Writing a Medical Device Technical File in South Korea — A Section-by-Section Guide

Key takeaway — The technical file has seven sections, but the review asks a single question: "do the numbers and statements inside this document agree with each other, and with the supporting evidence?" Before you learn the drafting conventions for each section, secure the three-way match between specification sheet, test report and labelling. That is where most deficiency letters come from.

Technical documentation is a rebuild, not a translation

It is tempting to assume that translating the manufacturer's specification sheet and manual gets you a technical file. It does not. Korean technical documentation requires information to be re-laid out according to the section framework prescribed by the Regulation on Approval, Notification and Review of Medical Devices — shape and structure, raw materials, intended use, performance, test specifications, directions for use, precautions for use, and so on. An overseas brochure is a document written to sell; technical documentation is a document written to be reviewed. Even with identical information, the review outcome changes depending on which section it lands in and what figures it is expressed with.

The reviewing body also depends on classification. Class II is normally handled by a technical document review body designated by MFDS (Ministry of Food and Drug Safety, formerly KFDA), while Class III and Class IV are reviewed by MFDS directly. The broader shape of the process is covered in the difference between notification, certification and approval.

One further point worth flagging. Technical documentation is not a write-once document — it becomes the baseline for change management after your licence is granted. If the raw materials, performance or directions for use you record today change later, the first question is whether that triggers a change approval or change certification. So "how tightly do we want to pin down specifications that may well change" is itself part of the drafting strategy.

What reviewers check, section by section

Section	What goes in it	The reviewer's question
Shape and structure	External form, dimensions, ratings, principle of operation	What is this product and how does it work?
Raw materials	Material and grade for each component	Which parts contact the body, and are they safe?
Intended use	Indications, target population	Does the evidence support the claimed scope?
Performance	Quantified performance figures	Is there a test report behind every number?
Test specifications	Test items, acceptance criteria, methods	Is the choice of standards justified?
Directions for use	Preparation → operation → storage	Does it guide safe use?
Precautions for use	Warnings, contraindications, adverse events	Does it match the manufacturer's IFU?

1. Shape and structure (including principle of operation) — This is where the reviewer first forms an understanding of the product. Provide external photographs or drawings, dimensions and weight, electrical ratings and component nomenclature, then describe in the principle-of-operation text "by what physical mechanism the product achieves its intended use." The goal here is accuracy, not length — if the principle described here does not connect logically to the intended use and performance sections, everything downstream wobbles. For multi-component products, list the main unit, accessories and consumables exhaustively; and if the ratings or dimensions stated here differ from the performance section or the test reports, that alone is a deficiency.

2. Raw materials — The heart of this section is the patient-contacting parts. Tabulate component name, material (chemical name or trade name) and grade; the contact site and contact duration determine the scope of biological safety evaluation under the ISO 10993 series. A thin raw materials table triggers a chain of requests for biological safety data. The practical bottleneck is obtaining material data from further up the supply chain, at the resin or raw material producer — it is common for even the device manufacturer to know only the supplier's trade name and not the chemical composition. For import projects, raising that data request at the very start of the project is central to holding the schedule.

3. Intended use — This section is the scope of your licence. The rule is to claim only as much as your performance data and supporting evidence cover, and changing a single word — "relief of" to "treatment of" — can shift both the product classification and the data package required. This is where you write what you can substantiate, not what marketing would like to say. A safe approach is to review the intended use wording of comparable products already licensed in Korea and settle your own wording against the scope of your evidence.

4. Performance — Quantitative expression is the rule. Not "excellent measurement accuracy" but "within ±X%," as a number. And every figure written here must be provable with a test report. Put the other way round: any figure you cannot prove is safer left out. The performance section is not where you boast; it is where you commit. Copying specifications straight out of a brochure, and letting in figures that appear nowhere in the test reports, is the single most common mistake in import projects.

