Effect of admission time on mortality in an intensive care unit in Mainland China: a propensity score matching analysis

Abstract

Introduction: The relationship between admission time and intensive care unit (ICU) mortality is inconclusive and influenced by various factors. This study aims to estimate the effect of admission time on ICU outcomes in a tertiary teaching hospital in China by propensity score matching (PSM) and stratified analysis.

Methods: A total of 2,891 consecutive patients were enrolled in this study from 1 January 2009 to 29 December 2011. Multivariate logistic regression and survival analysis were performed in this retrospective study. PSM and stratified analysis were applied for confounding factors, such as Acute Physiology and Chronic Health Evaluation II (APACHE II) score and admission types.

Results: Compared with office hour subgroup (n = 2,716), nighttime (NT, n = 175) subgroup had higher APACHE II scores (14 vs. 8, P < 0.001), prolonged length of stay in the ICU (42 vs. 24 h, P = 0.011), and higher percentages of medical (8.6% vs. 3.3%, P < 0.001) and emergency (59.4% vs. 12.2%, P < 0.001) patients. Moreover, NT admissions were related to higher ICU mortality [odds ratio (OR), 1.725 (95% CI 1.118-2.744), P = 0.01] and elevated mortality risk at 28 days [14.3% vs. 3.2%; OR, 1.920 (95% CI 1.171-3.150), P = 0.01]. PSM showed that admission time remained related to ICU outcome (P = 0.045) and mortality risk at 28 days [OR, 2.187 (95% CI 1.119-4.271), P = 0.022]. However, no mortality difference was found between weekend and workday admissions (P = 0.849), even if weekend admissions were more related to higher APACHE II scores compared with workday admissions.

Conclusions: NT admission was associated with poor ICU outcomes. This finding may be related to shortage of onsite intensivists and qualified residents during NT. The current staffing model and training system should be improved in the future.

Figure 1

Distribution of admissions per hour. APACHE II, Acute Physiology and Chronic Health Evaluation II.

Figure 2

Kaplan–Meier curves of survival differences among ICU patients. ICU survival of office hours (OH) admissions was significantlyhigher than that of nighttime (NT) admissions (a). Propensity score matching (PSM) analysis revealed that NT admissions corresponded to poorer ICU outcomes (b). The dashed lines represented nighttime (NT) admissions, and the solid lines represented office hours (OH) admissions.

Figure 3

Kaplan–Meier curves of ICU survival differences in stratified analysis. In the subgroups of emergency admissions (a), patients with greater Acute Physiology and Chronic Health Evaluation II (APACHE II) scores (>8) (b) and surgical patients (c), mortality was significantly reduced during office hours (OH) compared with nighttime (NT). On the contrary, admission time did not correlate with ICU outcome in nonemergency admissions (d), the subset with low APACHE II scores (≤8) (e) and medical patients (f). LOS_icu, length of stay in the ICU. The dashed lines represented nighttime (NT) admissions, and the solid lines represented office hours (OH) admissions.

Figure 4

Survival analyses for weekend and weekday admissions. No statistically significant difference was found between weekday and weekend admissions (a). Mortality rates between weekend and workday admissions were similar in the subgroups of nighttime (NT) (b) and office hours (OH) (c) admissions. LOS_icu, length of stay in the ICU. The dashed lines represented weekend admissions, and the solid lines represented weekday admissions.