Background:
Ventilator-associated pneumonia (VAP) represents the highest burden among all healthcare-associated infections (HAIs), with a particularly high rate in patients in neurosurgical ICUs. Numerous VAP risk factors have been identified to provide a basis for preventive measures. However, the impact of individual factors on the risk of VAP is unclear. The goal of this study was to evaluate the dynamics of various VAP risk factors given the continuously declining prevalence of VAP in our neurosurgical ICU. Methods: This prospective cohort unit-based study included neurosurgical patients who stayed in the ICU >48 consecutive hours in 2011 through 2018. The infection prevention and control (IPC) program was implemented in 2010 and underwent changes to adopt best practices over time. We used a 2008 CDC definition for VAP. The dynamics of VAP risk factors was considered a time series and was checked for stationarity using theAugmented Dickey-Fuller test (ADF) test. The data were censored when a risk factor was present during and after VAP episodes. Results: In total, 2,957 ICU patients were included in the study, 476 of whom had VAP. Average annual prevalence of VAP decreased from 15.8 per 100 ICU patients in 2011 to 9.5 per 100 ICU patients in 2018 (Welch t test P value = 7.7e-16). The fitted linear model showed negative slope (Fig. 1). During a study period we observed substantial changes in some risk factors and no changes in others. Namely, we detected a decrease in the use of anxiolytics and antibiotics, decreased days on mechanical ventilation, and a lower rate of intestinal dysfunction, all of which were nonstationary processes with a declining trend (ADF testP > .05) (Fig. 2). However, there were no changes over time in such factors as average age, comorbidity index, level of consciousness, gender, and proportion of patients with brain trauma (Fig. 2). Conclusions: Our evidence-based IPC program was effective in lowering the prevalence of VAP and demonstrated which individual measures contributed to this improvement. By following the dynamics of known VAP risk factors over time, we found that their association with declining VAP prevalence varies significantly. Intervention-related factors (ie, use of antibiotics, anxiolytics and mechanical ventilation, and a rate of intestinal dysfunction) demonstrated significant reduction, and patient-related factors (ie, age, sex, comorbidity, etc) remained unchanged. Thus, according to the discriminative model, the intervention-related factors contributed more to the overall risk of VAP than did patient-related factors, and their reduction was associated with a decrease in VAP prevalence in our neurosurgical ICU.Funding: NoneDisclosures: None
Retraction notice to “Testing for multiple bubbles in bitcoin markets: A generalized sup ADF test” [Japan World Econ. 46 (2018) 56–63]