Medical device development—Class I, II, and III

Building a medical device is an engineering problem. The regulatory pathway—510(k), De Novo, or PMA—determines the burden of evidence, but it does not change what engineers must actually do: define the device, understand its risks, make design decisions that are defensible under scrutiny, and capture those decisions in a form that regulators can evaluate. The submission is not the product. The device is the product, and the submission is a structured argument that the engineering was done right.

Most teams experience the regulatory process as two distinct phases: engineering, and then documentation. The documentation phase is where projects slow down, where institutional memory fails, and where engineers who built the device spend weeks reconstructing decisions they made months earlier. MANKAIND collapses those two phases into one continuous engineering workflow.

From design inputs to design outputs—without the translation gap

Design controls exist because the FDA recognized that poorly defined requirements are the upstream cause of most device failures. 21 CFR 820.30 mandates a structured development process not to create paperwork, but because structured development produces better devices. The challenge is that most teams treat design controls as a documentation obligation rather than an engineering discipline.

MANKAIND is built around the premise that design controls are engineering tools. When a team defines design inputs—intended use, user needs, performance specifications—MANKAIND uses that structured information to surface relevant standards, flag gaps in the risk analysis, and begin generating the traceability matrix that will eventually anchor the design history file. The team is not filling out forms. They are making engineering decisions, and the platform captures those decisions in a structured format that becomes the DHF.

For Class I devices, that traceability is lighter—but it still matters. Exempt devices can lose their exemption when modified, and engineering teams that lack clear records of original design intent are unable to assess whether a change crosses the line. MANKAIND maintains that record continuously, not as an end-of-project exercise.

Class II: the 510(k) argument is an engineering argument

A 510(k) is a substantial equivalence claim. The engineering team must demonstrate that their device has the same intended use and the same technological characteristics as a predicate—or, if the technological characteristics differ, that those differences do not raise new safety or effectiveness questions. That is an engineering analysis, not a legal one.

Selecting a predicate is a strategic decision that shapes the entire testing program. Accepting performance differences that require additional testing is an engineering trade-off. Determining which consensus standards apply—and whether to claim conformance or explain deviation—is an engineering judgment. MANKAIND supports those decisions directly. The platform understands the predicate landscape, maps applicable standards to device categories, and helps the engineering team construct the substantial equivalence argument before the testing program begins—not after it concludes.

This matters because 510(k) deficiency letters are overwhelmingly caused by testing programs that were designed without a clear understanding of what the submission needed to demonstrate. A performance test that uses the wrong acceptance criteria, or a biocompatibility evaluation that missed a contact duration threshold, sends the team back to the lab. MANKAIND helps engineering teams design the right experiments the first time.

Class III: PMA and De Novo demand deeper engineering evidence

PMA devices carry the highest risk classification and demand the most rigorous engineering evidence. Valid scientific evidence of safety and effectiveness is not a documentation standard—it is an engineering standard. The clinical investigation design, the statistical powering of bench performance studies, the failure mode analysis for high-risk use scenarios: these are engineering decisions that determine whether the PMA succeeds or fails.

De Novo is the pathway for novel, low-to-moderate risk devices without a predicate. It requires the engineering team to propose the regulatory framework—the special controls—that will govern their device class going forward. That is a sophisticated engineering and regulatory strategy exercise that MANKAIND supports by helping teams understand what special controls have been established for analogous device types and what performance criteria those controls typically require.

For PMA programs, MANKAIND maintains the connection between the pre-submission meeting strategy, the IDE study protocol, the non-clinical testing program, and the eventual module submissions. When the FDA asks a question in a major deficiency letter, the team can trace the answer back to the original engineering decision that generated the data—not spend three days searching email threads and shared drives.

The design history file as a living engineering record

The DHF is not a deliverable. It is a continuous record of engineering decisions. Teams that treat it as a deliverable produce DHFs that are incomplete, internally inconsistent, and disconnected from the actual development history. Teams that treat it as a living record—updated as design decisions are made, not reconstructed after the fact—produce DHFs that are both regulatorily adequate and genuinely useful for future development.

MANKAIND generates the DHF as a natural output of the engineering workflow. Design reviews, risk analysis updates, verification and validation records, and design change assessments are captured in a structured format that produces the submission-ready documentation without a separate documentation phase. The senior engineer who made the critical design decision six months ago does not need to reconstruct their reasoning. The platform recorded it at the time.

That is what engineering intelligence means in practice: a platform that makes the work of engineering visible, traceable, and defensible—across the full device classification spectrum, from Class I to PMA.