All-cause, premature, and cardiovascular death attributable to socioeconomic and ethnic disparities among New Zealanders with type 1 diabetes 1994–2019: a multi-linked population-based cohort study

Background New Zealand (NZ) research into type 1 diabetes mellitus (T1DM) mortality can inform policy and future research. In this study we aimed to quantify the magnitude to which ethnicity and socioeconomic disparities influenced mortality at the population level among people with Type 1 diabetes (T1DM) in Auckland, New Zealand (NZ). Methods The cohort data were derived from the primary care diabetes audit program the Diabetes Care Support Service (DCSS), and linked with national primary care, pharmaceutical claims, hospitalisation, and death registration databases. People with T1DM enrolled in DCSS between 1994–2018 were included. All-cause, premature, and cardiovascular mortalities were estimated by Poisson regression models with adjustment for population-level confounders. The mortality rates ratio (MRR) was standardized against the DCSS type 2 diabetes population. Mortality rates were compared by ethnic group (NZ European (NZE) and non-NZE) and socioeconomic deprivation quintile. The population attributable fraction (PAF) was estimated for ethnic and socioeconomic disparities by Cox regression adjusting for demographic, lifestyle, and clinical covariates. The adjusted slope index inequality (SII) and relative index of inequality (RII) were used to measure the socioeconomic disparity in mortalities. Results Overall, 2395 people with T1DM (median age 34.6 years; 45% female; 69% NZE) were enrolled, among whom the all-cause, premature and CVD mortalities were 6.69 (95% confidence interval: 5.93–7.53), 3.30 (2.77–3.90) and 1.77 (1.39–2.23) per 1,000 person-years over 25 years. The overall MRR was 0.39 (0.34–0.45), 0.65 (0.52–0.80), and 0.31 (0.24–0.41) for all-cause, premature and CVD mortality, respectively. PAF attributable to ethnicity disparity was not significantly different for mortality. The adjusted PAF indicated that 25.74 (0.84–44.39)% of all-cause mortality, 25.88 (0.69–44.69)% of premature mortality, 55.89 (1.20–80.31)% of CVD mortality could be attributed to socioeconomic inequality. The SII was 8.04 (6.30–9.78), 4.81 (3.60–6.02), 2.70 (1.82–3.59) per 1,000 person-years and RII was 2.20 (1.94–2.46), 2.46 (2.09–2.82), and 2.53 (2.03–3.03) for all-cause, premature and CVD mortality, respectively. Conclusions Our results suggest that socioeconomic disparities were responsible for a substantial proportion of all-cause, premature and CVD mortality in people with T1DM in Auckland, NZ. Reducing socioeconomic barriers to management and self-management would likely improve clinical outcomes.