Febrile neutropenia in EDs: the role of an electronic clinical practice guideline☆☆☆

Article Outline

• Abstract
• 1. Introduction
• 2. Methods
  • 2.1. Study design
  • 2.2. Study setting
  • 2.3. Study population
  • 2.4. Case definition
  • 2.5. Electronic clinical practice guideline
  • 2.6. Outcomes
  • 2.7. Data collection
  • 2.8. Statistical analysis
  • 2.9. Ethics
• 3. Results
  • 3.1. Sample and demographics
  • 3.2. Electronic clinical practice guideline use
  • 3.3. Management
  • 3.4. Disposition
• 4. Discussion
• 5. Conclusions
• Appendix A: Data Collection Form
• References
• Copyright

Abstract 

Purpose

Evidence-based clinical practice guidelines (CPGs) for managing febrile neutropenia (FN) are widely available; however, the integration of guidelines into routine practice is often incomplete. This study evaluated the uptake and clinical impact of implementing an electronic CPG on the management and outcomes of patients presenting with FN at 4 urban emergency departments (ED).

Methods

A retrospective chart review over a 3-year period at 4 hospitals in Edmonton, Alberta, was performed. Potentially eligible patient visits were identified by searching the Ambulatory Care Classification System database using International Classification of Diseases, 10th Edition, codes and ED physician diagnoses of FN. ED patients with fever (>38°C at home or in ED) and neutropenia (white blood cell count of <1000 cells/mm³ or a neutrophil count of <500 cells/mm³) who received an ED diagnosis of FN were included.

Results

From 371 potential cases, 201 unique cases of FN were included. Overall, the electronic CPG was used in 76 (37.8%) of 201 patient visits; however, there were significant differences in CPG utilization between hospitals. Clinical practice guideline usage was greatest at the University of Alberta Hospital (57%). This finding correlated with a decrease in time from triage to first antibiotic by 1 hour compared to the 3 control hospitals (3.9 vs 4.9 hours, P = .022).

Conclusions

The electronic CPG is a useful clinical tool that can improve patient management in the ED, and strategies to increase its utilization in this and other regions should be pursued.

Back to Article Outline

1. Introduction 

Febrile neutropenia (FN) is a potentially life-threatening condition that requires urgent attention and management in the emergency department (ED) [1]. ED management is becoming increasingly important as the trend toward outpatient chemotherapy treatments continues to grow [2]. There is widespread consensus that early initiation of broad-spectrum antibiotic treatment is a cornerstone in effective ED management of FN [3], [4].

Evidence-based clinical guidelines for managing FN have been developed and are widely available; however, the integration of guidelines into routine practice is usually incomplete [5], [6]. Reviews have shown that computer decision support systems (CDSSs) can be effective in narrowing gaps between “what we know” and “what we do” in clinical practice [7]. As a result, in 2001, the health informatics team at the University of Alberta Hospital (UAH) collaborated with other clinical colleagues to develop and implement an electronic clinical practice guideline (eCPG) for managing FN in the ED. Prior studies have assessed physician satisfaction with other eCPG decision support tools as well the sensibility of the tools; however, the impact of using an eCPG on important time targets and the clinical management of FN patients in the ED remains unproven [8], [9].

This study was designed to assess the uptake and clinical impact of the eCPG intervention on patients presenting with FN at 4 EDs in the urban health region of Edmonton, Alberta.

Back to Article Outline

2. Methods 

2.1. Study design 

A retrospective comparative cohort study was conducted to examine the management of patients with an ED diagnosis of “febrile neutropenia” between January 1, 2006, and December 31, 2008, using a chart review methodology.

2.2. Study setting 

This health region participates fully in the Ambulatory Care Classification System (ACCS) database maintained by the Alberta Health and Wellness. In Alberta, health records coders examine all ED charts and code information in a standardized manner, which is then submitted and included in the ACCS database. The ACCS database was used to identify eligible cases as it was the most accurate means of ensuring capture of all ED patients presenting with FN.

