Abstract
Background: Ethnic inequities in lung cancer in New Zealand are stark, with Māori (New Zealand's indigenous population) rates more than three times higher than European rates. New Zealand has a broad range of data that can be used to understand and monitor health inequities but limited understanding of how the use of different datasets and methods affects reported rates of health indicators. This study examined whether measured ethnic differences in lung cancer registration rates change depending on the population used and the source of ethnicity data.
Methods: We used data from the Integrated Data Infrastructure (IDI), a collection of linked, deidentified, New Zealand whole-population administrative and survey datasets. Lung cancer registrations were identified through the New Zealand Cancer Registry. We calculated age-standardised lung cancer registration rates (for Māori and sole European ethnic groups) and relative risks (Māori compared to sole European). These were compared for: six population selection methods (holding the source of ethnicity data constant); four sources of ethnicity data (holding the population constant); and ten combinations of population selection method and ethnicity data source.
Results: Population selection method and source of ethnicity data each had independent impacts on lung cancer registration rates and the Māori: sole European relative risk. Changing the population selection method had a larger impact than changing the source of ethnicity data. Different combinations of population selection method and ethnicity data source resulted in different age-standardised lung cancer registration rates, with 24.4% difference between the lowest and highest Māori rates and 11.7% between the lowest and highest sole European rates. The Māori: sole European relative risk varied by13.4% between the lowest and highest estimates.
Conclusions: Population selection method and source of ethnicity data have a measurable impact on lung cancer rates, especially for Māori, and on measures of inequity. Changes to these methods may disrupt the time series, obscure trends in inequities, and disrupt Māori rights to monitor the Crown. It is therefore critical that any methodological changes are undertaken with guidance from Māori. Transparency and consistency in IDI methods, and collecting high quality ethnicity data, are also priorities.