FDA 预提交（Pre-Sub）里怎么讨论检测方案

结论：预提交最值钱的用法，是把方案分歧提前暴露

预提交是与审评方在正式提交之前沟通的机会。多数企业用它讨论分类、临床路径、对比器械选择——这些当然重要，但检测方案的分歧同样会导致返工，而返工的周期不比其他环节短。

如果检测做完才发现方法选择不被认可、或者项目覆盖不足，样品可能已经消耗完，重新准备加重测，几个月就过去了。而在预提交里问一句，成本几乎为零。

哪些检测问题值得问

最后一行经常被忽略。 有些性能指标标准里没有给限值，需要企业自己设定并论证合理性。这个论证如果事后被质疑，很难补救；事前确认，风险就低得多。

提问怎么写

提问的质量直接决定答复的价值。几条经验：

一是给出你的方案和理由，再问是否认可。 开放式提问（「我们应该做哪些检测」）得到的答复通常很笼统。带方案的提问（「我们计划用某方法测某项目，理由是……，请确认是否合适」）能得到明确回应。

二是一次只问一件事。 把多个问题揉在一段里，答复可能只回应其中一部分。

三是提供足够的背景。 审评方需要了解产品才能判断方案是否合适。产品描述、预期用途、关键技术特征要说清楚。

四是明确你想要什么。 是要确认、要建议，还是要在两个方案之间选一个。说清楚才能得到对应的答复。

准备材料的要点

预提交材料里与检测相关的部分，建议包含：

产品的技术描述，重点是与检测相关的特征；计划的检测项目清单，逐项说明依据；对比器械的检测情况（如果能获得）；型号规格清单及覆盖论证思路；风险分析中识别的、需要通过检测验证的项目。

最后一项容易被漏掉，但它很有说服力。 把风险与检测项目对应起来，说明每项检测是为了验证什么风险控制措施，这个逻辑链条比单纯列清单要扎实。

答复怎么用

拿到答复之后，几件事要做：

逐条对照落实。 答复里认可的部分按原方案执行；有保留意见的部分调整方案；明确不认可的部分重新设计。

记录答复内容与后续处理。 在正式提交时，可以说明哪些方案是经过预提交沟通的，以及沟通的结论。这能减少重复讨论。

注意答复的性质。 预提交的答复通常是基于当时提供的信息给出的建议，不构成最终承诺。如果后续产品发生变化，之前的答复未必仍然适用。

不要过度解读。 答复中没有提到的部分，不等于被默认认可。只有明确回应的内容才有参考价值。

时间安排

预提交本身需要时间：准备材料、提交、等待答复、可能的会议。这段时间要算进项目计划。

建议的顺序是：产品设计基本定型、检测方案初稿完成之后提预提交，拿到答复再正式开始检测。 太早提，产品还在变，讨论的方案可能作废；太晚提，检测已经开始，答复来了要推翻重来。

如果时间实在紧张，可以考虑分开处理：不确定性高的项目等答复，确定的项目先做。这样既不浪费时间，也控制了返工风险。

常见的浪费

一是不问检测，只问临床。 把宝贵的沟通机会全用在一个方面。

二是问得太笼统。 得到的答复也笼统，没有实际指导价值。

三是问了但不采纳。 答复建议调整方案，企业觉得成本高就按原方案做，结果正式提交时被同样的问题挡住。

四是把预提交当成审批。 预提交是沟通不是批准，答复认可的方案在正式审查时仍可能被追问，只是概率低得多。

会议还是书面答复

预提交可以要求书面答复，也可以申请会议。两者各有适用：

书面答复适合问题明确、答案相对确定的情形，效率高；会议适合问题复杂、需要来回讨论的情形，比如产品形态特殊、方案有多种可能。

如果选择会议，检测方案的讨论要提前准备好材料，会上临时解释产品原理会浪费时间。建议把方案、依据、备选方案都写清楚，会上只讨论分歧点。

与国内申报的协同

如果产品同时要在国内申报，预提交阶段讨论的方案可以顺带考虑国内要求。有些项目两边都要做，方法如果能统一，一次检测两处使用。

这个协同要在方案设计阶段做，等一边测完再考虑另一边，往往因为方法或条件不同而无法复用。

一个容易被忽略的准备动作

提预提交之前，建议先查一查同类产品已上市的情况：它们走了什么路径、公开信息里能看到做了哪些检测、有没有可借鉴的方案。

这些信息能让你的提问更有针对性，也能避免问一些行业内已有共识的问题，把宝贵的沟通额度用在真正不确定的地方。

我们的做法

有委托方会在准备预提交时找我们讨论检测方案的可行性——哪些方法我们能做、需要多少样品、周期多长、有没有更合适的替代方法。这些信息对写预提交材料有用，因为方案要现实可行才有讨论价值。

