How to Handle Data Strategy Complexity During Software as a Medical Device (SaMD) Development
The medical industry has started to recognize the potential of medical software that makes use of machine learning for a variety of applications. As with all medical devices, development can proceed much more smoothly and effectively when regulatory concerns are addressed early in the process. In order to comply with medical device regulations and deliver a reliable product within timeline and budget constraints, medical device companies need to pay attention to their quality management systems before starting development on AI-based medical software. Regulations such as IEC 623O4 and IEC 623O4–1 dictate that companies use data from large-scale studies and clinical practice to substantiate any supplemental processes.
Supporting Claims for AI-based Software Devices
Accuracy and reliability are essential features in any medical device, especially ones that make use of new and unproven approaches like machine learning. Any claims about an AI-based software platform’s efficacy or capabilities must be examined by clinical and regulatory communities and also backed by well-designed research studies and real-world clinical evidence. Any 510K device can become a PMA unless the product development team says otherwise. Because of this, it is essential to conduct pre-submission procedures to confirm any claims that are necessary for general market acceptance.
Regulators' Point of View about Hardware and Software
There is essentially no regulatory difference between software or hardware-based medical devices. The same stringent QA and testing methods need to be used regardless. Baseline minimum requirements stated by IEC 62304 not only outline performance and reliability standards for these devices, but also suggest that companies create and implement a regulatory development plan, a clinical development plan including rigorous statistical analysis, and an endpoint software development plan.
Work With Regulators to Avoid Untimely Intervention
It’s important to involve regulatory officials early in the product development process, including during marketing reimbursement assessments. This will help device companies avoid errors in software and electronic devices. Because regulatory intervention can happen at any point in a product’s development cycle, it’s essential to clearly outline and adhere to the product development plan’s inputs, outputs, and risks. Discussing regulatory pathway alternatives as they relate to clinical pathways is key to developing informed judgements about a device’s quality.
All Devices Face the Same Scrutiny
While pre-submissions on devices of this type don’t need to cover the specifics of software or clinical protocols, this will be necessary at later points in the development process. Because of this, it’s important to find a trustworthy partner for software requirements including development tools, claims data, and monitoring technologies for the final release of the product. The FDA generally doesn’t differentiate by risk factor on medical software; high, medium, and low-risk devices are all subject to the same scrutiny, meaning that all clinical data must be justified, and the manufacturing process is governed by the same way for software and hardware products.
Identify Secondary Endpoints during Pre-submission
It is important to create clinical development plans in tandem with the regulatory roadmap in order to ensure compliance, but it’s also worth remembering that the regulatory procedure for 510(K) De Novo or PMA may change if the clinical research doesn’t involve the actual hardware or software device being developed. Animal testing is almost never required for devices of this type. During the pre-submission process, it may be discovered that secondary endpoints, rather than primary endpoints, will enable teams to build a stronger argument for the offering.
Start the Conversation Early
There’s no problem with starting a conversation with the FDA as long as you know what you want to push for in pre-sub and what you’re willing to give up. The first Q sub is essentially a handshake with the FDA, after which it’s possible to use retrospective data in real-world evidence to support clinical validation.
Clinical Validation is Essential
If you intend to use pre-sub to push the envelope in terms of software validation, thorough clinical validation is essential. The FDA generally anticipates 200 to 300 patients in a non-substantial-risk trial across a variety of settings and users. If the plan is to market your software device with third-party hardware, thorough clinical validation of the software is required. Some cameras may be exempt from class 1 510(K), but as soon as software enters the picture, clinical validation is required.
Essenvia is a regulatory lifecycle management tool designed to help you optimize the entire regulatory operations process. Our solutions include an innovative approach to building your pre-market submissions like 510(k) and MDR as well as post-market submissions like Regulatory Change Assessment and Post Market Surveillance. Essenvia is designed to reduce errors, streamline processes, and guarantee faster regulatory clearance for your medical device.
Schedule a demo now to find out how Essenvia can help you with your regulatory submissions.