American Journal of Respiratory and Critical Care Medicine

Rationale: The CDC introduced ventilator-associated event (VAE) definitions in January 2013. Little is known about VAE prevention. We hypothesized that daily, coordinated spontaneous awakening trials (SATs) and spontaneous breathing trials (SBTs) might prevent VAEs.

Objectives: To assess the preventability of VAEs.

Methods: We nested a multicenter quality improvement collaborative within a prospective study of VAE surveillance among 20 intensive care units between November 2011 and May 2013. Twelve units joined the collaborative and implemented an opt-out protocol for nurses and respiratory therapists to perform paired daily SATs and SBTs. The remaining eight units conducted surveillance alone. We measured temporal trends in VAEs using generalized mixed effects regression models adjusted for patient-level unit, age, sex, reason for intubation, Sequential Organ Failure Assessment score, and comorbidity index.

Measurements and Main Results: We tracked 5,164 consecutive episodes of mechanical ventilation: 3,425 in collaborative units and 1,739 in surveillance-only units. Within collaborative units, significant increases in SATs, SBTs, and percentage of SBTs performed without sedation were mirrored by significant decreases in duration of mechanical ventilation and hospital length-of-stay. There was no change in VAE risk per ventilator day but significant decreases in VAE risk per episode of mechanical ventilation (odds ratio [OR], 0.63; 95% confidence interval [CI], 0.42–0.97) and infection-related ventilator-associated complications (OR, 0.35; 95% CI, 0.17–0.71) but not pneumonias (OR, 0.51; 95% CI, 0.19–1.3). Within surveillance-only units, there were no significant changes in SAT, SBT, or VAE rates.

Conclusions: Enhanced performance of paired, daily SATs and SBTs is associated with lower VAE rates.

Clinical trial registered with www.clinicaltrials.gov (NCT 01583413).

Scientific Knowledge on the Subject

The Centers for Disease Control and Prevention released novel surveillance definitions for ventilator-associated events (VAEs) in early 2013. VAEs affect about 5% of ventilated patients and are strongly associated with increased duration of mechanical ventilation and increased hospital mortality risk. VAEs have been proposed as quality metrics but little is known about their prevention.

What This Study Adds to the Field

This is the first prospective study of the preventability of VAEs. Twelve intensive care units participating in a longitudinal, multicenter quality improvement collaborative focused on increasing the implementation of paired, spontaneous awakening trials and spontaneous breathing trials were able to reduce VAE rates by 37%.

Mechanically ventilated patients are at risk for multiple complications of critical care including pneumonia, acute respiratory distress syndrome (ARDS), pulmonary edema, thromboembolism, delirium, and atelectasis. Traditionally, surveillance for complications of mechanical ventilation has been limited to ventilator-associated pneumonia (VAP). This metric has been criticized, however, because it is subjective, labor intensive, prone to bias, and accounts for only a small fraction of intensive care unit (ICU) morbidity (13). In response to these concerns, the CDC collaborated with professional societies to develop new surveillance targets called “ventilator-associated events” (VAEs) (4, 5). The CDC’s National Healthcare Safety Network replaced their VAP definitions with VAE definitions in early 2013. Over 1,500 U.S. hospitals currently report VAE data to the CDC.

VAE definitions were designed to identify episodes of sustained respiratory deterioration after a period of stability or improvement. The VAE framework includes a hierarchy of surveillance targets (Figure 1). The primary target is called a “ventilator-associated condition” (VAC) and is defined as greater than or equal to 2 days of increased ventilator settings after greater than or equal to 2 days of stable or decreasing settings (4). There are then secondary criteria to flag the subset of VACs that might be infection-related ventilator-associated complications (IVAC) and those that might be pneumonias (4).

VAC and IVAC have been proposed as potential metrics for benchmarking and pay-for-performance programs. Most organizations have deferred using the definitions for these purposes, however, because there are very few data at present about whether, how, and to what extent VAEs are preventable. Indeed, some have hypothesized that VAEs are simply surveillance markers for preexisting morbid illnesses rather than preventable complications (6, 7). Data on the preventability of VAEs are therefore urgently needed.

We prospectively evaluated the preventability of VAEs. We reasoned that the best way to prevent VAEs would be to decrease exposure to mechanical ventilation by speeding time to extubation. Minimizing sedation and regularly assessing patients’ readiness for extubation have repeatedly been shown to speed liberation from mechanical ventilation (8, 9). Minimizing sedation may also have collateral benefits because high levels of sedation are independently associated with multiple infectious and noninfectious complications (10, 11). We therefore organized a multicenter collaborative to test the preventability of VAEs by enhancing the consistency, reliability, and coordination of daily spontaneous awakening trials (SATs) and spontaneous breathing trials (SBTs). Some of the results of this study have been previously reported in the form of an abstract (12).

Overview

We nested the collaborative within a prospective surveillance study of VAE epidemiology. The surveillance study included 20 ICUs affiliated with 13 academic and community hospitals. Twelve ICUs elected to participate in the collaborative (“collaborative units”) and eight did not (“surveillance-only units”). All units, however, collected the same data on all patients using the same definitions and entered the data into a common web-based data-entry system. We were therefore able to follow changes in VAE incidence, SAT and SBT performance rates, and other outcomes using comparable methods in both collaborative and surveillance-only units. We did not directly compare performance and outcome rates between collaborative participants and nonparticipants, however, because participation in the collaborative was not randomized and there were substantial differences between the units that elected to join the collaborative and those that did not (see the online supplement).

