Quality improvement (QI) continues to play an increasingly

Transcription

1 ORIGINAL ARTICLE Measurable Outcomes of Quality Improvement Using a Daily Quality Rounds Checklist: One-Year Analysis in a Trauma Intensive Care Unit With Sustained Ventilator-Associated Pneumonia Reduction Joseph DuBose, MD, Pedro G. R. Teixeira, MD, Kenji Inaba, MD, Lydia Lam, MD, Peep Talving, MD, PhD, Brad Putty, MD, David Plurad, MD, Donald J. Green, MD, Demetrios Demetriades, PhD, MD, and Howard Belzberg, MD Background: We have previously demonstrated that the use of a daily Quality Rounds Checklist (QRC) can increase compliance with evidencebased prophylactic measures and decrease complications in a busy trauma intensive care unit (ICU) over a 3-month period. This study was designed to determine the sustainability of QRC use over 1 year and examine the relationship between compliance and outcome improvement. Methods: A prospective before-after design was used to examine the effectiveness of the QRC tool in documenting compliance with 16 prophylactic measures for ventilator-associated pneumonia (VAP), deep venous thrombosis, pulmonary embolism, catheter-related bloodstream infection, and other ICU complications. The QRC was implemented on a daily basis for a 1-year period by the ICU fellow on duty. Monthly compliance rates were assessed by a multidisciplinary team for development of strategies for real-time improvement. Compliance and outcomes were captured over 1 year of QRC use. Results: QRC use was associated with a sustained improvement of VAP bundle and other compliance measures over a year of use. After multivariable analysis adjusting for age ( 55), injury mechanism, Glasgow Coma Scale score ( 8), and Injury Severity Score ( 20), the rate of VAP was significantly lower after QRC use, with an adjusted mean difference of 6.65 (per 1,000 device days; 95% confidence interval, 9.27 to 4.04; p 0.008). During the year of QRC use, 3% of patients developed a VAP if all four daily bundle measures were met for the duration of ICU stay versus 14% in those with partial compliance (p 0.04). The overall VAP rate with full compliance was 5.29 versus 9.23 (per 1,000 device days) with partial compliance. Compared with the previous year, a 24% decrease in the number of pneumonias was recorded for the year of QRC use, representing an estimated cost savings of approximately $400,000. Conclusion: The use of a QRC facilitates sustainable improvement in compliance rates for clinically significant prophylactic measures in a busy Level I trauma ICU. The daily use of the QRC, requiring just a few minutes per patient to complete, equates to cost-effective improvement in patient outcomes. Submitted for publication May 18, Accepted for publication September 4, Copyright 2010 by Lippincott Williams & Wilkins From the Division of Trauma and Critical Care at the Los Angeles County University of Southern California Medical Center, Los Angeles, California. Address for reprints: Joe DuBose, MD, 1200 North State Street, Inpt tower 2LC100, 5th floor, Los Angeles, CA ; jjd3c@yahoo.com. DOI: /TA.0b013e3181c4526f Key Words: Trauma, Intensive care, Quality improvement, Ventilatorassociated pneumonia, Prophylaxis. (J Trauma. 2010;69: ) Quality improvement (QI) continues to play an increasingly important role in health care, partly because of recognition that 44,000 to 98,000 Americans may die each year as a result of errors in hospital care. 1,2 Even using the lower, more conservative range of this estimate would place preventable adverse events as the eighth leading cause of death in the United States, surpassing that attributable to motor vehicle crashes. An improved awareness of these dangers has fueled healthcare reforms driven directly by government regulation and indirectly by market forces, with the latter impact manifesting changes both in Medicare reimbursement withholdings and private insurer incentives. The importance of QI is perhaps most apparent in the intensive care unit (ICU), where the level of acuity and complexity of care demand a substantial commitment to preventative practices. A growing abundance of best-practice measures, however, make QI initiatives difficult to implement and sustain. 3 However, the effective introduction of preventative interventions has been associated with reduction of such serious complications of care, such as catheter-related blood stream infection and ventilator-associated pneumonia (VAP). 