| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ek, MDek is an internist with the Jagiellonian University School of Medicine and Polish Institute for Evidence Based Medicine, Krakow, Poland. Ellen McDonald and France Clarke are ICU research coordinators, Carolyn Gosse is an ICU pharmacist, Roman Jaeschke is an intensivist, and Deborah Cook is an intensivist and Academic Chair of Critical Care Medicine at St Joseph’s Healthcare and McMaster University, Hamilton, Ontario, Canada.
Corresponding author: Deborah J. Cook, McMaster Health Sciences Center, Room 2C10, Departments of Clinical Epidemiology and Biostatistics and Medicine, 1200 Main St W, Hamilton, Ontario, Canada L8N 4A6 (e-mail: debcook{at}mcmaster.ca).
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionConclusionReferences |
|---|
Objective To characterize treatment of patients with presumed pneumonia in a tertiary care intensive care unit and to identify potential areas for improvement in care.
Methods In a prospective, cohort study, the sample consisted of all consecutive patients treated in an intensive care unit during a 3-month period. For patients with presumed pneumonia, data were collected on incidence of pneumonia, diagnostic investigations, microbial isolates, and antibiotics prescribed.
Results Of 194 admissions, 73 patients were treated for pneumonia: 47 had community-acquired pneumonia; 12 had hospital-acquired pneumonia; 12 had ventilator-associated pneumonia, both early (7) and late (5); and 2 had intensive care unit–acquired pneumonia. Approximately 71% of patients had microbiological tests performed. Among 54 microbial isolates, 51.9% were gram-positive bacteria, 31.5% were gram-negative bacteria, and 9.3% were Candida species. The most commonly used antimicrobials were quinolones (54 of 192 prescriptions) and cephalosporins (33); each patient received a median of 3 antibiotics.
Conclusions Most cases of pneumonia were community acquired. The most common causative organisms were gram-positive cocci. Four quality improvement strategies were rationalization of antibiotic use during rounds, nurses’ reporting of culture results, review of antibiotic appropriateness by a pharmacist, and redesign of the clinical information system.
VAP is the most serious type of nosocomial pneumonia that develops more than 48 hours after intubation in patients receiving ventilatory support. Case-control studies3 indicate that the estimated increase in ICU length of stay in patients with VAP is 4 to 9 days.
Many ICU quality improvement initiatives have been undertaken to better understand the incidence of pneumonia, management of patients with pneumonia, and prevention of VAP.4 In Canada, the epidemiology of VAP has been described on the basis of data generated from a randomized trial.5 However, inferences derived from these data may be limited because no information was collected on early hospital-acquired pneumonia and nonconsecutive patients were included (although consecutive eligible and consenting patients were enrolled in the randomized trial).
We hypothesized that under usual practice circumstances, the incidence of pneumonia would be higher than that observed in the national study, diverse diagnostic tests with variable yields would be used, and antibiotic prescribing would not always be optimal. Observational studies of consecutive patients with pneumonia are the best source of evidence for estimating the incidence of pneumonia in ICU patients and for understanding the tests and treatment the patients receive in the usual practice setting. Accordingly, our objective was to characterize the treatment of patients with presumed pneumonia in a tertiary care ICU. We determined incidence of pneumonia, diagnostic investigations undertaken, microbial isolates, and antibiotics prescribed. On the basis of those results, we discussed potential areas for improvement in care.
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionConclusionReferences |
|---|
| Nosocomial pneumonia is associated with a mortality risk of up to 30%.
|
A blinded, prospective, observational, single-center cohort study was done. The sample consisted of all consecutive critically ill patients admitted to a closed, university-affiliated medical-surgical ICU at St Joseph’s Hospital, Hamilton, Ontario, during a 3-month period from March to June 2002. Among these patients, those treated for presumed pneumonia were defined as those patients whose attending physician gave a positive response to the question, Are you treating this patient for pneumonia?
Patients were excluded if they were treated for pneumonia upon admission to the ICU but died within 24 hours of admission or if they were presumed to be colonized only with an organism that could be a causative agent of pneumonia. The ICU team had no knowledge of the hypothesis, objective, or conduct of this study. However, because we were simultaneously conducting a randomized trial of invasive versus noninvasive diagnostic strategies for late VAP, the screening question was asked daily by the research team for another purpose.
