Does geography explain ethnic inequalities in health in New Zealand?

Objective: To quantify the potential contribution of inter-district relative to intradistrict variation to the Maori disparity in life expectancy in 2000-02, by counterfactual modelling. Setting, data sources and methods: The setting was New Zealand's 21 health districts (District Health Boards, DHBs). All data (population estimates and life expectancy estimates) were sourced from Statistics New Zealand and relate to the 2000-02 period. Maori life expectancy (nationally) was recalculated under the counterfactual that Maori life expectancy in each DHB did not differ from total population life expectancy in the corresponding DHB (so eliminating intra-district variation). The difference between the observed total population and counterfactual Maori life expectancies therefore represents the contribution of inter-district variation to the Maori life expectancy disparity. Results: Observed total population and Maori life expectancies at birth in 2000-02, pooling sexes, were 78.7 and 71.1 years respectively, giving a total disparity of 7.6 years. Under the counterfactual, Maori life expectancy increased to 78.4 years (and total population life expectancy to 79.0 years). Inter-district variation was therefore estimated to potentially contribute only 0.6 years or 8% to the total Maori disparity. Allowing for imprecision, inter-district variation almost certainly accounts for less than 10.5% of the total disparity. Conclusion: Inter-district or geographic variation makes only a small contribution to the total Maori disparity in life expectancy. Adjustment or stanclardisation for district is not necessary when comparing Maori and non-Maori health outcomes. If the policy goal is to reduce ethnic inequalities in health, then the focus of policy (e.g. funding formulae) needs to be on factors directly linked to ethnicity, rather than on geographic variations in health and health care that have an impact on all ethnic groups more-or-less alike.