4 8 These benefits have been achieved by improving compliance with evidence-based practices, including those outlined in the Surviving Sepsis Campaign, 9 through the use of bundling and other innovative solutions. At our institution, a Quality Rounds Checklist (QRC) was developed to improve compliance with evidence-based preventative measures. We have previously reported that daily use of the QRC resulted in both an improved compliance with prophylactic goals and a decrease in adverse outcomes, including VAP. 10 Although these findings were encouraging, our initial study was limited by the brief duration of its design. The present examination was designed to determine the sustainability of daily QRC use and to establish a correlation between the subsequent compliance achieved and resulting clinical outcomes, during a 14-month period. The Journal of TRAUMA Injury, Infection, and Critical Care Volume 69, Number 4, October

2 DuBose et al. The Journal of TRAUMA Injury, Infection, and Critical Care Volume 69, Number 4, October 2010 Figure 1. LAC USC daily QRC. METHODS This was a prospective, observational study at a busy urban Level I trauma center conducted during a 14-month period. During the first month, the QRC was used, in a blinded fashion by a researcher not engaged in patient care, to identify baseline compliance with prophylactic measures in our trauma ICU. For the remainder of the study period, the clinical ICU fellow on duty used the QRC during rounds to determine compliance with prophylactic measures for every trauma patient on a daily basis. The QRC consists of 22 data points, including 16 prophylactic best-practice measures. These measures include VAP prevention bundle interventions, central line and other invasive device assessment, glucose control, nutrition, need for antibiotics, and code status (Fig. 1). The QRC served as both a cue to identify deficiencies for immediate correction and a method to prospectively monitor compliance. Data collected were entered into an Excel spreadsheet for subsequent statistical analyses of compliance results. These figures were then reviewed at monthly multidisciplinary meetings to identify systemic deficiencies and develop cohesive strategies to address areas in need of improvement. In addition to compliance rates, we also analyzed clinical outcome measures including mortality, ICU length of stay (ICU LOS) and hospital LOS, and morbidities, which include the development of pulmonary embolism (PE), deep venous thrombosis (DVT), gastrointestinal (GI) bleed, and VAP. DVT and PE were detected by duplex sonography and computed tomography with pulmonary angiography when the need for obtaining these tests was determined as clinically indicated by the care team. VAP was diagnosed using Centers for Disease Control and Prevention National Nosocomial Infectious Surveillance definitions 11 in the process of prospective surveys by a dedicated hospital epidemiologic team consisting of the trauma ICU Director, an epidemiology nurse, and board-certified infectious disease-attending physician. To develop comparison cohorts from the pre- and post-qrc implementation periods, trauma patients admitted to the ICU 48 hours during the period of QRC use were compared with a similar control cohort from the year before introduction of the QRC. Normally distributed, continuous variables were compared using Student s t test, and the Mann-Whitney U test 856 was used for those with an abnormal distribution. The 2 test or Fisher s exact test was used to compare categorical variables. Variables commonly associated with a higher incidence of VAP were entered into a multivariable logistic regression analysis to obtain adjusted rates of VAP calculated in cases per 1,000 ventilator days. The adjusted odds ratio, with 95% confidence intervals, was derived from logistic regression analysis after adjusting for possible confounders, and statistical significance was set at p All statistical analysis was performed using SPSS software, version 12.0, for Windows (SPSS, Chicago, IL). To better define the impact of compliance on clinical outcomes, patients within the study cohort were analyzed further in an attempt to identify a relationship between VAP bundle compliance and the incidence of VAP. All patients ventilated 48 hours were divided into partial and full compliance groups. For the purpose of our study, full compliance was defined as accomplishing all four of the VAP bundle measures within 24 hours of admission and maintaining this standard for the duration of ICU stay. After identifying these two cohorts, their basic demographics and clinical end points were then compared, using a p value 0.05 to determine statistical significance. RESULTS Compared with the month before implementation, compliance with the four measures of the VAP bundle remained significantly elevated at 3, 6, and 13 months postimplementation