大量历史数据在生化项目自动审核测试验证中的应用

The Usage of Large Amounts of Historical Data in the Test and Verification of Auto Verification

目的:探讨大量历史数据在生化项目自动审核规则测试和自动审核系统验证中的应用。方法:收集四川大学华西医院2014年1月-2015年10月经人工方式审核发出报告的所有患者及体检人群生化项目数据,共计样本数2 693 550个,测试数37 453 050个,应用编制的工具软件,对设定的生化项目自动审核规则进行测试及自动审核系统验证。结果:可快速得出单个项目和样本自动审核通过率与不通过率;样本不通过原因和分布;规则测试与系统验证数据可回顾、对比。结论:大量历史数据在生化项目自动审核规则测试和自动审核系统验证中,具有测试验证结果呈现快速、可靠、测试验证有效,操作方便,节约资源与成本,有利于自动审核系统建立和持续改进,是值得推荐的规则测试和系统验证方法。

Objective： To explore the usage of large amounts of historical data in the test and verification of auto verification. Methods： Collect all of the biochemical data which were manually reviewed during Jan. 2014 to Oct. 2015. There are 2,693,550 samples and 37,453,050 tests, and through the software, we tested, modified and verificated the auto verification rules. Results： After testing and verification the auto verification rules, we found the large amounts historical data could get the rate of pass and do not pass rapidly, the reason and distribution of the do not pass items, and filer the data, which is upper or lower than the reference or analysis range. Conclusion： The large amounts of historical data in the test and verification of auto verification is convenient, reliable, rapid and reproducible. It is effective to save the labor cost and improve the auto verification system continuous. The usage of historical data is recommended method during the test and verification of auto verification.