Association between evening admissions and higher mortality rates in the pediatric intensive care unit

Background. Previous research investigating the relationship between the time of admission and mortality rates has yielded inconsistent results and has not been conducted in the pediatric intensive care unit (PICU) patient population. Objective. To determine whether an association between the time of admission (weekday versus weekend and daytime versus evening) and the risk of death exists among pediatric patients included in a cohort of children admitted to a national sample of PICUs. Design/Methods. We analyzed retrospectively a cohort of consecutive admissions to 15 PICUs included in the Pediatric Intensive Care Unit Evaluations database. The odds of death were analyzed by using mixed-effects, multivariate, logistic regression, with clustering at the hospital level. The primary independent variables were admission to the PICU on a weekend and admission to the PICU during evening hours. The severity of illness was adjusted by using the Pediatric Risk of Mortality III probability of death score. Patients. All 20 547 emergency PICU admissions made between May 1995 and December 2001 were included in the analyses. Main Outcome Measures. The primary outcome was death within 48 hours after admission to the PICU. Results. Pediatric patients admitted to the PICU during evening hours had higher odds of death (odds ratio [OR]: 1.28; 95% confidence interval [CI]: 1.00-1.62) than did those admitted during daytime hours. Subgroup analyses revealed higher odds of death among patients admitted with shock (OR: 4.09; 95% CI: 1.65-10.1), with congenital cardiovascular disease (OR: 3.90; 95% CI: 1.37-11.1), or after cardiac arrest (OR: 1.80; 95% CI: 1.04-3.13). There was no association between mortality rates and the day of admission (weekend admissions versus weekday admissions). Conclusions. An increased risk of death exists for some pediatric patients admitted to the PICU during evening hours. It remains necessary to determine whether this finding results from differences in the structure of care, processes of care, or both.