注册申报资料里检测报告被质疑的常见问法

结论：提问的角度是固定的，可以提前防

审评人员看检测报告，关注的点相对集中：报告本身是否有效、内容是否覆盖、数据是否可信、结论是否成立、与其他资料是否一致。

这些角度是稳定的，所以提问也是可预期的。 在送检阶段按这几个角度自查一遍，能挡掉相当一部分补正。

六类常见问法

逐类说明

报告的资质与范围

关心的是出具报告的机构在这个项目上有没有相应资质。常见的漏洞是：机构整体有资质，但具体到某一项试验不在认可范围内。报告上会有相应标识，而申报方往往没有逐项核对。

委托前就应当确认：把项目清单给到实验室，请对方确认哪些在认可范围内、哪些需要分包、分包方是谁。这件事委托后再问，已经晚了。

项目覆盖是否完整

关心的是适用标准里该测的条款有没有都测。产生遗漏的典型原因是标准适用性判断出错——认为某条不适用，实际适用。

预防办法是做一张条款对照表：把适用标准的条款逐条列出，标明测了、不适用、或者不涉及，不适用的要写明理由。这张表本身可以作为申报资料的一部分，主动提供比等着被问要好。

样品的代表性

关心的是测的样品能不能代表申报产品。会被问到的情形包括：样品型号与申报型号不一致、样品是试制件而非定型产品、样品生产日期早于工艺定型时间。

预防办法是保证样品信息在各处一致：委托单、报告、申报资料上的型号规格、批号、生产日期要对得上。

型号规格覆盖

关心的是未送检型号凭什么认为符合要求。这是提问率较高的一类。

覆盖论证的基本逻辑是选取不利型号检测，说明其余型号不会更差。要注意的是不同项目的不利型号可能不同：强度项可能是尺寸大的，温升项可能是功率大的。笼统地说「选取典型型号」而不分项说明，容易被追问。

数据的可信度

关心的是报告里的数据是不是在规范条件下取得的。会看的包括：试验条件记录是否完整、使用的设备是否注明、方法依据是否明确、原始数据与结论是否对应。

这一类问题多数不是数据本身有问题，而是报告记录不够完整。委托时可以要求实验室在报告中体现关键条件，不要只给结论。

与其他资料的一致性

关心的是报告与技术要求、说明书之间对不对得上。这是最容易自查也最容易忽略的一类。

把三份文件并排，逐个核对重复出现的参数。不一致的情形包括：技术要求写的限值与报告判定依据不同、说明书标称值与报告实测值差距过大、标签信息与样品实物不符。

送检阶段的预防清单

把上面六类倒过来，就是送检前该做的事：确认实验室的认可范围覆盖你的项目；做标准条款对照表；核对样品信息与申报资料一致；分项做覆盖论证；要求报告体现关键试验条件；三份文件交叉核对参数。

这六件事加起来大概一两天的工作量，能挡掉的补正往返通常以月计。

被问到之后怎么回

如果补正已经来了，回复时有两个原则：

一是正面回答，不绕。 问的是覆盖论证，就把论证逻辑写清楚，不要用「参见某某资料」带过。

二是能补数据就补数据。 如果问题的根源是数据不足，用说理弥补通常不成功。判断补测的周期是否来得及，来得及就补，来不及要提前与相关方沟通时间安排。

主动提供比被动回答好

有几份材料，如果在首次提交时就主动附上，能明显降低被问的概率：标准条款对照表、型号规格覆盖论证、检测项目与技术要求的对应表。

这三份的共同点是它们回答的都是审评一定会想到的问题。等被问了再提供，中间要多走一轮补正；主动提供，审评看到逻辑完整，追问的动力就小。

制作这三份材料的工作量不大，多数信息在准备检测时已经产生，只是没有整理成文。

报告到手后的自查

拿到检测报告不要直接归档，建议先自查一遍：型号规格与申报一致吗；试验条件记录完整吗；有没有出现「不适用」而未说明理由的项目；结论页与明细页是否一致；报告上的标识与认可范围对得上吗。

这一步发现问题还来得及请实验室更正，提交之后再发现，更正报告要走完整流程，时间成本高得多。

一类特殊情形：报告来自境外实验室

如果检测报告由境外实验室出具，还会多一层问题：该机构的资质在境内如何认定、报告依据的方法版本与境内现行版本是否一致、数据是否需要转换。

这一类情形建议提前确认接受口径，不要等提交后才发现报告形式不被接受。确认的成本是一次沟通，不确认的成本可能是整套重测。

我们的做法

我们出报告时会尽量把关键试验条件写进去，而不只给一个合格结论。原因就是上面第五类——条件记录完整的报告，被追问的概率明显低。

对于用于注册申报的项目，我们也会在报告出具前确认样品信息与委托单一致，避免型号、批号这类低级但高发的不一致。

如果你收到了涉及检测的补正意见，可以把意见和报告一起发过来看看怎么回，或者直接联系：132 4819 8029。能力范围见服务介绍，流程见检测流程，更多内容见知识库。

English version

Conclusion. Reviewers examine test reports from a stable set of angles: is the report itself valid, is the coverage complete, are the data credible, does the conclusion hold, and is it consistent with the rest of the dossier. Because the angles are predictable, so are the questions, and self-checking against them before submitting samples removes a good share of deficiency letters.

Accreditation and scope. The question is whether the issuing body is accredited for that particular test. A common gap is a body that is accredited overall but not for one specific item. Confirm before commissioning: give the laboratory the item list and ask which items fall inside its accredited scope, which will be subcontracted, and to whom.

Completeness of coverage. The question is whether every applicable clause of the applicable standards was tested. Omissions usually arise from misjudging applicability. Build a clause-by-clause table marking each as tested, not applicable with a reason, or not relevant. That table can be submitted proactively rather than waiting to be asked for it.

Sample representativeness. The question is whether what was tested represents what is being registered. Problems arise when the sample model differs from the registered model, when pilot units were used instead of definitive product, or when the build date precedes process finalisation. Keep model, batch and date consistent across the test request, the report and the dossier.

Model and variant coverage. The question is why untested variants should be considered compliant. The logic is to test the least favourable variant and argue the others cannot be worse. Note that the least favourable variant differs by item: the largest size for strength, the highest power for temperature rise. A blanket statement about testing a typical model, without breaking it down by item, attracts follow-up questions.

Credibility of data. The question is whether the data were obtained under controlled conditions. Reviewers look for complete records of test conditions, identification of equipment, a clear method basis, and correspondence between raw data and conclusions. Most problems here are not with the data but with incomplete reporting. Ask the laboratory to record key conditions rather than issuing conclusions alone.

Consistency with other documents. Place the technical requirements, the instructions for use and the report side by side and reconcile every parameter appearing in more than one. Mismatches between stated limits and the acceptance basis used in testing are common.

A pre-submission checklist. Confirm accredited scope; build the clause table; reconcile sample information; argue coverage item by item; require key conditions in the report; cross-check parameters across documents. A day or two of work that typically prevents months of back and forth.

If a deficiency letter arrives. Answer directly rather than referring elsewhere, and where the root cause is missing data, supply data rather than argument. Judge whether the retest fits the timetable and communicate early if it does not.

How we handle it. We record key test conditions in reports rather than issuing bare conclusions, because completely documented reports attract materially fewer questions. For registration work we also verify that sample information on the report matches the request form before issue, to avoid the simple but frequent mismatches in model and batch.

Send us the deficiency letter together with the report and we will work through the response. Phone or WeChat: +86 132 4819 8029.