of the QRC (Fig. 2; Table 1). For comparison of clinical end points, the 570 patients admitted during the QRC implementation period were compared with the 577 trauma ICU admit controls from the year before QRC introduction. Demographics were compared between these two groups, revealing no statistically significant differences in clinically relevant demographics (Table 2). In addition, no statistically significant differences in the clinical end points of mortality, development of GI bleeding and PE, or ICU LOS and hospital LOS were identified (Table 3). However, there was a reduction in the occurrence of VAP between the pre-qrc and post-qrc groups. The overall number of identified VAP decreased from 33 in the year before QRC introduction to 25 during the year of implemen Lippincott Williams & Wilkins

3 The Journal of TRAUMA Injury, Infection, and Critical Care Volume 69, Number 4, October 2010 Measurable Outcomes of Quality Improvement Figure 2. One-year compliance with VAP bundle measures. HOB, head of bed; PUD, peptic ulcer disease. TABLE 1. Comparison Between Pre-QRC Compliance With VAP Bundle Measures in the Month Before Implementation and at 3, 6, and 13 mo After QRC Implementation Preimplementation 3 mo 6 mo 1 yr Percent of HOB 30 degrees or greater PUD prophylaxis DVT prophylaxis No continuous sedation/holiday PUD, peptic ulcer disease; HOB, head of bed. TABLE 3. Comparison of Outcomes Before and After QRC Implementation Pre-QRC (N 577) Post-QRC (N 570) p Deaths, % (n) 10.9 (63) 9.3 (53) 0.37 Gastrointestinal hemorrhage, % (n) 0.2 (1) 0.2 (1) 0.99 Pulmonary embolism, % (n) 0.7 (4) 0.9 (5) 0.72 ICU (d) SD Hospital (d) SD SD, standard deviation. TABLE 2. Comparison of Demographics and Clinical Characteristics Between Trauma Patients Before and After QRC Implementation Pre-QRC (N 577) Post-QRC (N 570) p Mean age, SD (yr) Age 55 yr, % (n) 20.5 (118) 22.3 (127) 0.45 Male, % (n) 79.4 (458) 77.0 (439) 0.33 Mechanism blunt, % (n) 78.3 (452) 77.9 (444) 0.86 Mean SBP SD (mm Hg), % (n) Hypotension on admission (mm Hg), 6.8 (39) 7.4 (42) 0.68 % (n) GCS score 8 on admission, % (n) 23.9 (137) 26.9 (153) 0.24 Mean ISS SD ISS score 20, % (n) 39.9 (230) 34.7 (198) 0.07 Revised Trauma Score Severe abdomen AIS ( 3), % (n) 21.0 (121) 22.3 (127) 0.59 Severe head AIS ( 3), % (n) 49.9 (288) 45.8 (261) 0.16 Severe chest AIS ( 3), % (n) 30.7 (177) 29.5 (168) 0.66 SD, standard deviation; SBP, systolic blood pressure; GCS, Glasgow Coma Scale, ISS, Injury Severity Score; AIS, Abbreviated Injury Scale. TABLE 4. QRC Effect on VAP No. VAPs VAP Rate (Per 1,000 Device d) Pre-QRC Post-QRC Adjusted mean difference (adjusted for age 55 yr, mechanism, GCS 8, and ISS 20); 95% CI ; p GCS, Glasgow Coma Scale, ISS, Injury Severity Scale; CI, confidence interval. tation. This correlated to a reduction in VAP incidence from to 8.74 (per 1,000 ventilator days). After multivariable adjustments for age 55 years, mechanism, Glasgow Coma Scale score 8, and Injury Severity Score 20, this difference proved a statistically significant improvement in VAP rates, with an adjusted mean difference of 6.65 (per 1,000 ventilator days; 95% confidence interval, ; p 0.008; Table 4). In our comparison of partial versus full compliance groups from the QRC implementation period, we noted a significantly lower incidence of VAP among those patients in whom full compliance with VAP bundle measures was obtained, i.e., from 13.4% to 3.5% (p 0.04). The corresponding VAP rates equated to 9.23 (per 1,000 ventilator days) for 2010 Lippincott Williams & Wilkins 857

4 DuBose et al. The Journal of TRAUMA Injury, Infection, and Critical Care Volume 69, Number 4, October 2010 the partial compliance group and 5.29 (per 1,000 ventilator days) for the full compliance cohort (Table 5). When compared with patients for whom only partial compliance was achieved, the full compliance cohort was also noted to have statistically significant reductions in duration of mechanical ventilation (6.2 days 4.5 days vs days 13.5 days; p 0.001), ICU LOS (9.4 days 7.7 days vs days 12.5 days; p 0.001), hospital LOS (16.1 days 11 days vs days 31.7 days; p 0.001), and hospital charges ($143,554 $100,971 vs. $311,930 $268,221; p 0.001; Table 6). TABLE 5. Comparing the Demographic and Clinical Characteristics Between Adult Trauma Patients Ventilated for 48 hr for Whom QRC Was Used Full Compliance (N 57) Partial Compliance (N 119) p Mean age SD (yr) Age 55 (yr), % (n) 15.8 (9) 20.2 (24) 0.49 Male, % (n) 78.9 (45) 79.0 (94) 1 Mechanism blunt, % (n) 64.9 (37) 82.4 (98) 0.01 Mean SBP SD (mm Hg) Hypotension on admission 5.3 (3) 11.8 (14) 0.17 (mm Hg), % (n) Respiratory rate SD, (per min) GCS score 