Class III — what actually changes when you cross from Class II into PMA territory

If you've built Class II devices and you're starting a Class III program, the instinct is usually to scale up what worked for Class II. That instinct is wrong in specific, predictable ways.

Class III is a different engineering discipline. PMA is a different regulatory framework. The document depth, the evidence standard, the FDA scrutiny, the internal quality system requirements all operate at a level that Class II practices can't sustain by just doing more of the same thing. Teams that treat it as an intensification of Class II spend the first 18 months of their PMA program accumulating documentation debt they then have to reconstruct.

What actually puts a device in Class III

FDA classifies based on the controls necessary to provide reasonable assurance of safety and effectiveness. Class I is general controls only — labelling, registration, listing. Class II adds special controls — performance standards, post-market surveillance, 510(k) clearance. Class III requires PMA because general and special controls don't provide adequate assurance of safety and effectiveness for the risk profile.

Devices land in Class III for three reasons. Explicit classification by FDA based on risk profile (implantable cardiac, mechanical heart valves, cochlear implants). Default classification for novel devices without a predicate (which is why De Novo exists — to reclassify genuinely novel low-to-moderate-risk devices out of the default Class III). High-risk diagnostic software making clinical decisions with direct patient consequence.

When Class III is the appropriate classification — not Class III by accident of novelty, but Class III by genuine risk — the PMA pathway is the evidence standard that corresponds to the consequence. It's not bureaucracy. It's the framework for demonstrating that a device with serious safety implications has been engineered to the depth the risk warrants.

The PMA pathway structure

PMA is an application submitted to FDA under 21 CFR Part 814. Required content includes a device description, alternative practices summary, marketing history, summary of studies, non-clinical performance data, clinical data, manufacturing information, labelling, environmental impact assessment, and bibliographic references.

The first formal FDA interaction is almost always a Pre-Submission (Q-Sub) meeting. Used properly, Q-Sub is the mechanism for aligning with FDA on the clinical study design (IDE-level questions), proposed endpoints, statistical analysis, technical questions — before they become PMA deficiencies. Skipping Q-Sub for a Class III program is a decision with consequences. I've seen sponsors try it. They generally regret it.

For devices requiring clinical investigation, the Investigational Device Exemption (IDE) is a separate FDA interaction — application, investigational plan, risk analysis, informed consent procedures, device manufacturing information. FDA has 30 days to review, may approve, approve with modifications, or disapprove. Clinical data collected under an IDE becomes the clinical section of the eventual PMA.

PMA review runs under a 180-day statutory clock. The clock doesn't run continuously. Major Deficiency letters stop it, and the sponsor's response restarts it. Total timeline — initial submission to approval — is typically 18 months to 3 years. Sometimes longer. Device complexity, clinical data strength, and the completeness of design control documentation all drive where in that range you land.

Design controls at the Class III level

21 CFR 820.30 applies to every Class II and Class III manufacturer, but the depth of implementation scales with device risk. At Class III, FDA reviewers read design control documentation carefully. This isn't rhetorical — they actually read it.

Design inputs get evaluated for completeness and traceability to user needs. Design outputs get evaluated for whether they're specified precisely enough to verify conformance. Verification testing gets evaluated for whether it actually addresses the acceptance criteria in the inputs. Validation gets evaluated for whether it confirmed the device meets user needs under actual use conditions with representative users.

Every gap in that chain is a potential deficiency. A design input that doesn't map to a user need raises questions about whether the design targeted the right problem. A verification test that doesn't reference the acceptance criteria it confirms raises questions about whether verification was disciplined. A design change late in development not evaluated for validation impact raises questions about whether the device being manufactured is the device that was validated.

Design reviews take on particular weight. Reviews have to include individuals independent of the stage being reviewed. For Class III, reviewers often include clinical advisors, risk management experts, manufacturing representatives. The review record — attendees, findings, action items, action item closure — is a DHF artefact that inspectors and reviewers actually examine.

Risk management at Class III depth

ISO 14971 applies to every medical device. At Class III, implementation depth reflects the consequence profile.

Hazard identification has to capture all reasonably foreseeable situations, not just the obvious ones. For implantables, that includes degraded-state operation, battery depletion scenarios, migration, infection, long-term material interactions, explant scenarios. For Class III SaMD, it includes model drift, adversarial inputs, data distribution shifts.

