Mortality by ethnic group to 2006: Is extending censusmortality linkage robust?

Abstract

Objective: To update trends in mortality by ethnic group from the New Zealand Census-Mortality Study (NZCMS), by additionally linking 2004-06 mortality records to the 2001 Census. To investigate possible bias from this extended linkage, especially for Pacific and Asian people who emigrate more frequently. Methods: Anonymous and probabilistic record linkage of 2004-06 mortality records with the 2001 Census was undertaken. Age-standardised 1-74 year old mortality rates by sex and age group, and for all-cause and selected causes of death, were calculated using the direct method for first 30 months post 2001 Census (2001-03) and second 30 months (2003-06). Results: Observed all-cause mortality r..