Charts were obtained from 4 EDs in the health region of Edmonton, Alberta: UAH, Royal Alexandra Hospital (RAH), Grey Nuns Community Hospital, and Misericordia Community Hospital. The UAH is an academic teaching hospital and has more experience with eCPG development, implementation, and clinical utilization than the other hospitals. The RAH and UAH are designated trauma hospitals and the UAH is the designated ED destination for patients receiving treatment at the largest cancer center (Cross Cancer Institute) in Northern Alberta.

2.3. Study population 

We screened adult (≥17 years of age) patients presenting to the ED receiving International Classification of Diseases, 10th Edition (ICD-10) codes of agranulocytosis or neutropenia (ICD-10 code D70). At the UAH, all eligible visits in the 3-year study period were screened. A random sample (n = 40) of patient visits from each of the other 3 hospitals were selected. We used the following data elements from the ACCS database: hospital number, registration date, age, sex, Canadian Emergency Department Triage and Acuity Score, discharge disposition, and final diagnosis.

The UAH was designated as the intervention hospital because the eCPG application was primarily developed and had the greatest clinical penetration at that site. The FN eCPG is one of more than 80 clinical decision support tools available in the eCPG application, and regional data indicate there are more than 9000 annual uses at the UAH site alone (D. Meurer, AHS; personal communication). The methods for knowledge dissemination at this site have included educational sessions, survey and feedback from the clinicians, and iterative changes approved by representatives of the participating clinical groups. The remaining 3 hospitals were designated as control hospitals as they were not developers and had less experience with the eCPG application. Characteristics of the hospitals and EDs from the 4 sites are detailed in Table 1.

Table 1. Characteristics of 4 hospitals participating in the Febrile Neutropenia study for the fiscal year 2008-2009

	Hospital types
	Intervention	Control 1	Control 2	Control 3
No. of ED visits	77 130	61 759	57740	48 993
No. of inpatient bedsa	650	678	267	259
No. of admissions	15 426	12 100	6074	5489
Percentage admissions of ED patients	20.00%	19.60%	10.50%	11.20%
Average ED LOSb (h)
All patients	7.9	9.6	6.9	6.3
Discharged patients	4.9	6.3	4.9	3.5
Admitted patients	22.4	22.9	25.7	26.6
Affiliated with cancer treatment center	Yes	No	No	No

LOS indicates length of stay.

2.4. Case definition 

A combination of criteria from the Emergency Medicine and Infectious Diseases literature defining FN was used: patients were included if they had an absolute white blood cell (WBC) count of less than 1000 cells/mm³ or a neutrophil count of less than 500 cells/mm³ [10], [11]. All patients were required to have a fever of higher than 38.0°C at home or in the ED in accordance with Infectious Diseases Society of America guidelines [11]. Patients were eligible for inclusion if an ED physician made a final primary or secondary diagnosis of FN and the patient's temperature and laboratory values met the criteria for FN. Patients were excluded if they were not seen by an ED physician because they were directly admitted to an inpatient ward, if they left without being seen by an ED physician or without completing their treatment, or if their medical records were not found. For patients with multiple ED presentations of FN during the study period, only the first encounter was considered.

2.5. Electronic clinical practice guideline 

The FN eCPG was developed and implemented at the UAH in 2001 by a team of clinicians and health informatics specialists. All eCPGs are reviewed and modified on an ongoing basis to ensure that the recommended therapies are consistent with up-to-date clinical evidence. The FN eCPG was updated in 2005 to reflect new antibiotic recommendations. Physicians at all 4 hospitals may access this eCPG via an Intranet web browser portal; however, its use has never been made mandatory and patient management practice remains at the discretion of the most responsible physician. Moreover, the physicians at all sites were unaware of the study at the time of the patient encounter.

