Anonymous record linkage using respondent-generated identification codes - a tool for health promotion research

Issue addressed: The problem of needing to match data in evaluation studies, while retaining anonymity of the respondent. Methods: Attempts were made to match data collected from approximately 1300 high school students in two anonymous surveys before and after an intervention by using a respondent-generated personal identification code. Results: Using components with the least missing data, 68.4 per cent of records were able to be matched exactly. The combination of components which enabled the highest percentage of exact matches were day of birth, month of birth, sex, initial of respondent's first name, and suburb/town of residence. Conclusions: This technique of record linkage can be readily applied in health promotion research, with the benefits of increased statistical power. So what? Using respondent- generated identification codes to link records can improve evaluations of health promotion interventions while preserving respondent confidentiality, and may be especially useful when working in sensitive areas. (author abstract)