Planning a Clinical Investigation? Don’t Underestimate the MHRA Notice of No Objection
We have seen a growing number of companies come to Hardian after submitting a clinical investigation application (Notice of No Objection, or NoNo as we affectionately call it at Hardian) to the MHRA and receiving significant questions, an objection, or being faced with the prospect of resubmitting.
Often, the underlying technology is good. The clinical study may be well thought through. The problem is that the documentation supporting the investigational device simply isn't at the level the MHRA expects.
This can come as a surprise, particularly to early-stage medical device companies. A clinical investigation can still feel like part of the development process: you are collecting evidence, learning about the product and perhaps several steps away from UKCA or CE marking. It is therefore easy to assume that the regulatory documentation can also be relatively early-stage.
Unfortunately, that isn't how an MHRA clinical investigation assessment works.
As we explained in our previous blog , ‘ Do I need to notify the MHRA about my clinical investigation?‘ by Dr Felicity Lock, the purpose of the MHRA assessment is not simply to decide whether your study protocol looks sensible. The MHRA needs sufficient evidence to determine whether an investigational device is safe enough to be used with real-world users.
That means your device documentation and underlying processes need to be considerably more mature than many manufacturers expect.
A NoNO isn't just another research approval
For a clinical investigation in Great Britain that requires notification, the manufacturer must notify the MHRA at least 60 days before the investigation begins. Importantly, that period runs from the first day after the MHRA has accepted a valid application, not simply from the date on which you first send something through IRAS.
A research team may have spent months developing a Clinical Investigation Plan, participant information sheets, consent forms and ethics documentation. All of that is important, but it does not replace the technical evidence the MHRA needs in order to assess the device itself.
The MHRA's current submission checklist includes, amongst other things:
the Clinical Investigation Plan and Investigator's Brochure;
device details;
the Essential Requirements or General Safety and Performance Requirements checklist;
risk analysis;
Instructions for Use and device labelling;
summaries and reports from bench and pre-clinical testing;
details of previous clinical experience;
the standards applied; and
for Software as a Medical Device, the software lifecycle documentation needed to demonstrate that the software has been developed, verified, controlled and released appropriately.
For a Great Britain-only investigation, the MHRA states that the manufacturer should already have the information necessary to demonstrate compliance with all the applicable Essential Requirements, rather than only the requirements that are specifically the subject of the clinical investigation.
That is a fairly high bar for something that might still be thought of internally as an "early study".
At Hardian, we often describe the documentation needed for a NoNO as being around 70–80% of the eventual Medical Device File. This isn't a regulatory percentage or an MHRA rule, and it will vary significantly depending on the device and investigation. It is simply a useful rule of thumb from the submissions we work on.
The point is that a NoNO application is rarely a handful of documents created specifically for a clinical study. Much of the underlying regulatory architecture of the device already needs to exist.
Software manufacturers can be particularly caught out
This is something we see frequently with Software as a Medical Device (SaMD).
Software teams are often very good at documenting development, but that documentation has not necessarily been created within a medical-device lifecycle. A product team might have roadmaps, tickets, technical specifications and test results spread across several development tools and assume that these collectively constitute a Software Development Plan.
From a regulatory perspective, that is not necessarily the same thing.
The MHRA's updated guidance is quite explicit here. For medical devices containing software, software should, be designed and maintained in accordance with the principles of IEC 62304, with any deviations strongly justified.
The MHRA lists, as a minimum:
a Software Development Plan;
a Risk Management Plan and Report, including software hazard analysis;
a Software Configuration Management Plan;
Software System Requirements Specifications;
a Software System Verification Plan and Report;
a documented software problem-resolution process; and
evidence that the completeness of the software release has been reviewed.
So, if the submission asks for a Software Development Plan, writing a retrospective document describing roughly how the developers built the product is unlikely to be a sufficiently robust approach.
The plan needs to make sense in the context of the software lifecycle. Requirements, risks, development activities, verification and release need to fit together.
The same applies to risk management. The MHRA specifically asks for a risk analysis and states a preference for this to be conducted to the international risk management standard EN ISO 14971:2019. It also expects the Essential Requirements or GSPR checklist to explain how applicable requirements have been addressed and to reference the evidence used to demonstrate this.
This is where submissions can start to unravel. It is usually not the absence of a document with the correct title that causes the biggest problem. It is inconsistency between the documents.
For example, a safety-related requirement appears in the software requirements but cannot be traced back to a risk control. A risk control is listed in the risk analysis but there is no verification evidence demonstrating that it works. The intended use described in the Clinical Investigation Plan differs subtly from the intended use used to develop the risk file. Or, an Essential Requirement is marked as satisfied, but the evidence referenced does not actually demonstrate it.
Individually, these can look like small documentation issues. Collectively, they make it difficult for an assessor to follow the safety argument for the device.
"We'll sort the regulatory documentation out after the study"
This is probably the most expensive misconception. The logic is understandable. Complete the clinical study first, gather the evidence, and then build the full regulatory documentation when you are closer to market. But for many investigational devices, particularly SaMD, the MHRA assessment requires you to bring a substantial amount of that work forward.
