Ethnic counts on mortality and census data (mostly) agree for 2001-2004: New Zealand Census-Mortality Study update.

BACKGROUND

The New Zealand Census-Mortality Study (NZCMS) previously demonstrated substantial undercounting of Maori and Pacific deaths on mortality data relative to census data for the 1980s and 1990s. The recent linkage of 2001-04 mortality data to 2001 census data allows us to determine whether any such 'numerator-denominator' bias persists.

METHODS

2001 census anonymously and probabilistically linked to 3 years of subsequent mortality data (82,404 eligible mortality records), allowing a comparison of ethnicity recording.

RESULTS

Using a 'total' definition of ethnicity, there was a close agreement of census and mortality counts: 7419 Maori on the 2001 census compared to 7536 Maori according to mortality data--a census to mortality ratio of 0.98; Pacific--2451 and 2493, ratio 0.98; Asian--1236 and 1215, ratio 1.02; non-Maori non-Pacific non-Asian--73,089 and 72,051, ratio 1.01. Using a 'sole' definition of Maori ethnicity, census counts were only 86% of mortality counts, indicating that mortality data is not recording as many people with two or more ethnic groups as would be expected based on census data. This 'sole' bias was more pronounced in the South Island.

CONCLUSION

There is now little bias in ethnic group counts between census and mortality data for a 'total' definition of ethnic group. Calculations of mortality rates by ethnicity using unlinked census and mortality data and a total definition of ethnicity should be unbiased. These results strongly support using the census definition of ethnicity on all health datasets.