The inclusion of medical devices onto the Therapeutic Goods Administration's (TGA’s) Australian Register of Therapeutic Goods (ARTG) requires manufacturers to demonstrate that their device meets specific safety and performance characteristics and has been designed and manufactured in accordance with the Essential Principles.
One key part of any submission package for ARTG Inclusion is the demonstration of Clinical Evidence for the device. This Clinical Evidence is also something that the manufacturer needs to update periodically throughout the lifecycle of the device.
With the emergence of new software based Medical Devices, regulators around the world have developed specific guidance on what is expected when submitting clinical evidence for these devices, and how the information should be evaluated. It is recognised that the technology behind software based medical devices and AI-based platforms are different to traditional devices, requiring specific information to be included when demonstrating clinical evidence.
The International Medical Device Regulators Forum (IMDRF) is a group of global regulators that have developed harmonised expectations for medical device manufacturers and published several recommendations and guidelines covering all aspects of medical device development. These documents have been adopted into regulations by member countries including Australia, Brazil, Canada, China, EU, Japan, Russia and Singapore.
By ensuring they are in compliance with the recommendations provided by the IMDRF/SaMD WG/N41FINAL:2017 Software as a Medical Device (SaMD): Clinical Evaluation, a manufacturer will meet the regulatory expectations in those countries listed above.
While the technology driving a software based medical device may be different from traditional medical devices, the expectations on clinical safety and performance are the same. Clinical evidence is the overarching term used to include the collection of clinical data and the documents generated during the evaluation of that data.
SaMD Clinical Evaluation Landscape
IMDRF/SaMD WG/N41Final:2017 Software as a Medical Device(SaMD): Clinical Evaluation
Data on the performance or safety of a device is generated through its use in a given healthcare setting. This may be either a literature review or clinical investigation.
Clinical Investigation
Systematic investigation or study in one or more human subjects, undertaken to assess the safety and/or performance of a medical device.
Note: 'clinical trial' or ' clinical study' is synonymous with 'clinical investigation’ and these terms are often used interchangeably.
Data can be generated on the safety and performance of a device through the use of clinical studies. These studies must be conducted with the approval of a Human Research Ethics Committee, and be governed by a Clinical Trial Protocol. The aim of clinical studies is to generate the data on how the device performs under real world conditions, which is an input into the evaluation of that data during Clinical Evaluation.
Literature Review
Review of clinical data that has already been generated and published in peer-reviewed literature is a good source of clinical data.
A literature review is conducted by defining what journals and repositories will be searched, the terminology and search terms used, and the how the literature will be reviewed and analysed.
The output of this will be a Literature Report that summarises the entire search method, results and conclusion.
This report must be endorsed by a scientifically competent or medically competent clinical expert.
Once a sufficient amount of clinical data is gathered, it is evaluated and documented. Clinical Evaluation of a software based medical device includes three elements:
Clinical Evaluation Process
IMDRF/SaMD WG/N41Final:2017 Software as a Medical Device(SaMD): Clinical Evaluation
1. Valid Clinical Association
For a new device to show it is developed with scientific principles, and will function as intended, the manufacturer must show a clear and valid clinical association between the output of the device and the clinical claim it is making.
For example, if the SaMD is intended to process the signal received from an ECG and make a diagnosis decision of atrial fibrillation, the manufacturer will need to provide clear and documented evidence that the ECG readings used to program the software are indicative of atrial fibrillation as accepted in peer reviewed literature, medically accepted guidelines or other body of evidence.
2. Analytical Validation
For a software based medical device to show that the output provided is correct, the manufacturer must perform robust verification and validation testing during the product development phase. This will provide the documented evidence that the device can reliably take an input, process the information and provide the output that is accurate, repeatable and reproducible.
The data used for this evaluation will be generated by ensuring that the software is developed in line with IEC 62304:2006 Medical Device Software – Software life cycle processes.
3. Clinical Validation
In order to show that a software based medical device will function as intended, the manufacturer must demonstrate that the device will provide a clinically meaningful output in real world use.
Clinical validation should be able to demonstrate that the software will be able to function as intended (intended use) in a real world setting on the patient or end user. This can be through collation of data from previous non-clinical development studies, or generation of data from new clinical studies.
It is a key part of any clinical evaluation and should be developed in line with the risk classification of the device.
Ongoing commitments to post-market clinical studies, or collection of post-market data and ensuring that a lifecycle approach to software development has been applied will ensure that the commitment to clinical validation is met.
All of the information gathered in the above three phases will form the overall Clinical Evaluation. This is collated into one document that forms the Clinical Evaluation Report for the device.
While the gathering of information during product development to verify clinical association, analytical and clinical validation, the requirement to continually review this information does not stop with the publication of the Clinical Evaluation Report.
The Clinical Evaluation Report is a living document that includes regular review of post market data, consideration of new testing, and other updates throughout the product life cycle. This must be maintained and provided to regulators as requested.
In summary, by ensuring that software based medical devices have been developed in line with IEC 62304:2006, enough clinical data is gathered during its development, and a process is in place for its continued evaluation post market, most companies are well on the way to ensuring a robust package of clinical evidence to support regulatory submission and ultimately, approval.
Contact us to explore your interest or needs in Australian and international medical device regulations.