinical reassessment before you can submit a meaningful response. Neither task becomes shorter because the forecast calls both of them “one review round”.
Plan who can triage a request, who can make the necessary technical decisions and who checks linked revisions before submission. Reserve access to those people rather than assuming they will be available during another project's launch or audit. Then distinguish your response submission date from the body's subsequent review and determination.
Do not allocate a universal number of rounds or assume that the end of your response period means closure. Use the body's actual requests, agreed response requirements and assessment process to update the plan. If new work threatens an agreed date, clarify the situation with the body while there is still time to make a documented plan.
Distinguish legal clocks from calendar time
An empirical survey, a contractual schedule and a legal assessment limit answer different questions. Keep their definitions separate rather than treating the shortest number as your expected certification time.
The original published text of Commission Implementing Regulation (EU) 2026/977 sets stage-specific timing provisions. Article 9 gives a general application date of 25 February 2027, while Article 8(1) excludes procedures with written agreements signed before that date from Articles 1 to 3. The cited official text is in Spanish.
Those provisions do not create an unconditional calendar deadline for certification. Article 2 uses distinct stage boundaries, and Article 3 addresses interruptions. Article 2(4) also distinguishes exhaustion of timing limits from the substantive certification decision. Reaching a time limit does not establish that your device conforms.
Before using a legal limit in your forecast, have your regulatory lead confirm the applicable text, agreement date, covered procedure and interruption rules. Your plan should still show elapsed calendar time and manufacturer work separately. A response period can consume calendar time even when an assessment clock does not run.
Use scenarios without inventing probabilities
A useful forecast shows what would change the date, not just a wide range with no explanation. Build a working scenario around the evidence and scheduling assumptions you currently support, then test specific disruptions.
In the illustrative example, the working scenario reaches internal readiness at week 14. An evidence-delay scenario reaches week 18 because the verification report arrives four weeks later. A scheduling scenario leaves readiness at week 14 but keeps the assessment start unconfirmed until the body provides a date. These scenarios describe different problems and require different action.
Do not assign them numerical probabilities unless you have a defensible basis. Likewise, avoid a blanket contingency that conceals known missing evidence. If you already know a test is outstanding, schedule that work directly rather than calling it contingency.
For a management update, carry three fields together. Show the next milestone, the assumption that controls it and the event that would trigger a new forecast. For this example, those fields could be “submission readiness at week 14”, “verification report available at week 8” and “laboratory confirms a later report date”. That is more useful than a green status marker beside an unsupported certification date.
What changes for a legacy device
A legacy device adds a separate question about the legal basis for continued market access. A certification forecast does not establish transition eligibility, and a projected assessment duration cannot extend a transition endpoint.
Keep the eligibility assessment and the certification plan connected but distinct. Your device-level regulatory review should establish whether the relevant transition provisions apply. Your project plan should then work backwards from the applicable endpoint, using confirmed milestones and explicit dependencies rather than subtracting a survey duration and calling the result a safe application date.
If the forecast leaves little room for evidence generation or scheduling uncertainty, escalate that conflict early. The decision may concern evidence scope, resources or commercial planning. It cannot rest on an assumption that a Notified Body will compress the assessment to fit your preferred date.
Questions to settle before you commit
Can you use the survey as a launch forecast?
Use it as context for the agreement-to-certificate interval, not as a forecast from today. Your launch plan also needs to account for work before agreement and applicable steps after certification. A certificate milestone and commercial availability should remain separate dates.
Can QMS and technical review run in parallel?
Some activities may overlap, but confirm that with the body and the relevant procedure. Parallel work only helps when both lanes have the evidence and people needed to progress. The slower required lane can still control the decision milestone.
When should you change the forecast?
Reforecast when a controlling assumption changes, including evidence availability, device scope, an agreed assessment slot or the work needed to answer a request. Keep the previous forecast and the reason for the change so you can distinguish new information from missed execution.
A credible timeline gives you decisions you can make now, dates you can substantiate and uncertainties you know how to resolve. That is the basis for a useful conversation about readiness, whether you are approaching a body for the first time or trying to understand an existing assessment plan.
