Geographic analysis of infant mortality in New Zealand, 1995–2008: an ethnicity perspective

Objective : To detect spatial clusters of high infant mortality rates in New Zealand for Māori and non‐Māori populations and verify if these clusters are stable over a certain time period (1995–2008) and similar between the two populations. Method : We applied the Kulldorff's spatial scan statistics on data collected by New Zealand Ministry of Health (1995 to 2008) at the territorial local authorities (TLA) level. Kappa coefficient was used to assess the concordance between clusters obtained for Māori and non‐Māori populations. T‐test analyses were conducted to identify associations between spatial clusters and two predictors (population density and deprivation score). Results : There are some significant spatial clusters of infant mortality in New Zealand for both Māori and Non‐Māori. The concordance of the cluster locations between the two populations is strong (kappa=0.77). Unsurprisingly, infant mortality clusters for both Māori and Non‐Māori are associated with the deprivation score. The population density predictor is only significantly and positively associated with clusters obtained for the non‐Māori population. After controlling for deprivation the presence of spatial clusters is all but eliminated. Conclusions : Infant mortality patterns are geographically similar for both Māori and Non‐Māori. However, there are differences geographically between the two populations after accounting for deprivation. Implications : Health services that can affect infant mortality should be aware of the geographical differences across NZ. Deprivation is an important factor in explaining infant mortality rates and policies that ameliorate its effects should be pursued, as it is the major determinant of the geographical pattern of infant mortality in NZ.