变更控制怎么触发重新检测

结论：判断的依据是影响路径，不是变更大小

变更控制流程走到「要不要重新检测」这一步，常见的做法是凭经验判断：改得小就不测，改得大就测。这个做法在简单情形下能用，遇到不典型的变更就容易出错。

可靠的做法是追影响路径：这次变更改变了什么物理量或过程条件，这些改变会传导到哪些产品特性，这些特性由哪些检测项目覆盖。追到最后，需要重测的项目自然就出来了。

决策路径

建议按这个顺序问：

第一问：变更改变了什么？ 要具体到物理层面——尺寸、材料成分、表面状态、工艺参数、软件逻辑。写「优化了结构」不算回答。

第二问：这个改变会影响哪些产品特性？ 从改变出发，列出可能受影响的特性。这一步要放宽，宁可多列。

第三问：这些特性由哪些检测项目覆盖？ 对照技术要求和检测报告，找到对应项目。

第四问：已有数据还能代表变更后的产品吗？ 如果不能，这个项目就需要重测。

第五问：有没有可以替代重测的证据？ 比如供应商的材料等同证明、理论分析、类似变更的历史数据。有充分替代证据的，可以不重测但要留论证记录。

常见变更与对应项目

这张表是起点不是终点。实际判断还要看变更的幅度——同样是改尺寸，改零点几毫米和改几毫米，影响可能完全不同。

替代证据怎么用

不是所有变更都必须重测。可用的替代证据包括：

材料等同论证。 新旧材料在成分、性能、加工方式上等同，可以沿用原数据。论证要有材料数据支撑，不能只凭供应商声明。

理论分析。 对于力学类变更，如果能通过计算说明变更后的余量仍然充分，可以减少实测。但计算要有依据，且通常需要一定的实测数据作为校核。

历史数据。 如果过去做过类似变更并有完整数据，可以作为参考。前提是「类似」要有明确定义。

风险分析结论。 通过风险分析说明该变更不引入新的风险，且不影响现有风险控制措施的有效性。

替代证据的使用要留记录，说明为什么这个证据足以替代实测。这份记录在审核和变更申报时都会被看。

判断记录该怎么写

变更评审的记录里，关于检测的部分应当包含：变更的具体内容（物理层面）；识别出的可能受影响特性；对应的检测项目清单；每个项目的处理决定（重测、沿用、用替代证据）及理由；重测项目的计划与实际结果。

理由是这份记录的核心。 只写决定不写理由，日后无法复核判断是否正确，也无法在类似情形下复用。

批量小变更的处理

实际生产中会积累很多小变更，逐个走完整流程成本很高。可行的做法是分层：

定义一个明确的清单，列出哪些类型的变更属于免评审范畴（通常是对产品特性无影响的，比如包装箱外标签排版）；其余变更走简化评审；涉及前面那张表里任一行的，走完整评审。

