3Residential mobility can have negative impacts on health, with some studies finding that residential 4 mobility can contribute to widening health gradients in the population. However, ethnically 5 differentiated experiences of residential mobility and the relationship with health are neglected in the 6 literature. To examine the relationship between residential mobility, risk of cardiovascular disease 7 (CVD) and ethnicity, we constructed a cohort of 2,077,470 participants aged 30+ resident in New 8Zealand using encrypted National Health Index (eNHI) numbers linked to individual level routinely 9 recorded data. Using binary logistic regression, we model the risk of CVD for the population stratified 10 by ethnic group according to mover status, baseline deprivation and transitions between deprivation 11 statuses. We show that the relationship between residential mobility and CVD varies between ethnic 12 groups and is strongly influenced by the inter-relationship between residential mobility and deprivation 13 mobility. Whilst residential mobility is an important determinant of CVD, much of the variation 14 between ethnic groups is explained by contrasting deprivation experiences. To reduce inequalities in 15 CVD within New Zealand, policies must focus on residentially mobile M ori, Pacific and South Asian 16 populations who already have a heightened risk of CVD living in more deprived areas. 17
Key words 18New Zealand; CVD; Ethnicity; Inequalities; Mobility; Migration; Deprivation; Record Linkage 19
Introduction 20Cardiovascular disease (CVD) and associated morbidities are among the leading causes of global deaths 21 (World Health Organisation, 2014). In New Zealand (NZ) there are marked variations between ethnic 22 groups in the prevalence of CVD (Blakely et al., 2004;Riddell et al., 2007;Jatrana and Blakely, 2008; 23 Kerr et al., 2008; Grey et al., 2010; Mehta et al., 2011; Perumal et al., 2012; Ker et al., 2015; Mehta et 24 al., 2014;Exeter et al., 2015;Wells et al., 2015). Between 1980 and 1999, while all ethnic groups 25 experienced reductions in CVD mortality, M ori and Pacific populations saw markedly smaller 26 2 reductions than non-M ori non-Pacific (nMnP) groups (Blakely et al., 2005). By 2007, these disparities 27 had not disappeared: M ori males and females almost invariably had the highest age-specific prevalence 28 of CVD across all age groups, as well as the highest age-standardised prevalence of CVD (7.41 29 compared to NZ's total population at 4.77, and 5.68 for the Pacific group) (Cheuk Chan et al., 2008). 30Stark differences in risk of CVD and CVD mortality between ethnic groups are not restricted to NZ. 31For example, rates of ischaemic heart disease amongst South Asian males are 30 to 40% higher than 32 rates amongst the UK's general population (Department of Health, 2001). In the US in 2013, Black 33 groups had 30% higher mortality from CVD than Whites, increasing to 113% higher CVD mortality 34 than Asians and Pacific Islanders (Singh et al., 2015). 35Exploring why ethnic in...