Intervention

We developed a consensus protocol for paired daily SATs and SBTs modeled after the work of Girard and colleagues (see the online supplement) (8). The protocol shifted responsibility for initiating SATs and SBTs from physicians to nurses and respiratory therapists using an opt-out model. Nurses and respiratory therapists screened all patients daily for eligibility for SATs and SBTs, and when appropriate, initiated one or both interventions. SATs entailed completely stopping all sedatives and, in patients without active pain, all narcotics. If a patient passed the SAT, the protocol encouraged keeping the patient off sedatives and/or narcotics; if a patient failed the SAT then sedatives and/or narcotics were restarted at half the preceding dose. SBTs entailed lowering positive end-expiratory pressure to less than or equal to 5–8 cm H2O with either continuous positive airway pressure or pressure support ventilation of less than or equal to 5 cm H2O for up to 2 hours. The protocol encouraged promptly extubating patients who passed SBTs. If the SBT was unsuccessful, therapists returned patients to their prior ventilator settings. Nurses and respiratory therapists were encouraged to coordinate efforts to perform SBTs off sedatives. Each hospital’s Institutional Review Board approved the study. The protocol was deemed quality improvement and therefore did not require informed consent.

The CDC Prevention Epicenters’ Wake Up and Breathe Collaborative

We organized a quality improvement collaborative for the 12 intervention units using the Institute for Healthcare Improvement’s All Teach All Learn framework (13). Each participating unit designated a physician, nurse, and respiratory therapist to serve as clinical champions and collaborative liaisons. We held an in-person kick-off meeting for champions at the CDC on April 24, 2012 to engage participants, review the consensus protocol, set unit-specific goals, and begin planning interventions. We taught Plan-Do-Study-Act cycles to encourage rapid prototyping of small tests of change. Clinical champions completed monthly reports detailing successes and challenges from the preceding month and setting goals for the coming month. We held monthly web conferences to review progress, share successes, and discuss challenges. We created a listserv to facilitate discussion between monthly meetings. We shared detailed performance data with each unit each month. We provided deidentified comparative statistics to allow units to compare themselves with peers. We held a second in-person meeting on October 30, 2012 to review progress, consolidate gains, and set further goals. Representatives from collaborative units helped develop the protocol and sought local approval in the months before the official kick-off of the care improvement collaborative.

We encouraged frontline clinicians to integrate documentation of SAT and SBT screenings and outcomes into existing processes, such as paper or electronic flow charts, daily ventilator bundle audits, and/or daily plan sheets. Research assistants collected basic demographic and clinical data for every patient ventilated for more than 1 calendar day including age, sex, unit, location of intubation, reason for intubation, initial Sequential Organ Failure Assessment (SOFA) score, admission and discharge dates, and vital status on discharge. They also documented daily minimum positive end-expiratory pressure, daily minimum fraction of inspired oxygen, temperature, and white blood cell count. Antibiotic exposures, pulmonary Gram stains, and culture results were collected on the subset of patients that met criteria for VAC. These data were all entered into a centralized data repository using a secure, web-based interface. Units provided ICD9 and DRG codes for each patient at the end of the study; we used these to derive comorbidities using the Elixhauser method (14, 15). Data collection for most variables began in November 2011 and continued through May 2013. Data collection for SAT performance rates began in January 2012 and continued through May 2013.

The primary outcome of the study was VAE risk per episode of mechanical ventilation (this is synonymous with the VAC risk per episode). We selected episodes rather than ventilator days as the denominator because the study intervention specifically aimed to reduce ventilator days; a disproportionate decrease in ventilator days relative to VAEs could create a misleading impression of static or increasing VAE rates. We did a secondary analysis of VAE risk per ventilator day for purposes of comparison, however, because infection control programs have traditionally used ventilator days as their denominator. Additional secondary outcomes included IVAC, possible and probable pneumonia, mean duration of mechanical ventilation, intensive care length-of-stay, hospital length-of-stay, hospital mortality, self-extubations, and reintubations within 24 hours. VAEs and episodes of mechanical ventilation were defined using CDC criteria and identified electronically using computer algorithms applied to manually collected data (16). The CDC defines an episode of mechanical ventilation as continual exposure to a ventilator for at least some part of consecutive calendar days; if a patient is disconnected from a ventilator for more than 1 calendar day then any subsequent mechanical ventilation is defined as a new episode.

Analysis

We assessed for cumulative changes across time in VAE risk, SAT and SBT rates, and other outcomes on a per-episode basis, using generalized mixed effects models to account for within-unit correlations. We used the Bernoulli distribution for binary outcomes and negative binomial distributions for counts. We originally intended to incorporate an inflection point corresponding to the collaborative kick-off meeting at the CDC on April 24, 2012. Inspection of the data, however, revealed substantial and significant increases in SAT, SBT, and proportion of SBTs performed with sedatives off in the months before the kick-off meeting (Figure 2), presumably attributable to unit champions’ participation in developing the protocol and garnering local support before the kick-off meeting. We therefore elected to fit linear trends for all outcomes across the entire study period without an inflection point. We adjusted each model for patient-level age, sex, reason for intubation, SOFA score, and Elixhauser index whenever possible. We used multiple imputation with chained equations to estimate missing SOFA covariates for individuals with incomplete baseline data (e.g., bilirubin not measured) (17, 18).