8 on admission, % (n) 36.8 (21) 37.8 (45) 0.9 Mean ISS SD ISS score 20, % (n) 47.4 (27) 49.6 (59) 0.78 Revised Trauma Score Spinal fracture, % (n) 14.0 (8) 19.3 (23) 0.39 Cervical fracture, % (n) 5.3 (3) 8.4 (10) 0.55 Thoracolumbar fracture, % (n) 8.8 (5) 14.3 (17) 0.3 Spinal cord injury, % (n) 5.3 (3) 5.9 (7) 1 Severe abdomen AIS ( 3), % (n) 35.1 (20) 26.9 (32) 0.27 Severe head AIS ( 3), % (n) 35.1 (20) 44.5 (53) 0.23 Severe chest AIS ( 3), % (n) 36.8 (21) 49.6 (59) 0.11 SD, standard deviation, SBP, systolic blood pressure, GCS, Glasgow Coma Scale, ISS, Injury Severity Scale, AIS, Abbreviated Injury Scale. TABLE 6. Comparing Outcomes Between Adult Trauma Patients With and Without Full QRC Compliance Full Compliance (57) Partial Compliance (119) p VAP, % (n) 3.5 (2) 13.4 (16) 0.04 Deaths, % (n) 17.5 (10) 11.2 (13) 0.25 Ventilator days (d), mean SD ICU LOS (d), mean SD Hospital LOS (d), mean SD Charge, mean SD ($) 143, , , , SD, standard deviation. DISCUSSION Born out of the revolution in industrial processes that occurred in the United States during the 1920s, quality improvement initiatives have evolved to be incorporated into a variety of different arenas. During this evolution, proponents of QI philosophies have often studied process improvement endeavors that are found to be successful in one field to innovatively modify them for use in another. In the same fashion, we have altered a common, successful practice used in the aviation industry for application in the realm of trauma critical care. Although not all doctors are pilots, and patients are certainly not planes, there are some important similarities between critical care and operating modern aircraft. The flight of contemporary airframes is a high-performance, complex endeavor in which overlooked safety measures significantly increase the risk for adverse outcome, including death. These facts were not lost on the B-17 test pilots of Boeing, who in 1935, were confronted with one of the most advanced flying craft their industry had ever seen. Faced with the daunting task of preparing this complex new plane for a safe flight, these pioneers developed pilot checklists consisting of consistent and reproducible step-by-step algorithms for takeoff, flight, landing, and taxiing. Similar to the B-17 of the 1930s, the modern ICU has become increasingly more advanced than its predecessors. Safe care of patients in this environment requires hundreds of daily actions by nurses and doctors, including medication administration, maintenance of invasive devices, and prophylactic interventions. To effectively and safely accomplish, these tasks requires significant manpower and attention to detail, often using innovative solutions and tools. In 2001, Pronovost et al. 12 described their use of a ICU daily goals form, which was associated with a 50% reduction in ICU LOS. In conjunction with a nationwide movement for quality improvement in health care, 13,14 many groups have followed suit in adopting checklists and other tools to aid them in improving adherence to best-practice measures in complex healthcare settings. In our previous study, 10 we described the design of the QRC and our use of this tool in a trauma ICU. With the institution of the checklist, we were able to achieve a statistically significant improvement in our compliance with a number of daily quality improvement measures. Subsequently, these gains were associated with a decrease in rates of both catheter-related blood stream infections and VAP in our patient population. Our approach somewhat differs from previously reported strategies, 15,16 in that the QRC is designed to be used by a surgical intensivist in the completion of their daily rounds, enabling the provider to correct deficiencies in real time. In addition, the data provided by the use of the QRC are collected on a continual basis and reviewed at a monthly multidisciplinary Trauma and Critical Care Quality Improvement Committee meeting. In this setting, systemic and policy changes for additional improvement can be discussed, developed, and implemented for all the ICU environments of our facility. One of the primary goals of this examination was to determine whether improvements in compliance with 2010 Lippincott Williams & Wilkins