Daily, a trained research coordinator (E.M. or F.C.) screened all consecutive ICU patients. Patients treated for presumed pneumonia were identified daily with the ICU team during bedside and radiological rounds. Data collected for each patient treated for presumed pneumonia included age and sex of the patient; admission and discharge dates; score on Acute Physiology and Chronic Health Evaluation II at the time of the ICU admission; reason for ICU admission; location before hospital or ICU admission; and cultures of endotracheal aspirate (ETA), bronchoalveolar lavage (BAL), pleural fluid, or blood done at the time treatment for presumed pneumonia began, 48 hours before initiation of treatment, and at any time afterward. In the case of microbial growth, we documented the organisms isolated. We also documented all antibiotics administered from 48 hours before the initiation of the treatment throughout the entire pneumonia treatment period in the ICU.
| More than 37% of patients admitted to the intensive care unit were treated for presumed pneumonia.
|
Weekly, the ICU pharmacist (C.G.) collected reports on microbial cultures and antibiotic treatment in duplicate to verify the data collected by the research coordinator. Any disagreements were resolved by review of the data, discussion, and consensus.
Pneumonia Classification
After data collection was complete, we classified patients treated for presumed pneumonia as having community-acquired pneumonia that was present at hospital admission or developed within the first 48 hours after admission, hospital-acquired pneumonia that developed more than 48 hours after hospital admission, ICU-acquired pneumonia that occurs in patients who are not receiving ventilatory support or who have been spontaneously breathing for more than 48 hours after extubation, early VAP (among patients receiving 2–5 days of ventilatory support), or late VAP (among patients receiving >5 days of ventilatory support). In instances of unclear classification, the original medical chart was reviewed by 2 investigators (J.B. and R.J.).
Statistical Analysis
Data were analyzed by using StatsDirect, version 1.9.14 (StatsDirect Ltd, Sale, United Kingdom). For continuous data, means and SDs or medians and interquartile ranges (IQRs), as appropriate, were calculated. Continuous variables were compared by using Mann-Whitney U tests; dichotomous variables by using 
2 analysis or the Fisher exact test.
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionConclusionReferences |
|---|
gives the baseline characteristics of all 194 patients in the sample. Compared with the other patients in the sample, the 73 patients (37.6%) treated for presumed pneumonia were older (median age 71 years, IQR 59–78 vs median age 65 years, IQR 52–76; P = .07), were more often admitted as medical patients (84.9% vs 58.7%; P < .001), and had a longer ICU stay (median 8 days, IQR 4–13 vs median 2 days, IQR 1–3; P < .001).
|
Diagnostic Tests and Microbial Isolates
During a 5-day period, from 2 days before until 2 days after the start of treatment for the presumed pneumonia, samples of blood, ETA, and BAL were cultured for 54.8%, 50.7%, and 11.0%, respectively, of the patients (Table 2). At least one sample was cultured for 71.2% of the patients. Of the blood, ETA, and BAL cultures, 17.6%, 44.6%, and 28.0%, respectively, were positive for microorganisms. The probability of a positive test result declined with time: for the period of 2 days before treatment, it was 47.3%; on the day treatment began, it was 21.5%; and on the following 2 days, it was 13.3% (Table 3).
|
|
). The most prevalent isolates were staphylococci (25.9%), streptococci (16.7%), gram-negative bacilli (25.9%; including Pseudomonas, Escherichia coli, and Klebsiella pneumoniae, 7.4% each, and Serratia marcescens, 3.7%), Candida albicans (7.4%), and Haemophilus influenzae (5.6%).
|
| Simple tests such as cultures of endotracheal aspirates and blood were not performed for about 30% of these patients.
|
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionConclusionReferences |
|---|
| The 4 quality improvement strategies pursued were rationalization of antibiotic use during rounds, nurses’ reporting of culture results, review of antibiotic appropriateness by a pharmacist, and redesign of the clinical information system.
|
These findings may reflect multiple decision makers in our teaching ICU acting both in response to and in the absence of microbiological data. The absence of culture data may represent test errors of omission, inadequate collection or processing of specimens, or decisions to forgo testing because of the perception that cultures would be negative for microorganisms in patients who had already received antibiotics. For example, we found that simple tests such as cultures of ETA and blood were not routinely performed to identify the organism responsible for pneumonia in almost 30% of the patients treated for presumed pneumonia.
