Sharing patient data: understanding anonymisation.
Sharing patient data: understanding anonymisation.
复制标题
共享患者数据：理解匿名化。
DOI：
10.1136/bmj.k2700
复制
发表时间：
2018
期刊：
影响因子：
--
通讯作者：
Affleck P
中科院分区：
文献类型：
--
作者：
Affleck P
英国医学杂志报道了英国公共卫生部的澄清，即它与William E Wecker Associates共享的数据是匿名的。[1]但是，将这些数据与澳大利亚重新识别的例子进行比较是误导性的。英国公共卫生部共享了一个汇总统计表，而澳大利亚的联邦卫生部则公布了个人数据。澳大利亚的数据被“去识别化”，这意味着强标识符被删除，但一些人可以通过将数据与其他信息相结合来识别。鉴于这些数据是纵向处方记录，对患者保密性的威胁是显而易见的。gov.uk/dataset/4d69e312-da1f-4a7d-819c-8eff414ae9f7/epidemiology-of-lung-cancer-tumours-in-england-2009-to-2013），并且仅包含八个数据列（肿瘤计数、性别、年龄段、死亡原因、种族、诊断年份、形态学和死亡年份）。值得注意的是，其中一些计数是个位数，这意味着只有一个人符合所有这些条件。从理论上讲，这意味着如果你已经掌握了大部分信息，这个人就可以被识别出来。此外，如果你知道一个人的性别，大致年龄，死亡原因，种族，诊断年份和癌症的形态分类，你也许可以找出他们的死亡年份。据推测，英国公共卫生部并不认为这是对患者保密性的威胁。
The BMJ reported on Public Health England’s clarification that the data it shared with William E Wecker Associates were anonymised. 1 But comparing these data with the Australian example of re-identification was misleading. Public Health England shared a table of aggregated counts, whereas the Federal Department of Health in Australia released individual data. The Australian data were “de-identified,” meaning that strong identifiers were removed but some people could be identified by combining the data with other information. Given that these data were longitudinal prescribing records, the threat to patient confidentiality is clear.The Public Health England dataset is publicly available (https://data. gov. uk/dataset/4d69e312-da1f-4a7d-819c-8eff414ae9f7/epidemiology-of-lung-cancer-tumours-in-england-2009-to-2013) and only contains eight data columns (count of tumours, sex, age band, cause of death, ethnicity, year of diagnosis, morphology, and year of death). Notably, some of the counts are in single figures, meaning that only one person met all of those conditions. Theoretically this means that the person could be identified if you possessed most of the information already. Also, if you knew a person’s sex, approximate age, cause of death, ethnicity, diagnosis year, and morphology classification of their cancer, you might be able to find out their year of death. Presumably, Public Health England did not judge this to be a threat to patient confidentiality.