And there is a good reason for that. Clinical evidence can tell you whether a device performs as intended in the investigation, but it shouldn't be the first time fundamental questions about device safety, software development, foreseeable risks or verification are properly addressed.
There is also a commercial reason to get it right.
The current MHRA fee for an initial clinical investigation notification for Class I, IIa and non-implantable/non-long-term-invasive Class IIb devices is £15,309. For higher-risk devices, including Class III devices, the initial submission fee is £32,016.
A resubmission is also chargeable: currently £11,701 for the former group and £22,678 for the higher-risk group.
There are fee waivers and payment easements available in certain circumstances, including provisions aimed at some smaller companies, so manufacturers should always check the current MHRA arrangements rather than assume the full fee applies to them.
But the wider cost of a poor submission is rarely just the MHRA fee. If the investigation is linked to grant milestones, recruitment windows, NHS sites, investor commitments or a planned regulatory submission, losing several months to regulatory rework can have a much greater impact.
And because the statutory assessment period begins only once an application has been validated, assuming that "we submitted 60 days before our study date" can create a nasty surprise if the application isn't ready to enter assessment.
The good news: this work shouldn't be disposable
There is a more positive way to look at all of this.
If the device you are taking into the clinical investigation is the device you ultimately intend to place on the market, a well-prepared NoNO submission can accelerate the regulatory work that follows.
Your risk management documentation doesn't suddenly become irrelevant when the study finishes. Neither do your software lifecycle documents, requirements, verification evidence, device description or standards assessment.
They become part of the evidence base that you continue to develop.
Of course, they will evolve. Clinical investigation results may identify new hazards or change the assessment of existing risks. Software may change following study feedback. Requirements may be updated, additional verification may be required and the final intended purpose may need refinement. But that is very different from starting again.
This is why we encourage clients to approach a NoNO as part of their regulatory development pathway, rather than as an isolated permission needed to get a study started.
If you build the documentation properly in the first place, you are not producing paperwork solely for the MHRA assessor. You are building the foundations of the evidence package that will ultimately support the device.
Where do NoNO submissions commonly go wrong?
There isn't one universal failure point, but the same themes come up surprisingly often in the submissions we review.
Sometimes the company has concentrated heavily on the clinical documents and underestimated how much technical documentation the MHRA will assess.
Sometimes the right documents technically exist, but they have been created independently and don't agree with one another.
Sometimes software documentation reflects normal commercial software development rather than IEC 62304.
Sometimes risk management has been treated as a standalone risk register rather than something that should connect hazards, risk controls, product requirements and verification.
And quite often, a company has tried to create the regulatory documents retrospectively immediately before submission.
Getting it right first time
None of this means that an investigational device needs to be a completely finished, market-ready product before it enters a clinical investigation. The entire point of the investigation may be to generate clinical evidence that does not yet exist.
But there is a difference between clinical evidence that still needs to be generated and fundamental device development and safety documentation that hasn't yet been done.
The MHRA's current guidance makes that distinction increasingly clear.
Before submitting, manufacturers should understand exactly what the MHRA expects for their type of device, identify which standards are relevant, and review the complete submission as one coherent evidence package rather than as a checklist of unrelated documents.
This is also where using a regulatory team that regularly works with MHRA clinical investigation submissions can be valuable. Often the most useful work happens before anything reaches IRAS: identifying gaps, making sure the intended use and clinical strategy make sense, getting the technical documentation into the right structure, and challenging inconsistencies before an MHRA assessor does.
At Hardian, our regulatory and clinical teams work together on NoNO submissions, from determining whether notification is required and designing the clinical investigation, through to reviewing or developing the technical documentation needed to support the investigational device.
We have also been increasingly called in to support companies after an unsuccessful or heavily queried submission. It can normally be recovered, but fixing a submission after the event is almost always more challenging than getting the foundations right from the start.
If you're planning a clinical investigation, it helps to think of the NoNO as more than a regulatory hurdle before the study can begin. Done properly, it is an opportunity to test the strength of your regulatory documentation early and build evidence that can continue to support the device as it moves towards market.
References and further reading
MHRA – Submitting a clinical investigation proposal for MHRA assessment. The key MHRA guidance for preparing a clinical investigation submission, including the supporting documentation expected, Essential Requirements/GSPR checklist, risk analysis, software documentation and applicable standards. Updated 6 August 2026. Read the MHRA guidance
MHRA – Clinical investigations in Great Britain. Guidance on when clinical investigations are required, notifying the MHRA, conducting an investigation in accordance with the UK Medical Devices Regulations and communicating with the MHRA during and after an investigation. Updated 23 July 2026.Read the MHRA guidance
MHRA – Determining if a clinical investigation is required. MHRA guidance and decision-making information for determining whether a proposed investigation needs to be notified under the UK Medical Devices Regulations 2002. Updated 23 July 2026. Read the MHRA guidance
MHRA – Current MHRA fees. Current fees for medical device clinical investigation notifications and resubmissions, including the different fee levels according to device classification. Updated 21 April 2026. View the current MHRA fees
Hardian Health – Do I need to notify the MHRA about my clinical investigation? Our previous overview of the MHRA Notice of No Objection process for SaMD, including when notification may be required, the application pathway and the documentation manufacturers should expect to prepare. Read the Hardian article