In this study, treatment for presumed pneumonia was routinely initiated without invasive bronchoscopic testing. The results of randomized trials to date have been conflicting about whether diagnosis based on cultures of ETAs or BAL fluids is superior. The results of 3 trials in Spain6–8 suggested that the ETA cultures are sufficient, whereas the findings in a recent trial in France9 indicated that use of BAL allows for more tailored antibiotic therapy and results in a lower 14-day mortality. In a recent Canadian trial, 740 nonimmunocompromised patients with suspected late VAP were randomized to BAL with quantitative cultures or ETA; these 2 diagnostic strategies resulted in similar mortality, duration of ICU and hospital stay, and use of antibiotics.10
This audit provided key information on practice patterns in our ICU, indicating diagnostic and therapeutic domains suitable for continuous quality improvement initiatives. We pursued 4 strategies in response to this information. First, every antibiotic is now reviewed and its use rationalized by the intensivist-led multidisciplinary ICU team on at least 1 of the 3 daily rounds. We instituted this procedure because although drugs were presumably reviewed each day, insufficient attention was paid to the appropriateness or duration of prescribing.
Second, during rounds, each patient’s bedside nurse is responsible for reporting to the ICU team the date, time, and result of all cultures in the past week; the nurse is also responsible for ensuring that required samples for cultures are obtained if the tests have not yet been done. We instituted this procedure because although some ICU team members were aware of recent culture results, a complete, accurate accounting of the sample sites, dates of sampling, and results was seldom provided on rounds.
Third, when the ICU team prescribes antibiotics, the ICU pharmacist now reconciles drug choice with microbiological isolates and the antimicrobial sensitivities of the isolates. We instituted this procedure because lack of clear disciplinary responsibility for following up on antimicrobial sensitivities before or after rounds had resulted in lack of informed decision making about management of patients with pneumonia.
Finally, our bedside clinical information system was redesigned to incorporate clearly displayed data on specimens ordered and obtained for microbiological tests, and the results of the tests, to enhance decision making. We instituted this procedure because having key data on numerous pieces of paper from computer printouts was unanimously viewed as a disincentive to timely informed decisions about plans of care.
Limitations of this study include the single-center design. Our results most likely are generalizable primarily to centers with a similar case mix and pattern of practice. Our goal was not to generate precise estimates of pneumonia rates. The study was noninterventional, with no protocol for diagnosing pneumonia, and one of our objectives was to identify the incidence of pneumonia as treated in the real-world setting. Our classifications of pneumonia were adapted from those in the guidelines of the American Thoracic Society and Infectious Diseases Society of America.11 Because quality improvement is incremental, and projects are challenging in the complex critical care environment, we began this study targeting pneumonia before moving on to identify and improve numerous other quality indicators.12
|
Conclusion |
|---|
|
TopAbstractMethodsResultsDiscussionConclusionReferences |
|---|
This quality improvement project was designed, implemented, and analyzed by 3 clinical leaders, an ICU nurse, a respiratory therapist, and a pharmacist, with strong support from the ICU physicians. A successful quality improvement program requires a clear ongoing commitment of the entire interdisciplinary ICU team. This project resulted in 4 multidisciplinary quality initiatives to improve ICU management of patients with pneumonia.
|
ACKNOWLEDGMENTS |
|---|
FINANCIAL DISCLOSURES
The study was supported by Father Sean O’Sullivan Research Center and an arm’s-length grant from Pharmacia Incorporated, which had no role in the design, conduct, analysis, or interpretation of the study.
To purchase electronic or print reprints, contact The InnoVision Group, 101 Columbia, Aliso Viejo, CA 92656. Phone, (800) 809-2273 or (949) 362-2050 (ext 532); fax, (949) 362-2049; e-mail, reprints{at}aacn.org.
Now that you’ve read the article, create or contribute to an online discussion about this topic using eLetters. Just visit www.ajcconline.org and click "Respond to This Article" in either the full-text or .pdf view of the article.
To learn more about preventing pneumonia in the ICU, visit http://ccn.aacnjournals.org and read the article by Hilinksi and Stark, "Memory Aide to Reduce the Incidence of Ventilator-Associated Pneumonia" (Critical Care Nurse, October 2006).
|
REFERENCES |
|---|
|
TopAbstractMethodsResultsDiscussionConclusionReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
