Record Details

Title Method to assess identifiability in electronic data files
Author Howe, HL
Secondary Authors Lake AJ, Shen T
Publication Type (Help) article
Journal Am J Epidemiol
Month Mar 1
Year 2007
Pages 597-601
Volume 165
Number 5
Publisher
Address
Note
URL http://aje.oxfordjournals.org/content/165/5/597.long
PubMed ID 17182982
NCI Id
EPub Date 2006 Dec 20
Citation Howe HL, Lake AJ, Shen T. Method to assess identifiability in electronic data files. Am J Epidemiol. 2007 Mar 1;165(5):597-601. EPub 2006 Dec 20. PMID 17182982. [http://aje.oxfordjournals.org/content/165/5/597.long.]
  BibTex    


Abstract

The authors developed the Record Uniqueness (RU) software program to assess electronic data files for risk of confidentiality breach based on unique combinations of key variables. The underlying methodology utilized by the RU program generates a frequency distribution for every variable selected for analysis and for all combinations of the variables selected. In addition, the program provides the regression coefficient that designates the relative contribution of each variable to the unique records on the data file. The authors used RU to evaluate a North American Association of Central Cancer Registries research data set with 4.67 million cases from 34 population-based cancer registries for 1995–2001. To illustrate the process and utility of RU, they describe the evaluation process of the confidentiality risk of adding a county-based socioeconomic measure to the research file. The RU method enables one to be assured of record confidentiality, provides flexibility to adjust record uniqueness thresholds for different users or purposes of data release, and facilitates good stewardship of confidential data balanced with maximum use and release of information for research. RU is a useful data tool that can quantify the risk of confidentiality breach of electronic health databases, including reidentifiability of cases through triangulation of information or linkage with other electronic databases.



Keywords

Keyword
assess
data
identifiability
method