Mortality in people with eating disorders presenting to the health system: A national population-based record linkage study.

BACKGROUND Most previous mortality research in eating disorders involves individuals attending specialist treatment services. Data linkage across jurisdictional health databases at a population level improves the generalisability of findings. AIMS To investigate mortality risk and causes of death for people with an eating disorder across a large geographic region using administrative health datasets. METHOD Using linked hospital, mental health and death records, a retrospective cohort study was conducted including individuals aged 10-59 years who received an eating disorder diagnosis during hospital-based encounters in Australia, over a 10-year period between 2010 and 2019. A contemporary cohort of people accessing community care only were also evaluated. Mortality rates and standardised morality ratios (SMR) compared to the general population were calculated for each state, and by sex and age groups. Cox regression models were used to assess the risk of sociodemographic characteristics on mortality. RESULTS Mortality in people hospitalised with an eating disorder (N=19,697) was more than four times higher than the general population (SMR: 4.54), and highest in people aged 30-39 years (SMR: 13.32). Men hospitalised for eating disorders had a higher risk of death. Mortality rates in anorexia nervosa were not higher than other eating disorder diagnoses. Almost three-quarters of deaths were caused by suicide/self-harm or cardio/respiratory illness. CONCLUSIONS People accessing hospital care with eating disorders in Australia have a higher risk of premature death regardless of age, sex or eating disorder diagnosis. Gender and age group disparities can inform policy and resource allocation and support the development of targeted interventions.