拿到答复之后，如果方案需要调整，我们也会配合重新评估项目清单和排期。答复中要求增加的项目，有些可能需要特定能力，早点确认能不能做、在哪做，比等到开测前才发现要从容。

如果你在准备预提交、检测方案部分想先过一遍可行性，可以把产品资料和初步清单发过来，或者直接联系：132 4819 8029。能力范围见服务介绍，送检要求见送检要求，流程见检测流程。

English version

Conclusion. A pre-submission is an opportunity to communicate with reviewers before formal filing. Most manufacturers use it for classification, clinical strategy and predicate selection, all of which matter. Disagreement about the testing plan causes rework on a comparable scale, however, and the rework takes just as long. Discovering after testing that a method choice is unacceptable or coverage inadequate may mean samples are consumed, preparation restarts and months pass. Asking the question in a pre-submission costs almost nothing.

Testing questions worth raising. Method selection, where several methods exist for one property and acceptance is uncertain. Item coverage, where it is unclear whether a particular item is required. Coverage justification, where a broad model range invites an argument to reduce testing. Substitute evidence, where literature or historical data might replace some testing. Evaluation of unusual designs, where standard methods do not fully apply. And acceptance criteria, where the standard sets no limit and the manufacturer must establish and justify one. The last is often overlooked and is hard to remedy afterwards; confirming in advance carries much less risk.

Writing the questions. The quality of the question determines the value of the answer. Present your proposal with reasoning and ask for confirmation, since open questions such as what testing should we perform attract general answers while a specific proposal draws a specific response. Ask one thing at a time, because several questions bundled into a paragraph may receive a partial answer. Provide enough background, since reviewers need to understand the device to judge the proposal, so describe the product, intended use and key technical characteristics. And state what you want: confirmation, advice, or a choice between two options.

Preparing the material. The testing portion should include a technical description focused on testing-relevant characteristics, the planned item list with the basis for each, information about testing of the predicate where obtainable, the model list with the coverage argument, and the risks identified in the risk analysis that testing is intended to verify. That last item is easily omitted and is persuasive: mapping risks to test items and explaining what control each test verifies produces a stronger argument than a bare list.

Using the response. Work through it point by point, implementing the plan where endorsed, adjusting where reservations are expressed, and redesigning where it is clearly not accepted. Record the response and what you did about it, so that the formal submission can state which elements were discussed and what was concluded, reducing repeated discussion. Note the status of the response: it is advice based on the information available at the time, not a binding commitment, and may not hold if the product changes. And do not over-read it: silence on a point is not endorsement, and only explicit responses carry weight.

Timing. The process itself takes time to prepare, submit and receive an answer, possibly with a meeting, and that belongs in the project plan. A sensible sequence is to submit once the design is substantially settled and a draft test plan exists, then begin testing after the response arrives. Submitting too early risks discussing a plan for a product that then changes; submitting too late means testing has begun and a response may require reversing it. Where time is tight, separate the work: hold the uncertain items pending the response and start the settled ones.

Common waste. Using the opportunity only for clinical questions. Asking so generally that the answer is equally general. Asking and then not following the advice because it costs more, only to be stopped by the same point at formal review. And treating the pre-submission as approval: it is communication, not authorisation, and an endorsed plan may still attract questions later, though far less often.

How we handle it. Clients sometimes discuss the feasibility of a test plan with us while preparing a pre-submission, covering which methods we can perform, how many samples are needed, how long it takes and whether a more suitable alternative exists. That information is useful when drafting, because a plan has to be realistic to be worth discussing. Once a response arrives, we help re-evaluate the item list and schedule where adjustment is needed; items added by the response may require particular capability, and establishing early whether and where they can be done avoids a scramble.

Send us the product information and your draft item list and we will review feasibility. Phone or WeChat: +86 132 4819 8029.