MDSAP 审核对检测文件的额外关注点

结论：一次审核，多套要求叠加

MDSAP 的特点是一次审核覆盖多个市场的监管要求。对检测文件而言，这意味着准备的口径要取各方要求的并集，而不是交集。

单一体系审核时，某项记录按一个口径准备就行；MDSAP 下，如果某个参与市场对该记录有额外要求，就要一并满足。这是文件准备工作量增加的主要来源。

与单一体系审核的差异

最后一行影响很大。 MDSAP 的审核流程是规定好的，审核员按既定路径推进，不太会因为你解释得好就跳过某个环节。所以准备工作要覆盖全流程，不能指望重点应付。

检测相关的关注点叠加

把几个常见的叠加点列出来：

设备校准的溯源要求。 各市场对溯源链的表述要求可能不同，准备时按较严的口径，把溯源关系写清楚。

供方（外部实验室）的控制。 有的市场对外部服务提供方的评价与监控有更具体的要求，评价记录要做实。

不合格品与风险的联动。 检测发现不合格之后，是否触发了风险评估、是否考虑了上市产品的影响，这条链在 MDSAP 下会被追得更紧。

上市后信息回到检测的闭环。 投诉或不良事件涉及性能问题时，有没有相应的检测复现与分析记录。

记录保存期限。 各市场要求不同，取较长的。

文件准备的实务做法

建议按这个思路组织：

先做一张要求对照表。 列出各参与市场对检测相关记录的具体要求，标出差异点。这张表做出来，才知道要按哪个口径准备。

统一口径而不是分别准备。 为不同市场准备不同版本的记录，维护成本高且容易出错。统一按较严口径做一套，各市场都能用。

把关联关系做显性。 MDSAP 很看重过程之间的衔接：检测发现问题如何进入 CAPA、CAPA 如何关联风险管理、风险管理如何影响设计。这些关联如果只存在于人的脑子里，审核时说不清楚。建议用表格或流程图把关联显性化。

准备追溯演练。 审核员会选一个产品批次往回追。自己先演练一遍，看链条上哪里会断。

常见的检测相关不符合项

从实务经验看，这几类出现得较多：

外部实验室的评价记录不充分，只有资质文件没有能力评价；检测记录与批记录之间的关联不清晰，无法确认某批产品对应哪份检测数据；不合格结果的处置未联动风险评估；设备校准的量程覆盖不足，且未被识别；方法变更后未重新确认；上市后发现的性能问题未回溯到相应的检测环节。

这几项的共同点是「关联断裂」——单看每个环节都有记录，但环节之间连不起来。

时间安排上的建议

MDSAP 的准备周期通常比单一体系审核长。检测相关的部分建议提前处理的有：

外部实验室的评价记录补齐，这项需要对方配合，临时办来不及；设备校准计划梳理，确认审核期间不会有设备处于过期状态；历史记录的抽查与整改，发现问题需要时间补；追溯演练，至少做两三次。

审核后的持续维护

通过审核不是终点。MDSAP 有监督审核，而且各市场的法规会变化。建议建立一个机制：定期回看要求对照表，法规有更新时评估对现有记录的影响。

这件事没有专人负责就会荒废，等到下次审核前才发现要求已经变了。

审核抽样的应对

MDSAP 的抽样有既定规则，审核员会按流程抽取一定数量的记录。这意味着「重点准备几份样板」的做法无效——抽到什么就看什么。

所以准备的思路应当是把整体水平提上来，而不是准备几份精品。做法是先自查一批随机记录，发现的共性问题批量整改，而不是挑几份做成范本。

跨部门的协调

检测相关的记录分散在质量、生产、研发、采购几个部门。MDSAP 追溯时会跨部门走，任何一个部门的记录断了，链条就断了。

建议在准备阶段做一次跨部门的联合演练：由一个人扮演审核员，按追溯路径逐个部门问过去。这个演练通常能发现部门交界处的空白——每个部门都以为对方在管。

文件语言的问题

MDSAP 涉及多个市场，文件语言要求可能不同。检测报告如果是中文出具，用于某些市场时需要翻译，而翻译件的效力和翻译资质可能有要求。

建议在委托时就明确需要的语言版本，由实验室直接出具英文报告或中英对照报告，比事后翻译更稳妥——事后翻译的版本需要额外证明翻译准确性。

审核发现的整改

MDSAP 的不符合项有分级，不同级别的整改时限和要求不同。检测相关的不符合项整改时，要注意两点：一是整改要覆盖同类问题而不只是被指出的那一处；二是有效性证据通常需要新产生的记录，而不是补齐旧记录。

