作者: Khaled El Emam , Saeed Samet , Jun Hu , Liam Peyton , Craig Earle
DOI: 10.1371/JOURNAL.PONE.0039915
关键词: Protocol (science) 、 Computer security 、 Cryptography 、 Privacy law 、 Confidence interval 、 Hash function 、 Contingency table 、 Linked data 、 Context (language use) 、 Medicine 、 General Biochemistry, Genetics and Molecular Biology 、 General Agricultural and Biological Sciences 、 General Medicine
摘要: Introduction In order to monitor the effectiveness of HPV vaccination in Canada linkage multiple data registries may be required. These not always managed by same organization and, furthermore, privacy legislation or practices restrict any linkages records that can actually done among registries. The objective this study was develop a secure protocol for linking from different and allow on-going monitoring vaccine effectiveness. Methods A protocol, using commutative hash functions multi-party computation techniques developed. This allows exact matching statistics on linked while meeting five practical requirements ensure patient confidentiality privacy. considered were: odds ratio its confidence interval, chi-square test, relative risk interval. Additional contingency tables, such as other measures association, added principles presented. time performance evaluated. Results The has acceptable scales linearly with size set table. worse case up 100,000 patients returned each query 16 cell table is less than 4 hours basic statistics, best under 3 hours. Discussion computationally been demonstrated context initiative impact assessment. generalized surveillance conditions, diseases, programs.