Bottom line first: samples and data can be shared, conclusions cannot
When several markets are pursued at the same time, the common misconception is "test once, then translate the report into a few versions." What can actually be reused across markets is the same batch of samples, the same set of test records, the same raw data. What cannot be reused is the applicability judgement written into the report. Target markets differ in the report formats they accept, in signing requirements, and in how the test basis has to be phrased. All of that has to be settled when the work is commissioned: how many report sets are needed, who reads each one, and which data may be cross-referenced.
Only one rule really decides whether this works: design the test programme against the broader set of requirements first, then add the market-specific deltas; do not push one market through to completion and then rework the design to accommodate the next one. Get the order wrong and you pay in fresh samples and a fresh slot in the schedule, and both are almost immovable late in a project.
Nail down three things before you commission anything
First, whether the product configuration is frozen. Frozen does not mean "the drawings are final." It means the material grades, suppliers, processing aids, cleaning and sterilization processes, software version and accessory combinations are all fixed and consistent with the production plan. If any one of those is still moving, running markets in parallel has no value, because the value comes from data reuse, and data reuse depends on the item under test being described identically in every market's dossier.
Second, whether the requirement matrix exists on paper. Not "we probably need biological evaluation" but an item-by-item list: which tests every market wants, which ones only one market wants, and which ones carry the same name but a different acceptance basis. The sponsor and the laboratory build this matrix together, and the earlier it is built the cheaper the project. Before commissioning, walk the list against sample and documentation requirements and close the obvious gaps first.
Third, how the reports will be used. Whether a report goes to a competent authority, to a customer, or into an internal file directly determines how it is split, what appears on the cover page, and whether raw traces are attached. Issuing one large report versus several separate reports from the same data set is not hard to change, but raise it after the reports are signed and you are into a revision process, and that costs real time.
What can be shared and what has to be arranged separately
The table below is the working draft used when a parallel project is set up. The left column is the decision dimension, the middle columns give the fork, and the right column is what it actually costs when the call is wrong.
| Decision dimension | Can be shared when | Must be arranged separately when | Cost of getting it wrong |
|---|---|---|---|
| Sample status | Finished-product status is identical and sterilization matches production | Market-specific sales configurations or accessory sets differ | The mismatch surfaces only after samples arrive; the whole batch of testing is void and has to be rescheduled |
| Biological evaluation | Materials, process and sterilization are frozen; contact type and contact duration category match | A market asks for additional evaluation endpoints | Evaluation under ISO 10993 run on non-final samples, later required to be repeated in full |
| Sterile barrier and packaging | Packaging materials, sealing process window and shipping unit are the same | Packaging varies with the sales unit and no grouping rationale was written | ISO 11607 related validation required size by size |
| Electrical safety | Electrical architecture, power supply solution and software version are identical | Power supply forms differ by market, or different collateral parts are invoked | An old report kept in use after a structural change; the report no longer matches the physical device and is rejected |
| Quality system interface | Same manufacturing site, same change control flow | Markets require different technical documentation structures | Sample lot numbers do not reconcile with batch records and the representativeness of the report is challenged |
| Report issuance | Raw data and test records are shared | Cover page, language and format issued per market | Changing the split after signature means a formal revision |
Biological evaluation: plenty of room to share, but it all rides on the documentation
Evaluation under the ISO 10993 framework starts with material characterization and risk assessment, not with putting a sample on the bench. Whether it can be shared across markets depends on whether you can file the same material and process description everywhere. In practice, work through this self-check:
- For every part that contacts the body, are the material grade and supplier fixed? Is any single part being dual-sourced?
- Are the release agents, cutting fluids and cleaning agents used in production on the list? These are routinely overlooked and they do move the evaluation conclusion.
- Do the sterilization method and the sterilization sequence applied to the test samples match production? Data generated on samples that never saw terminal sterilization will usually have to be repeated.
- Do the contact sites and contact durations stated in the instructions for use match the categorization in the evaluation plan? If the two disagree, the applicability of the evaluation does not hold up.
There is a very familiar failure here. To save time, the project team starts the evaluation on prototype parts, planning to "carry it over if the production material is the same." Then one part changes supplier at production launch; the grade is nominally identical but the processing aids are not, and the representativeness of the whole evaluation data set collapses. Fixing this at the plan stage costs nothing. Fixing it after the data exists means new samples, a new slot and a new invoice. For how the evaluation route is laid out, see biological evaluation services.
Sterile barrier and packaging: put the grouping rationale in the plan
Packaging validation splits into two lines: the integrity and seal performance of the sterile barrier system as supplied, and integrity after distribution and storage. Sharing across markets requires the packaging materials, the sealing process parameter window and the shipping unit to be the same.
