Predictors of successful discharge from intensive care units in older adults aged 80 years or older: A population-based study

Background: We determined the long-term outcomes of adult intensive care unit (ICU) patients and investigated the predictors of successful discharge for ≥6 months among older adults aged ≥80 years in Taiwan. Objectives: To identify the predictors of ICU admission in patients aged ≥80 years and the predictors of and optimal cutoff predictive discharge score (PDS) for ICU successful discharge in patients aged ≥80 years. Design: A population-based retrospective cohort study. Settings and participants: Medical records of 282,269 individuals aged ≥80 years collected from the Taiwan National Health Insurance Research Database from 2001 to 2013. Methods: Demographic and clinical parameters, Charlson's comorbidity index (CCI), hospital type, and post-discharge outcomes of ICU patients aged ≥80 years were obtained from their medical records. Multivariable logistic regression was used to analyze and identify the predictors of successful discharge and treatments received by critically ill patients aged ≥80 years admitted to the ICU. The optimal cutoff PDS for successful discharge in older adults were calculated by Youden Index. Results 65,756 ICU admissions were documented, of which 21% (n = 13,825) were for adults aged ≥80 years. The successful discharge rate among ICU patients aged ≥80 years (57.2%) was significantly lower than that among those aged 18–64 and 65–79 years (81.7% and 71.5%, respectively). Multivariable logistic regression analyses revealed the following predictors of successful discharge for ≥6 months after ICU admission in adults aged ≥80 years: younger age (adjusted odds ratio [OR] = 0.95, 95% confidence interval [CI] = 0.94–0.96), shorter ICU length of stay (adjusted OR = 0.90, 95% CI = 0.88–0.92), lower CCI (adjusted OR = 0.92, 95% CI = 0.90–0.93), and no life-sustaining treatments received (Cardiopulmonary resuscitation: adjusted OR = 0.75, 95% CI = 0.68–0.84; mechanical ventilation: adjusted OR = 0.63, 95% CI = 0.57–0.71; use of inotropic agents: adjusted OR = 0.37, 95% CI = 0.34–0.41). The optimal cutoff PDS in older adults was 6 (area under the receiver operating characteristic curve = 0.73, 95% CI = 0.72–0.74). Conclusion: This is the first population-based study investigating the post-discharge outcomes of ICU patients aged ≥80 years. Advanced age was a predictive factor of unsuccessful discharge from the ICU. Nevertheless, more than half of this vulnerable population survived for at least 6 months after discharge. Therefore, age should not be the sole criterion when considering ICU admission and deciding curative treatments for critically ill older adults. Comprehensive assessment and effective communication with patients and their families are also crucial in clinical decision-making for critically ill older adults.