关键是要有明确的分类规则并严格执行，而不是凭当事人判断。规则模糊时，倾向于简化处理是人之常情，风险就这样累积起来。

同时要定期回看累积的小变更。单个无影响，叠加起来可能已经改变了产品状态，这一点在延续注册或者监督检查时会显出来。

变更的分类标准怎么定

企业自己要有一份变更分类标准，明确什么级别的变更走什么流程。制定这份标准时的常见误区是按变更涉及的工作量分级，而应当按变更对产品特性的影响程度分级。

一个改动可能实施起来很简单（换一个供应商的同牌号材料），但影响可能很大；另一个改动实施很麻烦（重新设计包装盒结构），影响却有限。按工作量分级会把这两者放反。

变更与检测数据的版本对应

产品经过多次变更后，会积累多份检测报告，分别对应不同的产品状态。这时候要能说清楚：当前在产的产品状态，对应的是哪几份报告的组合。

建议维护一份「当前有效检测证据清单」，每次变更后更新：哪些项目沿用原报告、哪些换成了新报告。没有这份清单，几次变更之后就会出现说不清哪份报告代表当前产品的局面。

变更实施后的确认

变更批准并实施之后，还有一步常被省略：确认变更确实按批准的方案落地了。

实际中出现过的情况是，批准的是 A 方案，生产实施时因为某些原因改成了 B，而变更记录没有更新。此后所有基于 A 方案的验证结论，实际上都对不上产品。

建议在变更关闭前做一次实物确认，核对实际产品与变更方案一致。这一步成本很低，但能避免文件与实物脱节。

我们的做法

接到变更相关的检测委托，我们会要变更说明和原始报告。有原始报告才能判断哪些项目具备可比性、哪些方法条件必须保持一致。

实际中常见的情形是，委托方按自己的理解列了重测清单，我们看完变更说明后发现清单与变更的影响路径对不上——有的项目不必测，有的项目漏了。把影响路径梳理一遍，清单通常会有调整。

如果你手上有变更方案但不确定要重测哪些项目，可以把变更说明和原报告发过来一起理，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. When change control reaches the question of whether retesting is needed, the usual approach is judgement by experience: small change, no test; large change, test. That works for straightforward cases and fails for unusual ones. The reliable approach is to follow the path of influence: what physical quantity or process condition did this change alter, which product characteristics does that alteration propagate to, and which test items cover those characteristics. Follow that through and the items requiring retest emerge on their own.

A decision path. First, what did the change alter, stated physically: a dimension, a material composition, a surface condition, a process parameter, a software behaviour. Writing that the structure was optimised is not an answer. Second, which product characteristics could that alteration affect, listed generously at this stage. Third, which test items cover those characteristics, found by reference to the technical requirements and existing reports. Fourth, can existing data still represent the changed product; if not, that item needs retesting. Fifth, is there evidence that could substitute for retesting, such as a material equivalence argument, theoretical analysis, or historical data from a comparable change. Where substitution is adequate, retesting may be avoided but the reasoning must be recorded.

Typical changes and the items they touch. A change of material grade affects mechanical properties, biocompatibility and ageing, implying strength, biological evaluation and accelerated ageing. A dimensional change affects strength, stiffness and stability. A change of surface treatment affects contact properties and corrosion resistance. A change of electronic components affects clearance, temperature rise and emissions, implying electrical safety and EMC. A software change affects function and essential performance. A process parameter change affects consistency and process capability. A packaging material change affects barrier properties and transport protection. A sterilisation parameter change affects sterility assurance, material properties and residuals.

This mapping is a starting point rather than a conclusion, since magnitude matters: changing a dimension by a fraction of a millimetre and by several millimetres may have entirely different consequences.

Using substitute evidence. A material equivalence argument, supported by composition and property data rather than a supplier's assertion alone, may allow existing data to stand. Theoretical analysis may reduce measurement for mechanical changes, provided the calculation has a proper basis and is usually checked against some measured data. Historical data from a comparable change may serve, provided comparable is defined. And a risk analysis may show the change introduces no new risk and does not undermine existing controls. In each case, record why the evidence is considered adequate, because that record will be examined in audits and in change submissions.

What the decision record should contain. The specific content of the change at a physical level; the characteristics identified as potentially affected; the corresponding test items; the decision for each item, whether retest, rely on existing data, or use substitute evidence, with reasons; and the plan and results for items retested. The reasons are the core of the record. A decision without a reason cannot be reviewed later for correctness and cannot be reused in comparable situations.

Handling accumulated minor changes. Production generates many small changes, and running each through the full process is expensive. A tiered approach works: define an explicit list of change types exempt from review, typically those with no effect on product characteristics such as the layout of a shipping carton label; route others through a simplified review; and route anything touching the mapping above through full review. What matters is having explicit classification rules and applying them, rather than leaving it to individual judgement, because where rules are vague the human tendency is towards the simpler route and risk accumulates. Review the accumulated minor changes periodically as well, since individually harmless changes may together have shifted the product, which becomes visible at renewal or inspection.

How we handle it. For change-related work we ask for the change description and the original report, because the original tells us which items are genuinely comparable and which conditions must be held constant. A frequent outcome is that the client's retest list does not match the actual path of influence: some items are unnecessary and another has been missed.

Send us the change description and the original report and we will work through it. Phone or WeChat: +86 132 4819 8029.