5 The Journal of TRAUMA Injury, Infection, and Critical Care Volume 69, Number 4, October 2010 Measurable Outcomes of Quality Improvement evidence-based practices would prove sustainable beyond the 3-month period of our initial examination. There are several factors inherent to academic critical care environment that represent potential impediments to the accomplishment of this task. Changes in personnel and staff turnover represent significant challenges, as each new member of the team must be oriented to the goals and completion of the QRC. In our own experience, these concerns were magnified by the fact that, in the midst of the study, our entire ICU was relocated to a newly built facility with an increased number of ICU beds and new surroundings. To combat these adversities, new fellows were initiated at the beginning of the academic year and were trained in appropriate QRC use. This approach assured that each new fellow rapidly learned to complete the QRC for the entire trauma population in less than 1 hour on a daily basis. Regular in-servicing of nursing staff on the importance of prophylactic measures was also conducted. These efforts contributed to the creation of a culture in our ICU in which nurses, respiratory therapists, and other ancillary staff were appropriately educated in QRC measures and were better able to contribute in bringing deficiencies to the attention of the rounding personnel. Another major goal of our study was to examine the relationship between QRC implementation over a significant time period and subsequent clinical outcomes. The concept of care bundling as developed by the Institute for Healthcare Improvement 17 is well established and accepted. The potential efficacy of these practices in improving clinical outcomes has also been supported in literature. 6,7 One of the more commonly studied grouped interventions is the VAP prevention bundle, consisting of elevation of the head of bed 30 degrees, subglottic suctioning, daily sedation holiday with assessment for ventilation weaning, and both DVT and GI prophylaxis. By incorporating this bundle of interventions in our QRC, we were able to appreciate a subsequent decrease in the VAP rates in our patient population. Comparing the VAP rate for the year before QRC implementation with that during the same interval after initiation of daily tool use, an adjusted mean difference of 6.65 per 1,000 device days in favor of the latter period was observed. We, further, specifically examined the period of QRC use to determine a correlation between the degree of compliance with VAP bundle measures and the subsequent incidence of VAP. In this portion of our study, we found that if full compliance with all four measures was achieved within 24 hours of admission and sustained for the duration of the ICU stay, 3% of patients developed VAP. In contrast, among patients in whom these criteria were not achieved, 24% developed this significant morbidity. Further, these findings support the potential value of bundled VAP prophylactic interventions, while also highlighting the benefit that high levels of compliance with these interventions are likely to represent in both avoidance of morbidity and cost savings. Our study has several important limitations. In particular, our study design did not provide us adequate data that could be used to delineate why we were unable to achieve full compliance with VAP bundle measures in some patients. When we compared those patients for whom full compliance was achieved with those in whom this goal was not accomplished (Table 5), we found that they were similar in almost every regard, the exception being that those with partial compliance were more commonly injured by penetrating mechanisms. We surmise that these patients might have been subjected to longer periods of spine precautions, preventing head of bed elevation within the 24 hour period required to meet the definition of full compliance as defined by our design. Although the incidence of spinal fractures (Table 5) was the same between the two groups, this remains a possibility. The absence of DVT prophylaxis in the early phases of significant intracranial or solid organ injury might also have contributed to higher rates of partial compliance. The myriad of other factors that might contribute to a physician s decision to maintain spinal precautions beyond 24 hours were not adequately captured by our design. As our tool continues to evolve, we have focused additional effort on the more precise identification of the reasons why full compliance is not attained in each patient. By incorporating data on the medical rationale for these deficiencies into our ongoing-quality review and outcomes analysis, we hope to learn more about the reasons for and impact of partial compliance for these patients. However in this report,, the absence of this information represents a significant shortcoming. In addition, although we were able to show a decrease in VAP rates among our study population after QRC use, we were not able to demonstrate an associated improvement in other