Geographical and social inequalities in survival after incident acute myocardial infarction in Denmark: A spatial epidemiological cohort study based on nationwide register data

Abstract

Introduction Survival after acute myocardial infarction (AMI) is associated with receiving adequate and often time-dependent treatment as well as sociodemographic characteristics of the patient. Access to adequate AMI treatment might depend on where you live as distance to the nearest hospital differs substantially across the country. Moreover, where you live is to some degree influenced by your socioeconomic position and economical resources. The interplay between the geographical and social inequalities in survival after an AMI is not clearly understood and it remains unclear whether geographical patterns in survival can be explained by sociodemographic factors. This study aimed to examine whether the geographical patterns in survival after incident AMI can be explained by individual- and neighborhood-level sociodemographic characteristics. Methods The study population consisted of patients (≥ 30 years) with an incident AMI between 1 January 2005 and 31 December 2014 registered in the National Patient Register or the Danish Register of Causes of Death. The population was followed for a maximum of 365 days or until emigration, study end or death, whichever came first. The year 2005 was used as run-in period to ensure adequate results estimates across calendar years. Hence, the study period was 1 January 2006 to 31 December 2014. Poisson regression of incidence rates of all-cause mortality with a geographical component as a random effect was performed in R by use of the Bayesian Integrated Nested Laplace Approximations (INLA) method. The analysis included the individual-level variables age, sex, ethnicity, calendar year, cohabitation status, disposable household income, and educational level and the neighborhood-level variables population density, socioeconomic position, and ethnical composition of the neighborhood. Results The study population consisted of 95,274 patients with incident AMI contributing with 62,8210 person-years at-risk between 2006 and 2014. During the study period 28,194 deaths occurred within 365 days. Crude mortality rates varied substantially across the country as well as across neighborhood and individual-level characteristics, e.g., 53.2/100 person-years died within 365 days after an AMI in neighborhoods with low socioeconomic position, whereas this number was 36.7/100 person-years in high socioeconomic position neighborhoods. Results from the model only including calendar year and the geographically structured random effect component showed clear geographical patterns in survival after an AMI with especially the northern part of Jutland being a high-risk area. After additionally including sociodemographic characteristics of the population and neighborhood in the model the variation in survival after an AMI decreased across the country, but high-risk areas were still observed. Results on fixed effects from the adjusted model showed that persons with low educational level, low-income level and persons living in neighborhoods with low socioeconomic position had an increased incidence rate ratio (IRR) of dying within 365 days after AMI compared to the most educated and affluent population groups [IRRs (95% confidence intervals) of 1.29 (1.23–1.36), 1.26 (1.20–1.32) and 1.11 (1.05–1.17), respectively]. Furthermore, risk of dying within 365 days after an AMI decreased throughout the study period. Results from sensitivity analyses showed that the geographical patterns were more pronounced in out-of-hospital than in-hospital deaths. Conclusions Evident geographical and social inequalities in survival after an incident AMI exist in Denmark. Sociodemographic characteristics at the individual- and neighborhood-level explained part of the geographical variation in survival after an AMI; however, marked geographical patterns in survival after AMI persisted. Further studies are needed to explain the remaining geographical inequalities in survival after an AMI, especially with focus on out-of-hospital deaths.