Survival rates following medical intensive care unit admission from 2003 to 2013: An observational study based on a representative population-based sample cohort of Korean patients

The decision as to whether patients should be admitted to a medical intensive care unit (ICU), in the absence of information concerning survival rates or prognostic factors in survival, is often challenging. We analyzed survival trends in relation to hospital discharge and examined patient and hospital characteristics associated with survival following ICU care, using a sample of nationwide claims data in Korea from 2002 through 2013. The Korean government implements a compulsory social insurance program that covers the country's entire population, and the Korean National Health Insurance Service-National Sample Cohort (NHIS-NSC) data from 2002 based on this program were used for this study. The NHIS-NSC is a stratified random sample of 1,025,340 subjects selected from around 46 million Koreans. We evaluated annual survival trends using the Kaplan-Meier test. Analyses of the relationship between survival and patient and hospital characteristics were performed using Cox regression analyses. Employing a multivariate model, variables were selected using the forward selection method to consider the multicollinearity of variables. A total of 32,553 patients admitted to an ICU between 2002 and 2013 were identified among the eligible beneficiaries. The number of patients who had histories of ICU admission steadily increased throughout the study period, and patients older than 80 years constituted a progressively increasing proportion of ICU admissions, from 7.3% in 2002 to 16.9% in 2007 to 23.1% in 2013. The mean number of mechanical equipment items applied consistently increased, while no difference was observed in the trend for overall 1-year survival in patients following ICU treatment across the study period: the 1-year survival rate ranged from 66.7% (year 2003) to 64.2% (year 2010). Advanced age, cancer, renal failure, pneumonia, and influenza were all associated with heightened risk of mortality within 1 year. Our results should prove useful to older patients and their clinicians in their decisions regarding whether to seek ICU care, with the goals of improving the end-of life care and optimizing resource utilization.