In-hospital mortality among adults with autism spectrum disorder in the United States: A retrospective analysis of US hospital discharge data

A retrospective data analysis using 2004-2014 Healthcare Cost and Utilization Project Nationwide Inpatient Sample was conducted to examine in-hospital mortality among adults with autism spectrum disorders in the United States compared to individuals in the general population. We modeled logistic regressions to compare inpatient hospital mortality between adults with autism spectrum disorders (n = 34,237) and age-matched and sex-matched controls (n = 102,711) in a 1:3 ratio. Adults with autism spectrum disorders had higher odds for inpatient hospital mortality than controls (odds ratio = 1.44, 95% confidence interval: 1.29-1.61, p < 0.001). This risk remained high even after adjustment for age, sex, race/ethnicity, income, number of comorbidities, epilepsy and psychiatric comorbidities, hospital bed size, hospital region, and hospitalization year (odds ratio = 1.51, 95% confidence interval: 1.33-1.72, p < 0.001). Adults with autism spectrum disorders who experienced in-hospital mortality had a higher risk for having 10 out of 27 observed Elixhauser-based medical comorbidities at the time of death, including psychoses, other neurological disorders, diabetes, hypothyroidism, rheumatoid arthritis collagen vascular disease, obesity, weight loss, fluid and electrolyte disorders, deficiency anemias, and paralysis. The results from the interaction of sex and autism spectrum disorders status suggest that women with autism spectrum disorders have almost two times higher odds for in-hospital mortality (odds ratio = 1.95, p < 0.001) than men with autism spectrum disorders. The results from the stratified analysis also showed that women with autism spectrum disorders had 3.17 times higher odds (95% confidence interval: 2.50-4.01, p < 0.001) of in-hospital mortality compared to women from the non-autism spectrum disorders matched control group; this difference persisted even after adjusting for socioeconomic, clinical, and hospital characteristics (odds ratio = 2.75, 95% confidence interval: 2.09-3.64, p < 0.001). Our findings underscore the need for more research to develop better strategies for healthcare and service delivery to people with autism spectrum disorders.