We created additional models to assess the relationship between monthly unit-level SAT and SBT performance rates (predictors) and VAE risk, duration of mechanical ventilation, length-of-stay, and mortality (outcomes). We used logistic or negative binomial generalized mixed effect regression models as appropriate and adjusted for the same patient-level covariates as in our primary analysis. All analyses were executed using SAS v9.3 (SAS Institute, Cary, NC).

Role of the Funding Source

CDC scientists participated in the design, conduct, and interpretation of the study and helped edit the final manuscript.

The collaborative included four medical, two surgical, two cardiac, and four mixed medical-surgical units affiliated with seven hospitals. Collaborative units provided care for 3,425 consecutive episodes of mechanical ventilation during the study period. Patients’ characteristics by episode of mechanical ventilation are summarized in Table 1. Surveillance-only ICUs included one surgical, two cardiac, and five mixed medical-surgical units affiliated with six hospitals. Surveillance-only units provided care for 1,739 consecutive episodes of mechanical ventilation.

Table 1. Patient Characteristics by Episode among the 12 Intensive Care Units Participating in the CDC Epicenters’ Wake Up and Breathe Collaborative

 Collaborative Units (n = 3,425 Episodes)
Mean age (SD)62.8 (37)
Male1,942 (57%)
Mean SOFA score (SD)*9.3 (3.7)
Comorbidities 
 Congestive heart failure793 (23%)
 Chronic lung disease1,037 (30%)
 Diabetes1,117 (33%)
 Kidney disease883 (26%)
 Liver disease297 (8.6%)
 Cancer430 (13%)
 Peripheral vascular disease502 (15%)
Location of intubation 
 Prehospital emergency medical services101 (2.9%)
 Emergency department595 (17%)
 General medicine ward146 (4.3%)
 Operating room or recovery room625 (18%)
 Intensive care unit1,286 (38%)
 Outside facility435 (13%)
 Other237 (6.9%)
Reason for intubation 
 Altered level of consciousness379 (11%)
 Obstructive lung disease150 (4.4%)
 Cardiac arrest256 (7.5%)
 Community-acquired pneumonia173 (5.1%)
 Hospital-acquired pneumonia94 (2.7%)
 Pulmonary edema79 (2.3%)
 Sepsis157 (4.6%)
 Surgery675 (20%)
 Respiratory distress of unknown etiology1,462 (43%)
Intensive care unit type 
 Medical, four units1,047 (31%)
 Surgical, two units358 (10%)
 Cardiac, two units343 (10%)
 Mixed, four units1,677 (49%)
Days of mechanical ventilation 
 Total22,991
 Mean (SD)6.7 (9.2)
 Median (IQR)4 (3–8)
Intensive care length-of-stay in days 
 Mean (SD)11 (14)
 Median (IQR)8 (5–14)
Hospital length-of-stay in days 
 Mean (SD)22 (25)
 Median (IQR)15 (8–26)
Hospital mortality958 (28%)
Ventilator-associated events 
 VAC293 (8.5%)
 IVAC100 (2.9%)
 Possible pneumonia33 (1.0%)
 Probable pneumonia17 (0.5%)
 Possible or probable pneumonia50 (1.5%)

Definition of abbreviations: IQR = interquartile range; IVAC = infection-related ventilator-associated complications; SOFA = Sequential Organ Failure Assessment; VAC = ventilator-associated conditions.

*All 12 variables necessary to calculate SOFA scores were available for 74.4% of episodes. One of the 12 variables was missing for 24.4% of episodes, two variables were missing for 3.0% of episodes, and three variables were missing for 0.7% of episodes. We used multiple imputation to estimate values for missing variables.

Includes episodes that meet criteria for IVAC and possible or probable pneumonia.

Collaborative units significantly increased their frequencies of SATs, SBTs, and percentage of SBTs performed off sedatives (Table 2, Figure 2). SAT performance rates increased from 14% of days where indicated to 77% of days where indicated, corresponding to an increase from 5% of ventilator days to 21% of ventilator days (mean increase, 0.9%/mo; 95% confidence interval [CI], 0.7–1.1%). SBT performance rates increased from 49 to 75% of days where indicated, corresponding to 37% of ventilator days in the first month of the collaborative and 35% of ventilator days in the last month of the collaborative. The percentage of SBTs performed with sedatives off increased from 6.1 to 87% of SBTs (mean increase, 4.8%/mo; 95% CI, 4.6–4.9%).