clinical outcomes, such as peptic ulcer disease complications, occurrence of DVT or PE, or mortality. It is also important to note that our study population primarily consisted of trauma ICU patients; extrapolation of our results to other ICU populations requires additional investigation. Any endeavor in data collection relying so heavily on human input also lends itself to discrepancy and data lapses. In our experience, however, we found that heavy emphasis on the importance of the effort combined with the requirement of the fellows to present their data in a multidisciplinary setting at the end of their month in the trauma ICU created an environment in which each respective data point examined was recorded accurately with a frequency exceeding 95%. Finally, as with any quality tool, the QRC used by our group remains a dynamic instrument that requires continuous reevaluation in a multidisciplinary setting and education to remain a valid instrument of improved care. Successful implementation of a QRC or other evidencebased prophylactic tool demands collaboration and commitment from providers at every level and a shared dedication to creating an organized culture of patient safety and quality improvement. Some key factors that may be required for sustainability include strong leadership, education to all levels, and user-friendly tool with which to collect and analyze data on a daily basis. 18 Future challenges for these initiatives include the use of new technologies, including portable electronic devices, to assist in these endeavors. CONCLUSION Our results suggest that the daily use of a QRC to support compliance with evidence-based endeavors is a sus Lippincott Williams & Wilkins 859

6 DuBose et al. The Journal of TRAUMA Injury, Infection, and Critical Care Volume 69, Number 4, October 2010 tainable venture that may significantly improve VAP rates in a busy trauma ICU. Although the factors affecting the ability to achieve full compliance with prophylactic measures require further investigation, there also appears to be a relationship between bundle compliance rates and the incidence of VAP in this patient setting. REFERENCES 1. Organisation for Economic Cooperation and Development. Available at: Accessed February Institute of Medicine. To Err Is Human: Building a Safer Health System Available at: Accessed February Mittman BS. Creating the evidence base for quality improvement collaboratives. Ann Intern Med. 2004;140: Berenholtz SM, Pronovost PJ, Lipsett PA, et al. Eliminating catheterrelated bloodstream infections in the intensive care unit. Crit Care Med. 2004;32: McKee C, Berkowitz I, Cosgrove SE, et al. Reduction of catheterassociated bloodstream infections in pediatric patients: experimentation and reality. Pediatr Crit Care Med. 2008;9: Pronovost P, Needham D, Berenholtz S, et al. An intervention to decrease catheter-related bloodstream infections in the ICU. N Engl J Med. 2006;355: Resar R, Pronovost P, Haraden C, Simmonds T, Rainey T, Nolan T. Using a bundle approach to improve ventilator care processes and reduce ventilator-associated pneumonia. Jt Comm J Qual Patient Saf. 2005;31: Berenholtz SM, Milanovich S, Faircloth A, et al. Improving care for the ventilated patient. Jt Comm J Qual Saf. 2004;30: Dellinger RP, Levy MM, Carlet JM, et al. Surviving sepsis campaign: international guidelines for management of severe sepsis and septic shock: Crit Care Med. 2008;36: DuBose JJ, Inaba K, Shiflett A, et al. Measurable outcomes of quality improvement in the trauma intensive care unit: the impact of a daily quality rounding checklist. J Trauma. 2008;64:22 27; discussion Horan TC, Andrus M, Dudeck MA. CDC/NHSN surveillance definition of health care-associated infection and criteria for specific types of infections in the acute care setting. Am J Infect Control. 2008;36: Pronovost PJ, Berenholtz SM, Ngo K, et al. Developing and pilot testing quality indicators in the intensive care unit. J Crit Care. 2003;18: Leapfrog Group. Available at: Accessed February Wachter RM, Pronovost PJ. The 100,000 Lives Campaign: a scientific and policy review. Jt Comm J Qual Patient Saf. 2006;32: Craven DE. Preventing ventilator-associated pneumonia in adults: sowing seeds of change. Chest. 2006;130: Cocanour CS, Peninger M, Domonoske BD, et al. Decreasing ventilatorassociated pneumonia in a trauma ICU. J Trauma. 2006;61: ; discussion Institute for Healthcare Improvement. Available at: ihi Accessed December 1, Wall RJ, Ely EW, Elasy TA, et al. Using real time process measurements to reduce catheter related bloodstream infections in the intensive care unit. Qual Saf Health Care. 2005;14: Lippincott Williams & Wilkins