A program for sustained improvement in preventing ventilator associated pneumonia in an intensive care setting

Setting and study design

An interrupted time series study was conducted in a 38-bed medical-surgical intensive care unit (ICU) of a tertiary care, private hospital in São Paulo, Brazil. This is an open staffing model ICU where approximately 2,200 patients are admitted annually. This study was a quality improvement study that was approved by the Institutional Review Board of Hospital Israelita Albert Einstein (IRB). The requirement for informed consent was waived by our IRB in accordance with the Code of Federal Regulations and the Privacy Rule. This project was carried out after our previously published study from April 2007 to September 2008 [4]. Herein we report our observations for the period from October 2008 to December 2010 to evaluate whether the sustained implementation of the VAP bundle in our ICU could effectively reduce the incidence of VAP. All these hospital epidemiology data was analysed anonymously.

The VAP bundle included elevation of the head of the bed (HOB) (30–45 degrees); daily “sedation vacations” and assessment of readiness to extubate; peptic ulcer disease prophylaxis; and deep venous thrombosis/pulmonary thromboembolism (DVT/PE) prophylaxis for all ICU patients requiring mechanical ventilation. This ventilator bundle was monitored each weekday by an ICU nurse. She intervened in this process while performance monitoring was taking place at the bedside if non-compliance with an element of the bundle was detected (e.g., sedation was not stopped). We also intervened in other CDC evidence-based practices for prevention of ventilator associated pneumonia [3] including: 1) no routine changing of humidified ventilator circuits, 2) periodically draining and discarding condensate collecting in the ventilator tubing and, 3) changing the heat-and-moisture exchangers (HMEs) when they showed mechanical malfunction or became visibly soiled. These CDC process measures were audited twice yearly in a small sample of mechanically ventilated patients at random intervals.

Other interventions to control VAP in the ICU were implemented in October 2007 when oral decontamination with chlorhexidine 0.12% was introduced for all mechanically ventilated ICU patients [4]. In February 2008, the continuous aspiration of subglottic secretions (CASS) endotracheal tube was implemented for patients requiring mechanical ventilation and expected to require ventilation for longer than 24 hours [4].

Previously [4] we had compared the VAP bundle alone (phase 1), with the VAP bundle + oral decontamination with chlorhexidine 0.12% (phase 2), and the VAP bundle + oral decontamination with chlorhexidine 0.12% + continuous aspiration of subglottic secretions (CASS) endotracheal tube (phase 3). We then decided to analyze our performance after almost two years of all these interventions (including the VAP bundle) to determine whether this was a sustainable program for controlling ventilator associated pneumonia. We decided to extend data collection in phase 3 (VAP bundle + oral decontamination with chlorhexidine 0.12% + continuous aspiration of subglottic secretions (CASS) endotracheal tube) to evaluate this assumption (Figure 1). In summary, phases 1, 2 and 3 (from April 2007 to September 2008) are a consequence of our previous publication [4]. We extended data collection in this present manuscript (from October 2008 to December 2010) in phase 3. We provided monthly feedback on compliance with the bundle components to the ICU team (doctors, nurses and respiratory therapists). We also displayed posters in the ICU with bar charts showing compliance with the recommended procedures. These posters also showed the VAP rate as determined in surveys conducted by the Department of Infection Control and Hospital Epidemiology.

Definitions

VAP surveillance was performed by trained infection control specialists using the US Center for Disease Control and Prevention/National Healthcare Safety Network (CDC/NHSN) definition [12] in an independence way from the treating ICU team, the incidence of VAP was expressed as cases of VAP per 1,000 ventilator-days and the incidence of ventilator associated tracheobronchitis (VAT) was expressed as cases of VAT per 1,000 ventilator-days.

VAP was defined as the sum of the clinical criteria as described the presence of fever (temperature >38°C), new or increased sputum production, in combination with radiologic evidence of a new or progressive pulmonary infiltrate, leukocytosis, a suggestive Gram’s stain, and grow of bacteria (not necessarily) in cultures of sputum, tracheal aspirate, pleural fluid, bronchoalveolar lavage (BAL), or blood [12]. Per the CDC/NHSN definition, microbiological data are not necessary for the diagnosis.

VAT was defined as the presence of fever (temperature >38°C), new or increased sputum production, a microbiologically positive respiratory sample, and the absence of pulmonary infiltrates on chest radiography.