5. Test specifications — There is a priority order for selecting standards. If an MFDS-notified standard exists for that product category, it takes precedence; where none exists, apply international standards such as IEC 60601-1 and IEC 60601-1-2 (EMC); and for product-specific performance that neither covers, apply an in-house standard — with the rationale for each choice available on request. Write this section as a three-column structure of test item, acceptance criterion and test method, and check that each item matches the actual test conditions in the report. The conditions for relying on foreign test reports are covered in our guide to using overseas test reports.

6. Directions for use — Write this as a flow: preparation before use, sequence of operation, storage and maintenance after use. It is not a summary of the whole manual but a distillation of "the procedure for using this safely," and what you write here must not diverge from the labelling and package insert. For reusable products, including the cleaning and disinfection procedure and the replacement interval for consumables will cut down follow-up questions from the reviewer.

7. Precautions for use — Warnings, contraindications and adverse reactions. A warning that appears in the manufacturer's IFU but is missing from the Korean technical file is a deficiency; conversely, inserting content that does not appear in the IFU creates a consistency problem of its own. There is one principle: take what exists in the source documents (IFU, risk management file) and transpose it into the Korean format.

Beyond these seven sections, some products additionally require storage conditions and shelf life (particularly for sterile products and consumables), manufacturing method, and others. Which sections are mandatory varies by product category and classification, so the right first step is to check the drafting requirements for your specific category before filing.

Seven sections, one review — consistency

What actually consumes time in a technical file review is not the completeness of each individual section but the cross-checking between them. Figures drawn from the specification sheet (shape and structure, performance), the test results and conditions in the reports, and the wording destined for labelling (intended use, directions for use, precautions) — these three axes have to line up. A discrepancy as small as an electrical rating of DC 12V in the specification sheet against 12.6V in the test report will prompt a request for explanation, and if the explanation is thin, it leads to retesting.

The practical trick, then, is to invert the drafting order. Rather than writing the document first and reconciling test reports to it afterwards, write the performance and test specification sections from the figures in the test reports you already hold, and fix the intended use within that boundary. For items where you do not yet have a report, sharing the draft figures from your technical file with the test laboratory before commissioning the testing is how you lower the risk of having to retest.

When you receive a deficiency letter, you must submit the requested data within the period granted (normally 30-day increments, extendable on request). Because that period is excluded from the official processing time, each round of deficiencies pushes the whole schedule back accordingly. Class II certification is generally quoted at around three to four months, and this is mostly why the timeline feels longer in practice. For the details of the class-specific procedure, see our Class II certification guide.

Common reasons for rejection and deficiency letters

The specific reasons vary by product, but the patterns repeat with remarkable regularity. The following come up most often in practice.

• Numerical mismatches between specification sheet, test report and labelling (electrical ratings, dimensions, performance values)
• Qualitative language in the performance section — unsupported claims such as "excellent" or "fast"
• Overstated intended use — indications that the performance or clinical data do not support
• Insufficient raw material information for patient-contacting parts (upstream supply chain data not obtained)
• No rationale given for the choice of test standards — applying international standards without first checking whether an MFDS-notified standard exists
• Warnings or contraindications from the manufacturer's IFU omitted from the Korean precautions section
• Incomplete component list — accessories included in the actual commercial configuration missing from shape and structure

Checklist before you proceed

• Product category and classification confirmed; checked whether an MFDS-notified standard exists for that category
• Original manufacturer specification sheet, drawings, IFU and test reports obtained
• Cross-reference table built between the figures intended for the performance section and the figures in the test reports
• Patient-contacting parts list finalised; material data obtained, including from upstream raw material producers
• Intended use wording finalised — claimed scope confirmed against the scope of supporting evidence
• Draft labelling cross-checked against the technical file sections

---

A technical file is not a contest of volume but a contest of consistency. Thirty pages whose numbers agree will clear review faster than a well-written hundred. Send us the manufacturer's specification sheet and a list of the test reports you hold, and our free pre-review will identify where deficiencies are most likely to arise.