Risk estimation should be quantitative or semi-quantitative where clinical data supports it. Risk controls implemented in priority order — inherently safe design first, then protective measures, then information for safety. The benefit-risk analysis for Class III is central to the PMA clinical argument. Residual risk after all controls can still be acceptable if clinical benefit is sufficient, but the acceptance rationale has to be documented and defensible.

The risk management file is a live document for the device's lifetime. Post-market data feeds back, new hazards get identified, controls get reassessed. For Class III devices, post-approval studies often extend the risk file evolution for years after approval.

V&V depth — where the cost really lives

Verification and validation for Class III has to address not just performance to specification but whether the specifications were the right ones. That means validation under actual or simulated use conditions, with representative users or patient populations.

Simulated use testing includes usability engineering under IEC 62366 at summative-level rigor. For software-containing Class III devices, IEC 62304 Class C software lifecycle documentation — formal unit verification, coverage metrics, full integration and system testing, SOUP management. Biocompatibility per ISO 10993 for patient-contacting components. Electrical safety under IEC 60601 with applicable particular standards. All of this gets integrated into the PMA.

The traceability matrix for a PMA submission traces requirement to verification/validation activity, test results to requirements, design outputs to inputs, inputs to user needs. For complex devices with thousands of requirements, maintaining this traceability in spreadsheets becomes untenable. Teams that try to do it that way accumulate hours of reconciliation work per design iteration, and the reconciliation falls behind during crunch periods. Submission prep then has to absorb that accumulated debt, usually months of it.

What Class II habits don't scale

Informal design reviews work at Class II, generate 483s at Class III. Reviewer informality means records aren't contemporaneous enough for PMA scrutiny.

Spreadsheet-based requirements traceability works at Class II, breaks at Class III. Complexity exceeds what spreadsheets can maintain without drift.

Milestone-based risk file updates work at Class II, generate deficiencies at Class III. Risk file needs to be continuously current, not updated in batches before submissions.

Delayed design transfer activity works at Class II, creates pre-approval inspection findings at Class III. Manufacturing validation needs to be substantially complete before PMA submission.

Compiled DHF works at Class II — often. It doesn't work at Class III. The density of cross-references and the depth of audit scrutiny expose reconstruction in ways that a 510(k) review sometimes lets slide.

Where MANKAIND fits in a Class III program

The platform is built around continuous engineering record maintenance. Design inputs link to user needs and risk controls. Outputs link to inputs. Verification links to outputs. Validation links to user needs. Risk controls track across requirements and design elements. Every change surfaces downstream implications.

For a Class III program, the compounding effect is the main value. Every design change would normally require manual reconciliation across requirements, risk file, DHF, V&V records, and potentially regulatory submissions. In an integrated platform, the reconciliation is automatic because the relationships are structural. The documentation debt that usually accumulates over an 18-month PMA program doesn't accumulate the same way.

That's the difference between a PMA submitted on original timeline and a PMA submitted 9 months late after months of record reconstruction. The engineering work is often the same. The platform determines whether the documentation tracks with the work or whether it has to be rebuilt at the end.

Frequently asked questions about Class III medical device development

What is a Class III medical device?

Class III is FDA's highest device classification — reserved for devices that support or sustain human life, are of substantial importance in preventing impairment of human health, or present a potential unreasonable risk of illness or injury. Examples include implantable cardiac devices, mechanical heart valves, and high-risk diagnostic software.

What is the PMA pathway?

Premarket Approval (PMA) is the FDA pathway for Class III devices. It requires substantial evidence — typically from controlled clinical trials — that the device is safe and effective. A PMA application includes a full manufacturing section, a comprehensive design controls section, and a clinical section.

How long does a PMA review take?

The statutory review clock is 180 days, but practical timelines run 18 months to 3 years. FDA issues Major Deficiency letters that stop the clock until the manufacturer provides additional information. Device complexity, clinical data strength, and design controls documentation completeness all drive total duration.

How is Class III different from Class II?

Class III requires PMA with clinical evidence; Class II typically clears via 510(k) substantial equivalence to a predicate. Design controls are enforced at significantly greater depth for Class III, and FDA reviewers examine DHF completeness and risk management rigor with materially more scrutiny.