Microbiological methods

All isolates were identified by manual or automated methods and confirmed using the Vitek 2 system (bioMerieux Vitek, Inc., Hazelwood, MO).

Statistical analysis

The variables of interest were those that indicated compliance with the VAP prevention measures. We used segmented regression analysis of interrupted time series [13] to assess the changes in VAP before and after implementation of the ventilator bundle, oral decontamination with chlorhexidine 0.12%, and CASS endotracheal tube for patients requiring mechanical ventilation, according to the interventional phases (Figure 1).

We adjusted a segmented regression model that allowed us to analyze a reduction (or an increase) in VAP rate at each study phase separately: (1) ventilator bundle only, April 2007 to October 2007; (2) ventilator bundle + chlorhexidine, November 2007 to February 2008; (3) ventilator bundle + chlorhexidine + CASS endotracheal tube, March 2008 to December 2010 (Figure 1).

The intercept and slope are the two parameters which define each segment of a time series. The intercept is the value of the series at the beginning of a given time interval, the slope is the change of the measure (VAP rate) over a certain period (e.g., a month). A change in slope (β) is defined by an increase or decrease in the slope of the time step after the intervention, compared with the time step preceding the intervention. It is important to mention that this is not the same as constructing three models of simple linear regression, because the third partition parameters depend on the previous partitions’ parameters.

All tests of statistical significance were 2-sided with a significance level set at 0.05. All the data analyses were performed using SPSS 16.0 and SAS 9.1; SAS Institute Inc, Cary, NC, USA.

Compliance with process measures in each phase

In 2010, the analysis included 2,260 HOB elevation observations (91% compliance), 390 ventilator circuits without changes (99% compliance) and 390 observations of HMEs changes (94% compliance). Daily sedation vacations, gastric prophylaxis opportunities and DVT/PE prevention opportunities had 2,486 observations with 91% of compliance in all measurements (Table 1).

Incidence density of VAP and in-hospital mortality of VAP patients

The incidence density of VAP per 1,000 ventilator days in the ICU was 1.3 in 2009 (10,889 patient-days) and in 2010 the incidence was 2 (11,095 patients-days). The incidence density of VAT per 1,000 ventilator days in the ICU was 1.0 (10,889 patient-days) in 2009 and in 2010 the incidence was 2 (11,095 patient-days).

Mechanical ventilation days, ICU length of stay, ventilator utilization ratio, number of VAPs, number of VATs, ventilator-days, ventilator-free days, in-hospital mortality of VAP patients and in-hospital mortality ICU patients are shown in Table 1.

Getting to zero VAP for one or more months has occurred since 2009 when there was greater than 95% compliance with the ventilator bundle, oral decontamination with chlorhexidine 0.12% and continuous aspiration of subglottic secretions (CASS) (Figure 2). In addition to this we continued to evaluate ventilator circuits without changes, HMEs changes and ventilator-circuit-tubing condensate.

Segmented regression analysis (Figure 3) showed a statistically significant increase in VAP rate (β11 = +2.59; p < 0.001) in the first segment (ventilator bundle). The transition from the first segment (ventilator bundle) to the second segment (ventilator bundle + chlorhexidine) showed a significant decrease in VAP rate (β20 = −11.24; p < 0.001). The slope (β21) in the second segment was negative, indicating a reduction in VAP rate upon implementation of oral decontamination with chlorhexidine (β21 = −2.30 with p = 0.272). The transition from the second segment (ventilator bundle + chlorhexidine) to the third segment (ventilator bundle + chlorhexidine + CASS endotracheal tube) was not significant in VAP rate (β30 = −2.67 with p = 0.682). In the third segment the slope (β31) was practically zero, indicating that there was no reduction in VAP rate, which was maintained along the segment (β31 = 0.03 with p = 0.610).

Microbiological features

Pseudomonas aeruginosa accounted for over 40% of the gram-negative pathogens. The most prevalent pathogens overall were Pseudomonas aeruginosa and Acinetobacter baumannii. The majority of the pathogens were identified by tracheal aspiration. Only 25% of VAP patients (1/4) were investigate by BAL (Table 2).

The mortality rate was 100% for the patients with VAP in 2009 and 83.3% in 2010.

Mechanical ventilation

The total time of mechanical ventilation was 6,052 days, with a utilization rate of 28% in 2009 and 27% in 2010 (Table 1). As seen in Figure 4, our mechanical ventilation utilization rates have been reduced since 2000, with a 17% drop from 2000 to 2010.