Table 2. Changes in SAT and SBT Performance Rates among Collaborative Participants between November 2011 and May 2013

OutcomeFirst Month* (95% CI)Last Month* (95% CI)Average Change per Month (95% CI)Cumulative Change (95% CI)P Value
SATs performed as percent of days with SATs indicated14% (7.1 to 26)77% (61 to 87)+3.6% (3.3 to 4.0)+63% (57 to 69)<0.0001
SATs performed as a percent of ventilator days5.1% (3.5 to 7.5)21% (16 to 27)+0.9% (0.7 to 1.2)+16% (11 to 20)<0.0001
SBTs performed as percent of days with SBTs indicated49% (35 to 63)75% (64 to 84)+1.4% (1.0 to 1.8)+26% (19 to 34)<0.0001
SBTs performed as percent of ventilator days37% (31 to 43)35% (30 to 41)−0.08% (−0.010 to 0.05)−1.5% (−2.0 to −0.9)<0.0001
SBTs performed with sedatives off as percent of all SBTs6.1% (3.9 to 9.4)87% (81 to 92)+4.8% (4.6 to 4.9)+81% (79 to 84)<0.0001

Definition of abbreviations: CI = confidence interval; SAT = spontaneous awakening trial; SBT = spontaneous breathing trial.

*First month with available data was November 2011 for SBTs and January 2012 for SATs. Last month was May 2013.

Improvements in SAT and SBT performance rates were paralleled by significant decreases in VAE rates (Figure 3). The VAE rate (this is synonymous with the total VAC rate) went from 9.7 events per 100 episodes of mechanical ventilation in November 2011 to 5.2 events per 100 episodes in May 2013 (adjusted odds ratio [OR], 0.63; 95% CI, 0.42–0.97). The IVAC rate dropped from 3.5 to 0.52 events per 100 episodes (adjusted OR, 0.35; 95% CI, 0.17–0.71). The possible or probable pneumonia rate decreased from 0.88 to 0.52 per 100 episodes; however, this change was not statistically significant (adjusted OR, 0.51; 95% CI, 0.19–1.3). There was no change in VAE/VAC or IVAC risk when using ventilator days rather than episodes as the denominator (see Table E1 in the online supplement).

Decreases in VAE risk per episode of mechanical ventilation were similar when limiting the analysis to patients ventilated for greater than or equal to 4 days (adjusted OR for VAC, 0.71, 95% CI, 0.46–1.09; adjusted OR for IVAC, 0.36; 95% CI, 0.18–0.73; adjusted OR for pneumonia, 0.56, 95% CI, 0.21–1.48). There was no significant change over time in the monthly fraction of patients ventilated for less than 4 days.

Collaborative units observed improvements in other outcomes (Figure 4). Mean duration of mechanical ventilation dropped by 2.4 days (95% CI, 1.7–3.1 d), ICU length-of-stay by 3.0 days (95% CI, 1.6–4.3 d), and hospital length-of-stay by 6.3 days (95% CI, 4.0–8.6 d). There was no change in hospital mortality (OR, 1.1; 95% CI, 0.81–1.4). There was a significant increase in self-extubations per episode of mechanical ventilation (OR, 2.1; 95% CI, 1.1–3.9) but no change in reintubations within 24 hours (OR, 0.96; 95% CI, 0.66–1.4) as depicted in Figure E1. There was also no change in the frequency of reintubations within 2–7 days (data not shown).

On subgroup analysis, effect estimates were broadly similar for all unit types for all outcomes, although CIs were wide and sometimes no longer significant (see Table E2). Improvements were seen in six of the seven hospitals and 8 of the 12 units participating in the collaborative.

The VAE rate in December 2011 was strikingly high compared with all other months. On sensitivity analysis excluding this month, the decrease in VAC was no longer significant (adjusted OR, 0.76; 95% CI, 0.49–1.20) but the decrease in IVAC persisted (adjusted OR, 0.42; 95% CI, 0.20–0.90).

There were significant associations between monthly, unit-level SAT and SBT performance rates and length-of-stay, mortality, and VAEs (Table 3). Higher SAT performance rates were associated with significantly fewer VACs (OR, 0.20; 95% CI, 0.06–0.64), IVACs (OR, 0.04; 95% CI, 0.005–0.32), possible or probable pneumonias (OR, 0.03; 95% CI, 0.003–0.40), ventilator days (OR, 0.32; 95% CI, 0.25–0.43), ICU days (OR, 0.51; 95% CI, 0.40–0.66), hospital days (OR, 0.44; 95% CI, 0.34–0.57), and lower hospital mortality rates (OR, 0.32; 95% CI, 0.14–0.74). Monthly unit-level SBT performance rates were significantly associated with fewer VACs (OR, 0.15; 95% CI, 0.06–0.38) and ventilator days (OR, 0.79; 95% CI, 0.63–0.99). The percentage of SBTs performed off sedatives was associated with fewer ventilator days (OR, 0.89; 95% CI, 0.80–0.99) and hospital days (OR, 0.82; 95% CI, 0.74–0.91) but did not correlate with VAE rates.