What really goes wrong is size families. A product usually comes in several sizes and the packaging changes with them. If the plan does not spell out the grouping rationale, meaning which sizes form a group, which size represents the group, and why that coverage logic holds, the laboratory can only schedule each size independently, and both sample consumption and cost multiply. The rationale belongs in the plan with its supporting basis, not in a paragraph added after the report is out; an explanation with no test design behind it is generally not accepted at review. For ISO 11607 related validation arrangements, see packaging validation services.
Electrical safety: share the core, split the deltas
For active devices the core of the safety testing shares well, because the basic safety requirements differ far less between markets than people assume. The differences cluster in a few places: power supply form and plug type, the boundary for judging accessible parts, the language and marking requirements for accompanying documents, and the collateral and particular parts additionally invoked by each market.
The decision rule is direct. Identical electrical architecture, identical power supply solution and identical software version means the main body of IEC 60601-1 testing can be shared and the deltas tested separately; if any one of those three differs, that branch has to be split out in the plan. The classic mistake is a small structural change during testing, a swapped power supply module for example, that the project team decides "does not affect safety," so the earlier report stays in use. That is a report that no longer matches the physical device, and once review notices, the whole report is worthless.
The quality system is not a report, but it decides whether the report is believed
The point of ISO 13485 here is not "having a certificate," it is being able to explain where the samples came from. Reviewers ask three things: can the submitted samples be traced to a batch record; did the product undergo a design change during testing; was the impact of that change on completed testing assessed? In a company whose system actually runs, all three answers are already on file. In a company where it does not, the answers get invented on the spot, and invented answers do not survive follow-up questions.
One concrete action: when samples ship, send the lot number, the manufacturing date and batch information, and an internal confirmation that design changes are frozen for the duration of testing. The confirmation does not have to be elaborate, but it will save you when someone starts asking.
Sample submission and scheduling: the wrong order means doing it twice
| Step | Recommended practice | Common mistake | Consequence |
|---|---|---|---|
| Requirement mapping | Build the multi-market requirement matrix in one pass, marking each item shared or market-specific | Adding requirements as you go | Repeated sample top-ups and inconsistent sample batches |
| Sample preparation | Prepare against the combined upper bound of sample consumption, from one production run | Prepare for a single market's consumption | A second sample batch, and questions about the representativeness of the two batches |
| Plan approval | One master plan, with market deltas written as annexes | A separate standalone plan per market | Conclusions phrased inconsistently; reports hard to keep self-consistent |
| Sequencing | Non-destructive tests first, destructive tests after | Reverse order | Samples consumed early, nothing left for the remaining tests |
| Reporting | Issue the master report first, derive market versions from it | Edit several reports in parallel | Version confusion, revision needed after signature |
One detail on sample quantity deserves emphasis. For parallel projects, over-prepare rather than prepare exactly enough. A top-up means a new batch, and a new batch means writing a justification for consistency with the previous one, a document that costs more than the extra samples would have.
When to give up on running in parallel
Parallel work is not always worth it. If any of the following applies, lock down one lead market, get it through, and extend from there:
- the design is still in an active change phase and changes arrive faster than testing progresses;
- materials or suppliers are not frozen, so the object of the evaluation can shift at any moment;
- the technical requirements of the target markets diverge so much that the shared portion is small;
- samples are hard to obtain and you cannot assemble the quantity that parallel work needs.
Forcing it usually produces the same pattern: test, change, scrap a set of data, repeat, with a total cost above running the markets in sequence. The test is those three conditions: design frozen, changes controlled, samples available. If any one fails, do not run in parallel.
On choosing a laboratory and on accreditation
Many people look only for an accreditation mark. It is more useful to check whether the specific tests fall inside the accredited scope, and whether the reports can cover every market you intend to file in. This has to be said plainly: an accreditation mark only demonstrates that the laboratory holds the corresponding technical competence within its accredited scope, and it is not a commitment as to market access outcomes. Any claim that accreditation amounts to a promise of approval is unfounded, because the access decision is made by the competent authority of the target market.
Parallel projects also touch on schedule planning and on what to do when results are disputed. Each of those has its own logic and is covered separately in testing knowledge rather than here.
Let us map the requirement matrix and the sampling plan with you
If you already have a target-market list and a product structure description, we can start with a parallel feasibility review: which tests can be merged, how many samples to prepare, how the reports get split, and at which step the market deltas are added. Settle that at the plan stage and there is very little rework afterwards. SUNGO Lab (Shanghai Shage Medical Technology Co., Ltd.) is a medical device testing laboratory accredited by CNAS, CMA and IAS (USA), with laboratories in Shanghai and Hefei. Call +86 132 4819 8029 to talk through your product, or request a quote and we will come back with a sample list and a schedule proposal matched to your market combination.