Manufacturing transfer readiness: the 12 things most medical-device teams miss

21 CFR 820.30(h) on design transfer is one paragraph. The readiness work behind a successful transfer is the largest single integration effort in a regulated-product program — covering design output completeness, process validation under 820.75, supplier readiness under 820.50, and inspection posture across the design history file and device master record.

This is a working note on twelve specific gaps practitioner teams should close before transfer kicks off. The list is not exhaustive, but it covers the gaps most often surfaced — at the worst possible time — by an inspector, an auditor, or a contract manufacturer's first PQ batch.

What design transfer actually is

Design transfer is the activity of translating the device design into production specifications. The regulation (21 CFR 820.30(h)) does not specify how — it specifies the property: the design must be correctly translated. In practice, the burden of demonstrating correct translation falls on the design output package: drawings, specifications, software builds, process descriptions, work instructions, and inspection criteria.

The transfer is correct when manufacturing can build a unit that meets the design output specifications without consulting design engineering for clarification. Anything less is an incomplete transfer — and incomplete transfers are usually not detected at the moment they happen, but at the first deviation, the first PQ batch, or the first audit finding.

Gaps 1-4 — Design output package

1. Design output completeness against the BOM

Every BOM line has a controlled drawing or specification at a released revision. The most common version of this gap is engineering notation that worked at bench scale — a shared component number, an implicit material callout, a dimension that was held by hand-fit rather than specified — that does not survive the move to a contract manufacturer. Treat the design output package the way a supplier would: if a drawing does not include the dimensions, materials, finishes, and inspection criteria a supplier needs to quote, build, and inspect, it is incomplete.

2. Tolerance stack-up against process capability

Critical-to-quality dimensions have tolerances justified by capability data from the actual process or a representative one. The most common pattern: bench-built parts hold tolerances that the production process (injection molding, sheet-metal forming, automated dispense) cannot reliably hold. The fix is process capability data — Cpk targets at the drawing level for high-risk tolerances, validated against the supplier's actual process capability before transfer commits.

3. Inspection criteria locked before transfer

Each inspectable design output has an inspection method, sample size, and AQL or comparable acceptance criterion in the design output package. Negotiating inspection criteria with QC after transfer is a common pattern that produces inconsistent acceptance decisions and inspection records that do not trace to the design output. The acceptance criterion is part of the design output, not a downstream QC decision.

4. Design FMEA and process FMEA cross-reference

Each failure mode in the design FMEA maps to a process FMEA risk control or to a design feature that prevents the failure mode at the device level. Failure modes that have neither are gaps — either the FMEA is incomplete or the design and process do not yet jointly address the failure mode. Cross-referencing the two FMEAs surfaces the gap before transfer.

Gaps 5-8 — Process readiness

5. IQ/OQ/PQ scope for each non-verifiable process

21 CFR 820.75 requires validation of processes whose results cannot be fully verified by inspection or test. Common examples include sterilization, lyophilization, certain bonding processes, certain laser-welding operations, and most software-controlled processes where output verification is impractical. Each non-verifiable process needs a defined IQ/OQ/PQ scope, success criteria, and sample sizes — agreed before transfer kicks off, not negotiated during PQ execution.

6. Device master record (DMR) per 21 CFR 820.181

The DMR contains or references device specifications, production process specifications, quality assurance procedures, packaging and labeling specifications, and installation/maintenance/servicing specifications. Build the DMR incrementally during transfer, not retroactively after launch. A DMR assembled before launch reads as an inspection deliverable rather than a working production document.

7. Test method validation (TMV)

Test method validation confirms that each non-standard test method actually measures what it claims to. Skipping TMV is a frequent cost-and-time shortcut that produces verification data that cannot be defended at audit — "how do you know your test method produces accurate results?" is the standard auditor question, and a verification report without a TMV is the standard wrong answer.

8. Representative-build batch before formal transfer

A pre-PQ build of representative parts on production-intent processes catches the gap between bench and production before formal verification commits. Representative builds surface gaps (process settings that drift, fixtures that interfere with operators, cycle times that conflict with throughput targets) at the lowest cost — before they become PQ deviations.

Gaps 9-12 — Supplier and QMS readiness

9. Supplier qualification per 21 CFR 820.50

Each supplier of a critical component is qualified per the firm's purchasing controls procedure: capability assessment, quality agreement, audit cadence, approved supplier list entry. Critical components have at least one qualified backup or a documented sole-source justification. Sole-source on a critical component without a documented mitigation plan is an inspection finding waiting to happen.

10. Change control during the transfer phase

Transfer-phase changes are common — DFM findings, supplier substitutions, process tweaks. They need to route through 820.30(i) design changes and 820.40 document controls without re-opening the design output baseline more than necessary. The pattern that works: a tight change-control board active during transfer, with explicit criteria for which changes return to engineering for re-verification and which can be handled within manufacturing-engineering scope.

11. DHF built incrementally for the transfer phase

Per 21 CFR 820.30(j), the DHF must contain or reference the records necessary to demonstrate the design was developed in accordance with the approved design plan. The transfer phase generates a substantial portion of the design history — design output revisions, transfer activities, deviations, supplier qualification records, process validation records. Building the DHF incrementally during transfer is faster and cleaner than assembling it afterward.

12. Inspection-readiness rehearsal

Before any scheduled inspection, walk an experienced auditor through the DHF, DMR, supplier files, and process validation records. Identify gaps and close them before an actual investigator does. The rehearsal also surfaces narrative gaps — places where the records are correct but the program owner cannot explain them coherently — which are a different category of inspection risk and equally addressable.

Practitioner summary

Manufacturing transfer is the most expensive seam in a regulated-product program — the place where late discoveries cost the most because the design baseline is locked, the supplier is engaged, and the launch timeline is committed. The 12 gaps in this note cover the patterns most often surfaced too late: design output incompleteness, process capability mismatches, FMEA gaps, IQ/OQ/PQ scoping, DMR/DHF curation, and supplier readiness.

Programs that work through these gaps before transfer kicks off treat transfer as the activity it is — a controlled translation of a complete design package into a controlled production process — rather than as a discovery phase that surfaces what should have been resolved during design output.

For the manufacturing-transfer service area, see Manufacturing Transfer & Commercialization. For where the transfer seam fits in the broader lifecycle, see Why medical device development breaks down between user needs, engineering, and manufacturing.