The history and future of record linkage in the ONS longitudinal study

Article type: Research Article

Authors: Smith, Jillian

Affiliations: Office for National Statistics, 1 Drummond Gate, London, SW1V 2QQ, UK

Note: [] Paper presented at the Joint UN-ECE/Eurostat Work Session on Registers and Administrative Records for Social and Demographic Statistics, Geneva, 1–3 March 1999.

Abstract: The Office for National Statistics Longitudinal Study (LS) is a large dynamic record linkage database, with regular updates of information from censuses and vital events, such as births and deaths, covering one percent of the population of England and Wales. It now encompasses over twenty-five years' data and is supported within a ‘safe setting’ for analysis on a wide range of subjects. The LS does not carry identification information such as name and address. It uses the National Health Service Central Register (NHSCR) to provide the means of linkage between censuses and vital events and the LS database. Events are added to the LS, either through flagging mechanisms at NHSCR or by trawling for LS events in the national event registers. They are added to the database regularly to maintain a dynamic dataset for analysis. Every ten years, since 1971, a large project is mounted to add the census sample into the LS database. To date, each census sample has been linked manually using the NHSCR, but plans for the 2001 Census link are to complete it using computer matching techniques. Initial tests have shown that this will be an efficient method of linkage. Keystones to the success of the LS are the up-to-date register systems at NHSCR and the extensive security systems which preserve confidentiality. The study now has a long history of linkage and confidentiality techniques.

DOI: 10.3233/SJU-1999-162-308

Journal: Statistical Journal of the United Nations Economic Commission for Europe, vol. 16, no. 2-3, pp. 197-205, 1999