2.6. Outcomes 

This study examined both informatics-related outcomes focused on physician behavior and ED workflow, as well as FN patient clinical outcomes. The primary outcome for the informatics component was the proportion of patient visits in which the eCPG was used, compared between the intervention and control hospitals. Secondary outcomes included changes in eCPG use over time and time flow related to ED management of FN patients, particularly the time interval from triage to the initial administration of antibiotics. The primary outcome of interest for the clinical component of the study was change in patient outcomes (length of ED stay, disposition). Investigations (cultures, imaging) and treatments (antibiotics, granulocyte colony-stimulating factor (GCF)) used in management are also reported.

2.7. Data collection 

A code book containing operational definitions of all variables was developed before data collection. Data were collected from ED charts, from the regional electronic health records (Alberta NetCare), and from the online Emergency Department Information System database by 3 trained data abstractors using a standardized form. The abstractors were aware of the study hypothesis; however, the main clinical outcomes were obtained from administrative data to reduce bias. Administrative databases data were used to record patient age and sex, time and date of patient presentation, time and date seen by the emergency physician (EP), time of discharge, disposition, and final diagnosis. All remaining data were abstracted from the ED patient charts. Interobserver agreement and reliability were assessed from a random selection of 29 charts using the κ statistic and the intraclass correlation coefficient (ICC). Variables analyzed included time and date of patient presentation, time and date seen by EP, time and date of first vital signs, and time and date of antibiotics administered. Other variables included patient age and sex, use of the eCPG, value of first WBC count, value of first Polymorphonuclear leucocyte count, antibiotics used, disposition, and final diagnosis.

2.8. Statistical analysis 

Data analyses were performed using Stata Statistical Software Release 10.0 (Stata Corporation, College Station, TX). Descriptive analyses included proportions for categorical variables and medians with interquartile ranges for continuous variables as most of the data did not follow a normal distribution. Bivariable analyses were performed using χ² or Mann-Whitney U tests according to the nature of the variables.

2.9. Ethics 

The protocol for this study was approved by the health research ethics board of the University of Alberta and patient consent was not required. Physicians were unaware that a study would be conducted at the time of the patient encounter.

Back to Article Outline

3. Results 

3.1. Sample and demographics 

In total, 371 unique visits were identified as potential cases based on ICD-10 criteria and 201 patient visits were included for review in this study. Exclusions occurred as follows: 54 patients received ED diagnoses other than FN, 48 patients were directly admitted to hospital, 1 patient left without treatment, and 57 patients did not meet the inclusion criteria. Medical records were not found for 10 patients. The interobserver agreement for major study outcomes (e.g., time seen by ED physician, time to first antibiotic, disposition, and final diagnosis) among all 3 data abstractors was excellent (κ/ICC >0.9). The κ was lower (0.6) for the variable “use of other antibiotics” because it was a free text entry, unlike other fixed variables such as use of piperacillin-tazobactam or ceftazidime. These outcomes were selected based on relevance to the study objectives and with the purpose of representing a broad spectrum of variable types.

There were 128 visits at the intervention hospital and 73 visits at the control hospitals (16 at RAH [control 1], 25 at GNH [control 2], 32 at Misericordia Community Hospital [control 2]). Patient demographics and vital signs upon presentation to the ED were similar between both groups. The median temperature taken at ED presentation was 38.2°C at the intervention hospital and 38.0°C in the control hospital group. The median absolute neutrophil count was 0.1 × 10³ cells/mm³ for both groups (see Table 2).