补齐旧记录是常见的错误做法。 事后补填的记录在时间逻辑上站不住，反而会加重问题性质。正确的做法是承认缺失、说明原因、并用之后的记录证明流程已经改正。

我们的做法

作为外部实验室，我们能配合的主要是两件事：提供完整的资质与认可范围文件，供委托方完成供方评价；以及在报告中保留足够的信息（方法、条件、设备、日期），使委托方能把报告与自己的批记录关联起来。

报告信息不全是关联断裂的常见起点。 报告上如果没有明确的样品标识，委托方就无法证明这份报告对应哪批产品。所以委托时把样品标识信息写清楚，对后续的体系审核有直接帮助。

如果你在准备 MDSAP 审核，检测相关的记录有不确定的地方，可以把具体要求发过来一起看，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. The defining feature of MDSAP is that one audit covers the regulatory requirements of several markets. For testing documentation this means preparing to the union of those requirements rather than their intersection. Under a single-scheme audit a record prepared to one standard suffices; under MDSAP, if any participating market imposes an additional requirement on that record, it must also be met. That is the main source of extra preparation work.

Differences from a single-scheme audit. The basis is the quality system standard plus the regulatory requirements of each market rather than the standard alone. Record requirements accumulate market-specific additions. Some markets require deeper traceability. Some have particular post-market reporting expectations. And the audit process itself follows a defined route with defined sampling, which matters considerably: the auditor works through a prescribed path and is unlikely to skip a step because an explanation was persuasive, so preparation must cover the whole process rather than selected high points.

Where requirements accumulate for testing. Equipment calibration traceability, where the way the traceability chain must be expressed can differ between markets, so prepare to the stricter reading and state the chain explicitly. Control of external providers, where some markets are more specific about evaluation and monitoring, so evaluation records must be substantive. The link between nonconforming results and risk, where whether a nonconformity triggered risk assessment and consideration of marketed product is pursued more closely. The loop from post-market information back to testing, where complaints or adverse events involving performance should show corresponding reproduction and analysis records. And record retention periods, where the longest requirement governs.

Practical preparation. Build a requirements comparison table listing what each participating market expects of testing-related records and marking the differences; only with that table can you know which standard to prepare to. Prepare one set to the strictest reading rather than separate versions per market, since maintaining parallel versions is expensive and error-prone. Make relationships explicit, because MDSAP attends closely to how processes connect: how a test finding enters CAPA, how CAPA connects to risk management, how risk management feeds design. Relationships existing only in people's heads cannot be explained during audit, so document them as tables or flow diagrams. And rehearse traceability, because the auditor will pick a batch and work backwards; running that exercise yourself shows where the chain breaks.

Frequent testing-related findings. Insufficient evaluation of external laboratories, with accreditation certificates on file but no capability assessment. Unclear linkage between test records and batch records, so that no one can confirm which data belong to which batch. Nonconforming results handled without linking to risk assessment. Calibration ranges not covering actual measurements, with the gap unidentified. Methods changed without revalidation. And post-market performance problems not traced back to the corresponding testing. What these share is broken linkage: each step has records, and the steps do not connect.

Timing. Preparation generally takes longer than for a single-scheme audit. Address early the items that depend on others: completing external laboratory evaluations, which needs their cooperation and cannot be arranged at short notice; reviewing calibration schedules so no instrument falls out of validity during the audit period; sampling and correcting historical records, since remediation takes time; and rehearsing traceability at least two or three times.

Maintaining afterwards. Passing is not the end. Surveillance audits follow and market regulations change. Establish a mechanism to revisit the requirements comparison table periodically and assess the effect of regulatory updates on existing records. Without a named owner this lapses, and the change is discovered only when the next audit approaches.

How we handle it. As an external laboratory we can help with two things: supplying complete accreditation and scope documentation for the client's supplier evaluation, and retaining sufficient information in reports, covering method, conditions, equipment and dates, so that reports can be linked to the client's batch records. Incomplete report information is a common starting point for broken linkage: without clear sample identification on the report, a client cannot demonstrate which batch it belongs to. Providing precise sample identification when commissioning therefore pays off directly in later audits.

Send us the specific requirements you are working to and we will review. Phone or WeChat: +86 132 4819 8029.