The geography of diabetes in Auckland: the influence of the socio-spatial environment on the prevalence of diabetes

Abstract

Introduction Type 2 diabetes is one of the most common non-communicable diseases in New Zealand. Approximately 176,000 people in New Zealand (5% of adults) have been diagnosed with diabetes, and rates have increased over the last 15 years. Diabetes is known to vary by age group, ethnicity, area-deprivation, District Health Board (DHB) and across geographical regions. However few studies from New Zealand and internationally have investigated the influence the socio-spatial environment has on the prevalence of diabetes. Aims This thesis first investigates the prevalence of type 2 diabetes in Auckland Region, by sociodemographic determinants (age, gender, ethnicity, deprivation), and geographically by Census Area Unit (CAU), Territorial Authority (TA) and District Health Board (DHB). Second, this thesis aims to determine the association between diabetes prevalence and two socio-spatial determinants: access to unhealthy foods and recreational facilities in neighbourhoods. Method Using encrypted National Health Identifiers, data from the Primary Health Organisation (PHO) Enrolment Dataset and the National Minimum Dataset (NMDS) were linked to identify patients diagnosed with type 2 diabetes. For inclusion into the study, participants had to be aged 30 years and above, be enrolled in an Auckland Region PHO between 1 July and 30 September 2011. Additionally, patients had to have complete demographic, residential neighbourhood (census Meshblock), and type 2 diabetes status information. Measurement of the socio-spatial environment focussed on neighbourhood accessibility to unhealthy food outlets and recreational facilities at Meshblock level. Locality information of unhealthy food outlets and recreational facilities were obtained from the Zenbu directory of businesses in New Zealand. Accessibility was based on travel distance (metres) from population weighted centroids of Meshblocks to destinations with 1,000m for food outlets and 1,500 for recreational facilities. Levels of accessibility were indexed into quintiles from quintile 1 (low access) to quintile 5 (high access). Poisson regression analyses were conducted to investigate a patient’s likelihood of having diabetes, controlling for demographic and socio-spatial determinants and results were presented as Incidence Rate Ratios (IRRs) and 95% confidence intervals (CI). Findings This thesis used a study population of 746,568 participants, from which 63,761 patients were identified as having type 2 diabetes (9% prevalence for the Auckland region). After controlling for socio-demographic and socio-spatial factors, females in the study population were 13% less likely to have diabetes (IRR 0.87; 95% CI 0.86-0.88) than males. The likelihood of diabetes increased with age, peaking among those aged 75-79 years. Compared with the people aged 55-64 years, participants aged 30-34 years were 82% less likely (IRR 0.18; 95% CI 0.17-0.19), to have diabetes while participants aged 75-79 years were 61% more likely to have diabetes (IRR 1.61; 95% CI 1.56-1.65). Indian (IRR 3.48; 95% CI 3.39-3.58), Pacific Island (IRR 3.31; 95% CI 3.23-3.39) and Māori (IRR 2.37; 95% CI 2.29-2.45) participants were significantly more likely than the New Zealand/Other ethnic group to have diabetes. There were increasing IRRs across the New Zealand Index of Deprivation (NZDep 2006) quintiles, with the most deprived areas (quintile 5) having nearly twice the likelihood of diabetes (IRR 1.92; 95% CI 1.86-1.99) than residents living in least deprived areas. Across DHBs, residents of Counties Manukau DHB portrayed a 10% higher likelihood for diabetes (IRR 1.09; 95%CI 1.06-1.12) than residents of Auckland DHB, whereas no statistically significant differences were observed for Waitemata DHB. A similar burden of diabetes was evident at the TA level. Controlling for socio-demographic and socio-spatial factors, neighbourhoods with high accessibility to unhealthy foods were associated with 30% lower likelihood for diabetes (IRR 0.70; 95%CI 0.64- 0.77), however there was no statistically significant association between accessibility to recreational facilities and diabetes prevalence. Conclusion This thesis found increased risk for diabetes among participants aged 65 years and older, Maori, Pacific and Indian groups and for people living in the most deprived neighbourhoods of Auckland. Additionally, the thesis observed differences in diabetes rates, across geographical regions at the level of MB, CAU, TA ad DHB. This was the first study in New Zealand to investigate the association between the socio-spatial environment and diabetes. Surprisingly, there was a negative association between diabetes and increased neighbourhood accessibility to unhealthy food outlets. The need for a routinely updated is overdue in New Zealand and this thesis calls for urgent need to develop a routine diabetes register in New Zealand. Other key recommendations proposed in this thesis include improving geographic identifiers in health datasets and improving resources for physical activity opportunities in areas of high need.