Moss S M, Smith J A, Nicholas D S
South and West Cancer Intelligence Unit, Highcroft, Winchester.
Public Health. 1997 Mar;111(2):101-6. doi: 10.1016/s0033-3506(97)90010-x.
Electronic linkage between pathology data sources and other information systems has not realised its full potential benefits due to the poor quality of histopathology coding. This study showed that 38% of a sample of 158 pathology reports were coded accurately. Of the incorrectly coded reports, 25% had the potential to distort published cancer incidence figures. The incidence figures of the most common cancers are less likely to be affected by coding errors. Areas in which all errors, both topographical and morphological, could have significant impact include examining resource allocation at directorate level and adjusting outcome indicators for casemix. This study concludes that electronic linkage between histopathology systems and cancer registries is not sufficient to improve the quality of registration data. As cancer registries become more dependent on computerised information provided through hospital information systems, registries need to be aware of poor quality data and put in place appropriate quality assurance measures specifically tailored to support electronic cancer registration. Purchasers and providers need to be aware that incomparable datasets could be produced if other bodies use computerised pathology datasets without first validating the data.
由于组织病理学编码质量不佳，病理数据源与其他信息系统之间的电子链接尚未充分实现其潜在效益。本研究表明，在158份病理报告样本中，38%被准确编码。在编码错误的报告中，25%有可能扭曲已公布的癌症发病率数据。最常见癌症的发病率数据受编码错误影响的可能性较小。所有错误（包括地形学和形态学错误）可能产生重大影响的领域包括在科室层面审查资源分配以及调整病例组合的结果指标。本研究得出结论，组织病理学系统与癌症登记处之间的电子链接不足以提高登记数据的质量。随着癌症登记处越来越依赖通过医院信息系统提供的计算机化信息，登记处需要意识到数据质量不佳的问题，并制定专门为支持电子癌症登记而量身定制的适当质量保证措施。购买者和提供者需要意识到，如果其他机构在未首先验证数据的情况下使用计算机化病理数据集，可能会产生不可比的数据集。