Table 3. Multivariable Analysis of the Association between Unit-Level Monthly Performance Measures and Outcomes*

 Monthly Unit Level SAT Rate Odds Ratio/Relative Risk (95% CI)P ValueMonthly Unit Level SBT Rate Odds Ratio/Relative Risk (95% CI)P ValueMonthly Unit Level SBTs Done with Sedatives Off Odds Ratio/Relative Risk (95% CI)P Value
Ventilator-associated events      
 Ventilator-associated conditions0.20 (0.06–0.64)0.0070.15 (0.06–0.38)<0.0011.1 (0.68–1.8)0.71
 Infection-related ventilator-associated complications0.04 (0.005–0.32)0.0030.35 (0.08–1.61)0.181.2 (0.55–2.5)0.68
 Possible or probable pneumonias0.03 (0.003–0.40)0.0070.24 (0.03–1.92)0.181.2 (0.4–3.5)0.73
Lengths of stay and mortality      
 Duration of mechanical ventilation, d0.32 (0.25–0.43)<0.00010.79 (0.63–0.99)0.050.89 (0.80–0.99)0.03
 Intensive care length-of-stay, d0.51 (0.40–0.66)<0.00011.01 (0.81–1.3)0.920.91 (0.82–1.0)0.06
 Hospital length-of-stay, d0.44 (0.34–0.57)<0.00011.04 (0.8–1.3)0.770.82 (0.74–0.91)<0.0001
 Hospital mortality0.32 (0.14–0.74)0.0081.1 (0.60–1.9)0.810.85 (0.62–1.2)0.31

Definition of abbreviations: CI = confidence interval; SAT = spontaneous awakening trial; SBT = spontaneous breathing trial.

Analyses adjusted for patient level indicators including unit, age, sex, reason for intubation, and Sequential Organ Failure Assessment score. Monthly SAT and SBT rates calculated as number of SATs or SBTs performed divided by total ventilator days.

*Estimates for dichotomous outcomes derived from binomial distributions and presented as odds ratios; estimates for continuous outcomes derived from negative binomial distributions and presented as relative risks.

Among surveillance-only units, there was no significant change in SAT performance rates (9.1% of ventilator days in the first month vs. 11% of ventilator days in the last month; mean increase 0.1%/mo; 95% CI, −0.1 to +0.3%). There was a modest increase in SBT performance rates (from 22% to 24% of ventilator days; mean increase, 0.1%/mo; 95% CI, +0.03 to +0.2%) and a modest increase in the percentage of SBTs performed without sedation (from 39% to 67%; mean increase, 1.6%/mo; 95% CI, 1.3–1.9%). There was no change, however, in the overall risk of VAEs, IVAC, or possible and probable pneumonias. Clinical characteristics, SAT and SBT performance rates, and outcomes for patients in the surveillance-only units are provided in Tables E3–E5.

The CDC developed VAE definitions to replace VAP as a possible quality measure for ventilated patients. Multiple studies have shown that VAEs strongly predict poor outcomes (1926). Until now, however, there has been little evidence that VAEs are preventable. Our study suggests that increasing the frequency of paired, daily SATs and SBTs can significantly lower VAE rates. These gains were further associated with shorter ventilator, ICU, and hospital stays.

Our findings are concordant with evolving critical care practice standards. The Society of Critical Care Medicine’s Pain, Analgesia, and Delirium Guidelines encourage minimizing sedation and early extubation (27). We show that in addition to decreasing duration of mechanical ventilation and length-of-stay, these interventions can also lower VAE rates. VAE surveillance may therefore be a useful metric to help hospitals to monitor the impact of their efforts to improve sedation and extubation practices.

We were able to significantly increase SAT and SBT rates in collaborative units despite the fact that all participants already had policies encouraging daily SATs and/or SBTs in place before joining the collaborative. National and statewide surveys affirm that SAT and SBT performance rates are often low despite policies encouraging their performance (2831). This underscores the fact that there can often be substantial mismatch between official policies and actual performance rates. We attribute the collaborative units’ increases in SAT and SBT performance to three factors: (1) rigorously measuring and reporting actual performance, (2) continually feeding back local and comparative performance rates, and (3) working assiduously to change providers’ perceptions about sedation. Collaborative discussions were often dedicated to challenging assumptions about required levels of sedations, encouraging providers to attempt SATs and SBTs on a broader array of patients, and exploring alternative strategies to calm patients without administering additional sedatives.

The decrease in VAE risk per episode of mechanical ventilation but not per ventilator day suggests that the primary mechanism of the intervention was reducing total duration of exposure to mechanical ventilation rather than decreasing risk of complications on any given day of mechanical ventilation. Reducing duration of mechanical ventilation can lower VAE rates in two ways: by decreasing time at risk for complications and by increasing the fraction of patients ventilated for less than 4 days (because VAE criteria are only applicable to patients ventilated for ≥4 d). On subgroup analysis, we found that the intervention decreased VAE rates among patients ventilated for greater than or equal to 4 days and there was no significant change in the fraction of patients ventilated for less than 4 days. This suggests that additional strategies to speed liberation from mechanical ventilation, such as sedation protocols, automated weaning algorithms, early mobility, and conservative fluid management, may also lower VAE rates (3235).

These observations beg the question of whether surveillance for VAEs adds anything over surveillance for mean duration of mechanical ventilation alone. Duration of mechanical ventilation is attractive insofar as it is simple, ubiquitous, and clinically intuitive. Duration of mechanical ventilation is a very coarse metric for quality assessment, however, because it is highly dependent on case mix and does not indicate whether and when any given patient suffered a complication. VAE surveillance, by contrast, identifies a concrete complication that can serve as a focus for root cause analyses to identify possible opportunities to improve care. The evidence that most VAEs are complications includes the following: (1) by definition, VAEs mark trajectory changes in mechanical ventilation from stability or improvement to deterioration (4); (2) VAEs have been consistently associated with increased duration of mechanical ventilation and higher hospital mortality risk compared with matched control subjects (1926); (3) clinical reviews of VAEs suggest that most are caused by pneumonia, hypervolemia, atelectasis, and/or ARDS (19, 21, 25); and (4) this study and others suggest that improvements in care can lower VAE rates (22, 36).

