Assessment of changes’ significance in accordance with MDR Article 120(3)

In line with agreed arrangements for notification of changes between the manufacturer and the notified body according to the AIMDD/MDD (e.g., contractual relationships, approved procedures) changes and their implementation will be verified by the notified body as part of the surveillance activities or following a manufacturer’s submission for prior approval.

The outcome of this verification will determine whether a certificate in accordance with AIMDD/MDD remains valid according to Article 120 MDR. To use this derogation from Article 5 MDR manufacturers are not allowed to make significant changes in design or a significant change in the intended purpose. In case of doubt whether a change is significant they should ask their notified body (MHRA).

Directives to the new EU Regulations

Guidance for manufacturers and Notified Bodies on reporting of Design Changes and Changes of the Quality System (section 5.1)

Criteria for “substantial” changes

The manufacturer must report to the MHRA any plan of “substantial” or “significant” changes before implementation.

Product changes[1] should be considered substantial if the change may affect:

• The conformity with the essential requirements and/or
• The indications and/or contraindications and/or warnings determined by the manufacturer to be appropriate to ensure the clinical performance of the device.

When determining whether or not a particular product change is “substantial” following considerations should be made (the list is not exhaustive):

• Changes of the intended purpose and/or the performance specification of the device?
• Are new hazards introduced which have not previously been addressed?
• Are risks associated with existing hazards affected?
• Does the change alter the details on intended use given in the design/type approval dossier submitted to the MHRA?
• Does the change trigger a need to alter the indications or contraindications for use or warnings necessary to ensure safety and efficacy for the intended use of the device?
• Does the change mean that the device will have different end users or be used in a different manner?
• Does the change mean that the clinical data/performance evaluation data for the original device is not sufficient to assure conformity of the changed device with the required characteristics and performance?
• Is the change a direct result of actions taken related to concerns arising from Post Market Surveillance including incidents/recalls/complaints?

Changes to design specifications

Changes to the design specifications, may be substantial if they affect the indications for use or the performances of the device or raise new safety and performance issues.

If the response to any of the following questions is yes, then it is likely that the design change is substantial

1. Does the design change affect the indications or contraindications for use or warnings necessary to ensure safety and performance for the intended use of the device?
2. Are further clinical data necessary to support the safety and performance of the altered device?

In cases where the change consists only of tightening of design specifications within specified tolerances and where there is no creation of new features, the change is not considered to be substantial.

Changes to software

Many changes to a device’s software will require an evaluation and acceptance by the MHRA.

The following would be considered substantial changes:

• a software change, which impacts the control of the device, that may alter the diagnosis or therapy delivered to the patient;
• an alteration in software that modifies an algorithm impacting the diagnosis or the therapy delivered;
• a software change that impacts the way data is read or interpreted by the user, such that the treatment or diagnosis of the patient may be altered when compared to the previous version of the software;
• a software change that replaces previously required user input to a closed loop decision;
• addition of a new feature to the software that may change the diagnosis, or the therapy delivered to the patient;
• a software change that incorporates a significant change to the operating system on which the software runs. If the software is modified to correct an error (for example, a change in algorithm), for which there is a safety risk to the patient if the error is not corrected, this software change may require an evaluation and approval by the MHRA.

HRA

1. Where the AI app algorithm is being used in operation in a service,as a service evaluation, but has changed sensitivity and specificity, we would advise to re-examine whether those changes mean that, in effect, the nature of the activity has now become research. In other words, the reasons behind those changes or the effects, might herald a different type of project that is more akin to research.

See this useful table to help with that determination Defining Research Table Oct 2017 (hra-decisiontools.org.uk) – see also Is my study research? (hra-decisiontools.org.uk).

The kinds of things to consider are whether the changes indicate that the purpose of the review has now expanded into research e.g., because generalisable findings are being sought from the analysis of the data post-changes (i.e. falling outside of the original intended use of the AI app algorithm).

In particular:

• why a new version and any revisions that change the algorithm intermittently are being carried out?
• does that purpose indicate that the review of the findings goes beyond pure service evaluation?
• Is any new type of analysis being performed on the data being generated beyond a service evaluation designed to define/judge current care?

2. Where the AI app algorithm is being used in primary care outside its original intended use, also changing the sensitivity, again I would want to know whether the reasons for the change.

If the reason goes beyond the clinical into e.g., generating a hypothesis, that would indicate that the changes are straying into research activity that should be notified on IRAS where NHS data are being used (and require REC if that data is identifiable of the patients). Again, this will require a case-by-case assessment of the facts.

In either case, if the facts now suggest that research is being carried out, that could trigger a need to submit on IRAS. This HRA toolkit, designed for students, but useable by all as a simple navigable PowerPoint to download, is a great roadmap to use to determine when and what type of approvals may be required from an HRA perspective: Student research toolkit – Health Research Authority (hra.nhs.uk).

The Medicines and Healthcare products Regulatory Agency (MHRA) is the designating and competent authority in the UK.

An approved body is an organisation that has been designated by the MHRA to assess whether manufacturers and their medical devices meet the requirements set out in the Medical Devices Regulations 2002 (SI 2002 No 618, as amended) (UK MDR 2002).

Manufacturers can apply to any UK approved body and once they have the necessary certification their products can be sold anywhere in Great Britain (England, Wales and Scotland). Following an appropriate assessment, the approved body will issue relevant certification allowing manufacturers to place UKCA marking on their products and place them on the market in Great Britain.