Table 2. Characteristics of patients with FN according to treatment at intervention vs control hospital sites

	Intervention hospital (n = 128)	Control hospitals (n = 73)	P
Patient history
Age (y), median (IQR)	51 (40, 65)	57 (47, 68)	.09
Female sex, n (%)	56 (43.7%)	28 (38.4%)	.46
Allergy to any medications, n (%)	50 (39.1%)	29 (39.7%)	.97
Features at ED presentation
Vital signs
Pulse (/min), median (IQR)	109 (96, 122)	112 (96, 122)	.88
Respiratory rate (/min), median (IQR)	20 (18, 22)	18 (18, 20)	.17
Systolic blood pressure (mmHg), median (IQR)	118 (107, 135)	117 (108, 132)	.91
Diastolic blood pressure (mmHg), median (IQR)	72 (63, 84)	70 (64, 81)	.33
Features of FN at ED presentation
Absolute WBC count (×103 cells/mm3), median (IQR)	0.8 (0.4, 1.3)	0.9 (0.6, 1.2)	.29
Absolute neutrophil count (×103 cells/mm3), median (IQR)	0.1 (0, 0.3)	0.1 (0, 0.3)	.81
Temperature (°C), median (IQR)	38.2 (37.2, 38.8)	38.0 (37.1, 38.5)	.18

IQR indicates interquartile range.

3.2. Electronic clinical practice guideline use 

Overall, the eCPG was used for 76 (37.8%) of 201 patients. Fig. 1 summarizes the proportion of eCPG use by hospital site over the 3-year study period. No evidence of eCPG use could be found at 2 control hospitals and use was low (19%) at the other control hospital. At the intervention site, most FN patients (57%) were treated using the eCPG and there was no evidence of a trend of increasing or decreasing use over time.

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 1. 

  Proportion of FN eCPG use between 2006 and 2008 according to hospital site.

3.3. Management 

Investigations such as ECG (46.9% vs 31.5%, P = .03) and blood culture (96.1% vs 93.1%, P = .04) were performed more frequently in patients presenting to the intervention hospital. The most common antibiotic administered at the intervention hospital was piperacillin-tazobactam, whereas ceftazidime was the most commonly used among the control hospitals. Emergency physicians used G-CSF for approximately 1 in 4 FN patients; however, the percentage of patients receiving the agent did not differ between intervention and control sites (28.9% vs 21.9%, P = .27). Of note, the FN eCPG prompts the EP to order G-CSF after consultation with the hematologist or infectious diseases physician. The FN eCPG also provides access to specific G-CSF information to the EP at the point of ordering.

Fig. 2 summarizes ED times for FN patients. Times from triage to room placement and from triage to physician assessment were not significantly different. However, at the intervention hospital, there were reductions in median time from triage to first consultation (3.8 vs 5.0 hours, P = .001) and time from triage to first antibiotic (3.9 vs 4.9 hours, P = .02). The subgroup of physicians at the intervention hospital who elected to use the eCPG achieved slightly different median times from triage to first antibiotic compared to the subgroup of physicians not using the eCPG (3.8 vs 4.2 hours); however, this difference was not statistically significant (P = .31). Table 3 also summarizes ED management for FN patients.

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 2. 

  ED times of patients with FN according to treatment at intervention vs control hospital sites.

Table 3. ED management of patients with FN comparing treatment at intervention vs control hospital sites

	Intervention hospital (n = 128)	Control hospitals (n = 73)	P
ED investigations
Blood cultures done, n (%)	123 (96.1%)	68 (93.1%)	.04⁎
ECG done, n (%)	60 (46.9%)	23 (31.5%)	.03⁎
Chest x-ray done, n (%)	112 (87.5%)	62 (84.9%)	.77
ED treatments
patients receiving piperacillin-tazobactam, n (%)	84 (65.6%)	10 (13.7%)	<.001⁎
patients receiving ceftazidime, n (%)	21 (16.4%)	44 (60.3%)	<.001⁎
patients receiving G-CSF, n (%)	37 (28.9%)	16 (21.9%)	.27

3.4. Disposition 

In both the intervention and control hospital groups, the vast majority of FN patients were either admitted to hospital or transferred to the Cross Cancer Institute. The time from triage to admission or transfer was not significantly different. Patients presenting to control hospitals were more likely to be discharged home (15.1% vs 7.0%, P = .04) (Table 4).