An insight from our study is that “episodes” might be preferable to “ventilator days” as the denominator for reporting VAE rates when the care improvement strategy is directed toward decreasing ventilator days. The large and consistent decreases in ventilator, ICU, and hospital lengths-of-stay suggest that our collaborative did improve patient care yet we only observed significant decreases in VAE rates calculated as events per episode not as events per ventilator day. This finding is consistent with the literature on preventing catheter-associated urinary tract infections where interventions designed to decrease the number of catheter days can lead to paradoxical increases in infection rates expressed as the risk per catheter day (37).

Similarly, measuring SAT and SBT performance rates as a percentage of ventilator days may be more informative than measuring SAT and SBT performance rates as a percentage of “days where indicated.” We observed substantial increases in SAT and SBT performance rates measured as the percentage of days where indicated but much more modest increases (or in the case of SBT, no increase) when measuring performance as the percentage of ventilator days. Reporting SAT and SBT rates as the percentage of days where indicated can be misleading because it is possible to improve the rate without necessarily changing care by assigning contraindications more liberally or by improving documentation of contraindications. Ventilator days, by contrast, are objective.

Limitations of our study include the lack of randomized control units and early adoption of the collaborative intervention before the anticipated kick-off date. It is therefore possible that the observed decreases in VAE rates were caused by preexisting secular trends or secondary to initiatives unrelated to our intervention. Our study includes five pieces of evidence that help mitigate these concerns: (1) analyses were adjusted using detailed patient level data including age, sex, reason for intubation, comorbidity index, and SOFA score; (2) a multivariable analysis affirmed the strong protective association between SATs and VAEs and other outcomes; (3) parallel data from units outside the collaborative showed no changes in VAE rates; (4) VAEs were identified objectively using electronic data; and (5) surveys of units before and after the collaborative did not reveal any new quality improvement interventions. Furthermore, even if the decrease in VAE rates was caused by something other than the collaborative intervention, the primary goal of this study was to assess the preventability of VAEs rather than to test the effectiveness of SATs and SBTs per se. The observed decrease in VAE rates suggests VAE rates can be lowered.

In summary, we demonstrate that better performance of coordinated daily SATs and SBTs can decrease VAE rates and improve patients’ outcomes. Future studies should explore whether additional interventions targeting the conditions most frequently associated with VAEs can decrease VAE rates further. The conditions most commonly associated with VAEs include pneumonia, pulmonary edema, atelectasis, and ARDS (19, 21, 25). Potential interventions could therefore include head-of-bed elevation, early mobility, low tidal volume ventilation, conservative fluid management, and conservative transfusion thresholds (33, 36, 3840). Combining these interventions into a new ventilator bundle and using VAE surveillance to track the bundle’s impact has further potential to improve outcomes for ventilated patients.

The authors thank all the physicians, nurses, and respiratory therapists from the hospitals that participated in this study (Alamance Hospital, Duke Raleigh Hospital, Duke University Hospital, Durham Regional Hospital, Durham Select Hospital, Durham Veterans Affairs Medical Center, Hospital of the University of Pennsylvania, Lawrence-Memorial Hospital, Melrose-Wakefield Hospital, Missouri Baptist Medical Center, North Shore Medical Center [Salem], North Shore Medical Center [Union], and Stroger Hospital).

