近日,我们跟踪研究美国食品安全的网友李和平翻译报道,美国FDA正式宣布采用新的电子筛查系统保障进口食品安全,并就这一系统的启用对出口食品企业提出应对建议。本文仅代表网友个人观点,登载此文出于传递更多信息之目的,并不意味我们赞同其观点或证实其描述,文章内容仅供参考,本文如需转载需经我们许可并保留原始链接。
近日,美国食品药品管理局(FDA)正式宣布采用新的电子筛查系统保障进口食品安全。该系统称作:规范进口货物动态风险评价预测系统(Predictive Risk-based evaluation for Dynamic import Compliance Targeting: PREDICT,简称“预测”)。通过对几百个数据库中的现有数据实施风险筛查,该系统能够将所发现的相关信息反馈到FDA的决策系统中,即接收进口货物决策系统中、并为相关进口货物的“自动放行”或“扣留、抽样检测”决策提供参考建议。在具体运行过程中,“预测”系统筛查范围涉及下列内容:
进口食品自身风险等级(如:水产品原料具有较高的安全风险);
相关进口公司(如:生产商、运输商)或进口食品是否仍处于进口预警通报中;
相关公司(如:生产商、运输商、进口商)及食品的历史记录情况,具体包括:召回记录、抽样检查结果、口岸现场检查结果以及过去的进口抽样结果;
进口食品相关生产商、出口商、进口商、委托人的记录;
进口食品相关产品、生产商、及国外所在地相关信息资源(如:洪水或酷暑等气候条件变化情况);及
食品进口过程中发现的其他非异常数据情况。
进口食品筛查系统(PREDICT)同时会考虑风险结构及其他相关因素。例如,对每个具体进口项目(批)次,“预测”系统都给出一个量化风险数值;基于该量化风险数值,“预测”系统将对该进口食品货物发出“可以通行”通知、或标明该货物具有潜在风险需要送交FDA检查人员对相关文件进行手工处理、并可能进行现场检查和抽样。另外,“预测”系统将取代FDA现行的进口货物筛查管理系统(OASIS)。由于现行系统采用手工方式对生产商、进口商、或产品相关信息进行检索,在一定程度上更依赖审查人员的记忆来完成工作,因此在决定具体进口食品货物是否实施“放行”或“扣留、抽检”工作中存在人为的差异性、缺乏规范性。目前,“预测”系统已经在加州洛杉矶口岸完成了测试,在纽约口岸正在进一步运行和完善中,不久将在美国其他口岸推广运用。 预计在2010年夏季结束前,“预测”系统将在美国全国范围内完成调试。此外,“预测”系统将明显提高进口食品的自动“通关(无需进一步检验)”数量。洛杉矶口岸的测试结果显示,“自动通关(无需进一步检验)”数提高了60%。
面对美国正在采用、即将推出的进口食品安全筛查系统(PREDICT系统),检验检疫机构建议食品出口企业从以下几个方面调整工作:1.认真对待每一次出口机会。向美国出口食品是一个长期、具有联续性的工作;不能抱有侥幸、不负责任的态度,要确保每一批输美食品具有良好的质量;2.慎重选择出口商、代理商。新形势下,出口商、代理商的声誉也会影响到输美食品的顺利出口;3.食品生产企业要注重日常生产。“预测”系统中明确提出要综合考虑食品生产企业的各种信息,在信息高速发达的今天,发生在世界任何角落的一件事情都有可能被放大、缩小或被利用;所以,食品出口企业一定要珍惜自身的日常声誉;4. 出口食品企业要树立集体意识、国家意识。食品出口工作是一项充满挑战的工作,某个产品出现问题后不仅仅会影响到该产品的声誉,更重要的是会波及到行业、甚至国家的整体形象。
附:英文原稿如下
FDA launches PREDICT screening system for food imports
The U.S. Food and Drug Administration (FDA) has announced the launch of a new electronic screening tool for food imports: Predictive Risk-based evaluation for Dynamic import Compliance Targeting (PREDICT). PREDICT is a risk-based screening system that automatically scans hundreds of databases for information that may affect FDA’s admissibility determination. For example, PREDICT will check for the following:
A product’s inherent risk rating (e.g., raw seafood has a higher inherent risk);
Whether a company (e.g., producer, shipper) or product is subject to an import alert;
The compliance history of the relevant companies (e.g.,producer, shipper, importer) and product, including recalls, inspection results, and results of field exams and sample analyses of previous entries;
The admissibility history of the producer, exporter, importer, and consignee;
Open source intelligence about the product, producer, and foreign locale from which the product originated (e.g., extreme weather conditions such as floods or extreme heat); and
Anomalies in entry data.
The system also looks for patterns and associations. PREDICT then issues a risk score for each entry line. based on the score, the system will either issue a “may proceed” notice or flag the entry as potentially risky and send it to an FDA reviewer for manual processing and possible examination and sampling.
PREDICT will replace the admissibility screening function of OASIS, FDA’s legacy system. Under OASIS, checking of other databases for information about a manufacturer, importer, or product had to be done manually and often relied on the memory of the reviewer.
PREDICT was pilot tested in Los Angeles, Calif., is currently being implemented in New York, and will be deployed in the rest of the country over the next several months with complete installation expected by the end of summer. It is anticipated that PREDICT will significantly increase the number of automated “may proceed” decisions issued for food imports. In pilot testing in Los Angeles, the percentage of “may proceed” decisions jumped to about 60%.
原文链接:http://www.fda.gov/ForIndustry/importProgram%20/ucm172743.htm