1. Klein Klouwenberg PM, Ong DS, Bos LD, de Beer FM, van Hooijdonk RT, Huson MA, Straat M, van Vught LA, Wieske L, Horn J, et al. Interobserver agreement of Centers for Disease Control and Prevention criteria for classifying infections in critically ill patients. Crit Care Med 2013;41:23732378.
2. Stevens JP, Kachniarz B, Wright SB, Gillis J, Talmor D, Clardy P, Howell MD. When policy gets it right: variability in U.S. Hospitals’ diagnosis of ventilator-associated pneumonia. Crit Care Med 2014;42:497503.
3. Melsen WG, Rovers MM, Groenwold RH, Bergmans DC, Camus C, Bauer TT, Hanisch EW, Klarin B, Koeman M, Krueger WA, et al. Attributable mortality of ventilator-associated pneumonia: a meta-analysis of individual patient data from randomised prevention studies. Lancet Infect Dis 2013;13:665671.
4. Magill SS, Klompas M, Balk R, Burns SM, Deutschman CS, Diekema D, Fridkin S, Greene L, Guh A, Gutterman D, et al. Developing a new, national approach to surveillance for ventilator-associated events. Crit Care Med 2013;41:24672475.
5. Klompas M. Complications of mechanical ventilation—the CDC’s new surveillance paradigm. N Engl J Med 2013;368:14721475.
6. Shorr AF, Zilberberg MD. Nature (and the ICU) abhors a VACuum. Chest 2012;142:13651366.
7. Lilly CM, Ellison RT III. Quality measures for critically ill patients: where does ventilator-associated condition fit in? Chest 2013;144:14291430.
8. Girard TD, Kress JP, Fuchs BD, Thomason JW, Schweickert WD, Pun BT, Taichman DB, Dunn JG, Pohlman AS, Kinniry PA, et al. Efficacy and safety of a paired sedation and ventilator weaning protocol for mechanically ventilated patients in intensive care (Awakening and Breathing Controlled trial): a randomised controlled trial. Lancet 2008;371:126134.
9. Strøm T, Martinussen T, Toft P. A protocol of no sedation for critically ill patients receiving mechanical ventilation: a randomised trial. Lancet 2010;375:475480.
10. Jackson DL, Proudfoot CW, Cann KF, Walsh T. A systematic review of the impact of sedation practice in the ICU on resource use, costs and patient safety. Crit Care 2010;14:R59.
11. Nseir S, Makris D, Mathieu D, Durocher A, Marquette CH. Intensive care unit-acquired infection as a side effect of sedation. Crit Care 2010;14:R30.
12. Klompas M, Andersen DJ, Trick W, Babcock H, Sinkowitz-Cochran RL, Ely EW, Jernigan JA, Magill S, Lyles RD, O’Neill C, et al. The preventability of ventilator-associated events: the CDC Prevention Epicenters' Wake Up and Breathe Collaborative. Presented at IDWeek 2014, October 8–12, 2014, Philadelphia, PA. Abstract 46339.
13. Institute for Healthcare Improvement. The Breathrough Series: IHI’s collaborative model for achieving breakthrough improvement. Diabetes Spectrum 2004;17:97101.
14. Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care 1998;36:827.
15. van Walraven C, Austin PC, Jennings A, Quan H, Forster AJ. A modification of the Elixhauser comorbidity measures into a point system for hospital death using administrative data. Med Care 2009;47:626633.
16. Centers for Disease Control and Prevention. Ventilator-associated event protocol. 2014 [accessed 2014 March 25]. Available from: http://www.cdc.gov/nhsn/acute-care-hospital/vae/index.html
17. van Buuren S, Boshuizen HC, Knook DL. Multiple imputation of missing blood pressure covariates in survival analysis. Stat Med 1999;18:681694.
18. Raghunathan TE, Lepkowski JM, van Hoewyk J, Solenberger P. A multivariate technique for multiply imputing missing values using a sequence of regression models. Surv Methodol 2001;27:8595.
19. Klompas M, Khan Y, Kleinman K, Evans RS, Lloyd JF, Stevenson K, Samore M, Platt R; CDC Prevention Epicenters Program. Multicenter evaluation of a novel surveillance paradigm for complications of mechanical ventilation. PLoS One 2011;6:e18062.
20. Klompas M, Magill S, Robicsek A, Strymish JM, Kleinman K, Evans RS, Lloyd JF, Khan Y, Yokoe DS, Stevenson K, et al.; CDC Prevention Epicenters Program. Objective surveillance definitions for ventilator-associated pneumonia. Crit Care Med 2012;40:31543161.
21. Hayashi Y, Morisawa K, Klompas M, Jones M, Bandeshe H, Boots R, Lipman J, Paterson DL. Toward improved surveillance: the impact of ventilator-associated complications on length of stay and antibiotic use in patients in intensive care units. Clin Infect Dis 2013;56:471477.
22. Muscedere J, Sinuff T, Heyland D, Dodek P, Keenan S, Wood G, Jiang X, Day A, Laporta D. The clinical impact and preventability of ventilator-associated conditions in critically ill mechanically ventilated patients. Chest 2013;144:14531460.
23. Klompas M, Kleinman K, Murphy MV. Descriptive epidemiology and attributable morbidity of ventilator-associated events. Infect Control Hosp Epidemiol 2014;35:502510.
24. Prospero E, Illuminati D, Marigliano A, Pelaia P, Munch C, Barbadoro P, D’Errico MM. Learning from Galileo: ventilator-associated pneumonia surveillance. Am J Respir Crit Care Med 2012;186:13081309.
25. Klein Klouwenberg PM, van Mourik MS, Ong DS, Horn J, Schultz MJ, Cremer OL, Bonten MJ; MARS Consortium. Electronic implementation of a novel surveillance paradigm for ventilator-associated events. Feasibility and validation. Am J Respir Crit Care Med 2014;189:947955.
26. Lilly CM, Landry KE, Sood RN, Dunnington CH, Ellison RT III, Bagley PH, Baker SP, Cody S, Irwin RS; UMass Memorial Critical Care Operations Group; UMass Memorial Critical Care Operations Group. Prevalence and test characteristics of national health safety network ventilator-associated events. Crit Care Med 2014;42:20192028.
27. Barr J, Fraser GL, Puntillo K, Ely EW, Gélinas C, Dasta JF, Davidson JE, Devlin JW, Kress JP, Joffe AM, et al.; American College of Critical Care Medicine. Clinical practice guidelines for the management of pain, agitation, and delirium in adult patients in the intensive care unit. Crit Care Med 2013;41:263306.
28. Mehta S, Burry L, Fischer S, Martinez-Motta JC, Hallett D, Bowman D, Wong C, Meade MO, Stewart TE, Cook DJ; Canadian Critical Care Trials Group. Canadian survey of the use of sedatives, analgesics, and neuromuscular blocking agents in critically ill patients. Crit Care Med 2006;34:374380.
29. Miller MA, Krein SL, George CT, Watson SR, Hyzy RC, Iwashyna TJ. Diverse attitudes to and understandings of spontaneous awakening trials: results from a statewide quality improvement collaborative. Crit Care Med 2013;41:19761982.
30. Patel RP, Gambrell M, Speroff T, Scott TA, Pun BT, Okahashi J, Strength C, Pandharipande P, Girard TD, Burgess H, et al. Delirium and sedation in the intensive care unit: survey of behaviors and attitudes of 1384 healthcare professionals. Crit Care Med 2009;37:825832.
31. Leone M, Ragonnet B, Alonso S, Allaouchiche B, Constantin JM, Jaber S, Martin C, Fabbro-Peray P, Lefrant JY; AzuRéa Group. Variable compliance with clinical practice guidelines identified in a 1-day audit at 66 French adult intensive care units. Crit Care Med 2012;40:31893195.
32. Mehta S, Burry L, Cook D, Fergusson D, Steinberg M, Granton J, Herridge M, Ferguson N, Devlin J, Tanios M, et al.; SLEAP Investigators; Canadian Critical Care Trials Group. Daily sedation interruption in mechanically ventilated critically ill patients cared for with a sedation protocol: a randomized controlled trial. JAMA 2012;308:19851992.
33. Schweickert WD, Pohlman MC, Pohlman AS, Nigos C, Pawlik AJ, Esbrook CL, Spears L, Miller M, Franczyk M, Deprizio D, et al. Early physical and occupational therapy in mechanically ventilated, critically ill patients: a randomised controlled trial. Lancet 2009;373:18741882.
34. Mekontso Dessap A, Roche-Campo F, Kouatchet A, Tomicic V, Beduneau G, Sonneville R, Cabello B, Jaber S, Azoulay E, Castanares-Zapatero D, et al. Natriuretic peptide-driven fluid management during ventilator weaning: a randomized controlled trial. Am J Respir Crit Care Med 2012;186:12561263.
35. Lellouche F, Mancebo J, Jolliet P, Roeseler J, Schortgen F, Dojat M, Cabello B, Bouadma L, Rodriguez P, Maggiore S, et al. A multicenter randomized trial of computer-driven protocolized weaning from mechanical ventilation. Am J Respir Crit Care Med 2006;174:894900.
36. Mekontso Dessap A, Katsahian S, Roche-Campo F, Varet H, Kouatchet A, Tomicic V, Beduneau G, Sonneville R, Jaber S, Darmon M, et al. Ventilator-associated pneumonia during weaning from mechanical ventilation: role of fluid management. Chest 2014:146:5865.
37. Wright MO, Kharasch M, Beaumont JL, Peterson LR, Robicsek A. Reporting catheter-associated urinary tract infections: denominator matters. Infect Control Hosp Epidemiol 2011;32:635640.
38. Serpa Neto A, Cardoso SO, Manetta JA, Pereira VG, Espósito DC, Pasqualucci MdeO, Damasceno MC, Schultz MJ. Association between use of lung-protective ventilation with lower tidal volumes and clinical outcomes among patients without acute respiratory distress syndrome: a meta-analysis. JAMA 2012;308:16511659.
39. Hébert PC, Wells G, Blajchman MA, Marshall J, Martin C, Pagliarello G, Tweeddale M, Schweitzer I, Yetisir E. A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. Transfusion Requirements in Critical Care Investigators, Canadian Critical Care Trials Group. N Engl J Med 1999;340:409417.
40. Muscedere J, Rewa O, McKechnie K, Jiang X, Laporta D, Heyland DK. Subglottic secretion drainage for the prevention of ventilator-associated pneumonia: a systematic review and meta-analysis. Crit Care Med 2011;39:19851991.
Correspondence and requests for reprints should be addressed to Michael Klompas, M.D., M.P.H., Department of Population Medicine, 133 Brookline Avenue, 6th Floor, Boston, MA 02215. E-mail:

*These authors contributed equally to this study.

Supported by the Centers for Disease Control and Prevention (grant #3 U54CK000172-01S1).

Author disclosures are available with the text of this article at www.atsjournals.org.

Author Contributions: M.K., D.A., W.T., H.B., M.P.K., L.L., E.W.E., J.J., S.M., K.K., E.L., D.S., V.F., R.A.W., and R.P. were responsible for conception, design, study implementation, interpretation of findings, and obtaining funding. R.S.-C., R.L., C.O'N., B.T.K., E.A., M.C.B., and C.E.C. made critical contributions to study implementation. C.B., J.L., and M.V.M. contributed to data acquisition, analysis, and study management. L.L., and K.K. conducted all statistical analyses. M.K. drafted the initial version of the manuscript. All authors helped revise and refine the manuscript.

Originally Published in Press as DOI: 10.1164/rccm.201407-1394OC on November 4, 2014

This article has an online supplement, which is accessible from this issue’s table of contents at www.atsjournals.org

Comments Post a Comment




New User Registration

Not Yet Registered?
Benefits of Registration Include:
 •  A Unique User Profile that will allow you to manage your current subscriptions (including online access)
 •  The ability to create favorites lists down to the article level
 •  The ability to customize email alerts to receive specific notifications about the topics you care most about and special offers
American Journal of Respiratory and Critical Care Medicine
191
3

Click to see any corrections or updates and to confirm this is the authentic version of record