Table 4. Comparison of outcomes of patients with FN at intervention vs control hospital sites

	Intervention hospital (n = 128)	Control hospitals (n = 73)	P
Disposition
Admitted or transferred, n (%)	118 (92.2%)	62 (84.9%)	.47
Time from triage to admission/transfer (h), median (IQR)	14 (11.1, 16.3)	12.8 (10.1, 15.8)	.96
Discharged home, n (%)	9 (7.0%)	11 (15.1%)	.04⁎
Time from triage to discharge (h), median (IQR)	4.4 (4.3, 6.3)	6.4 (4.9, 10.9)	.14
Died in ED, n (%)	1 (0.8%)	0	N/A

^⁎ P < .05, statistically significant.

Back to Article Outline

4. Discussion 

Computer decision support systems have the potential to improve and standardize practice, especially for rare but serious conditions [12]. One such tool for clinicians are care maps derived from high-quality evidence (individual trials, systematic reviews) or combined evidence summaries from published CPGs. In this study, the penetration and effectiveness of an eCPG for FN were examined. The results illustrate there was a clear difference in eCPG utilization between the intervention hospital and the control hospitals. The increased usage at the UAH was likely the result of greater physician experience and familiarity with the eCPG application as well as recognition that these tools represented the recognized standard of care. In addition to the FN eCPG, UAH EPs routinely used eCPGs specifically designed for patients with other presenting conditions such as chest pain, asthma, or cellulitis. The clinical group of UAH EPs was also actively involved in revising existing eCPGs as well as in developing new eCPGs.

The eCPG application (Web based, modularly designed CDSS containing a collection of interactive eCPGs) was originally developed in 2000 to 2001 and made available to clinical users at the UAH in April 2001. The eCPG application was updated in October 2005 and disseminated to other sites in April 2006. During the dissemination process, all sites were provided with eCPG application access, educational presentations, and a demonstration from the development team. Unfortunately, there was no plan with appropriate resources to implement ongoing support for the clinical users of the application nor were there any resources to support any eCPG modifications to meet specific site-based needs. This may have contributed to the poorer integration of eCPGs in patient management [12]. In contrast, the greater experience with eCPGs and the ongoing interaction with the health informatics development team may have contributed to improved eCPG utilization at the intervention hospital, resulting in the large disparity observed in eCPG use between hospital sites. This finding is consistent with previous reports in the literature that CDSS implementation is more successful when the study authors are also actively involved in developing and supporting the tool [13]. In this case, clinician members of the development team were strong advocates encouraging their clinical colleagues to use the eCPG tools. Finally, eCPG champions have emerged at the UAH, whereas the other sites have no such identified or funded individuals. This may have also contributed to the lower uptake of the eCPG application at other sites.

There were no significant differences in presenting patient demographics and vital signs between the intervention and control hospital patients. This suggests that selection bias was not involved in determining which hospital site FN patients presented to. We found that the median time from triage to first antibiotic was approximately 1 hour faster at the intervention site compared to the control hospitals. This finding is both statistically significant and clinically significant. Previous studies have demonstrated that delays in antibiotic administration in serious infections result in poorer patient outcomes [14], [15]. Although no defined benchmark for acceptable time to antibiotic for FN patients has been described in the literature, the reduction of time to antibiotic administration by an hour reported in this study indicates an important improvement in clinical management [16]. In addition to reducing the time to antibiotic initiation, greater eCPG use at the UAH also correlated with greater compliance in prescribing piperacillin-tazobactam as a first-line treatment. Given that this recommendation was updated in 2005 based on a review of current evidence, local antibiotic resistance patterns, and higher complication rate among FN patients using the previously recommended combination of ampicillin and tobramycin, eCPG use at the UAH appears to have improved adherence to these recommended and evidence-based prescribing patterns [17].

We elected to analyze hospital-level data rather than the eCPG utilization data because the control and intervention hospitals were inherently different with regard to follow-up after eCPG implementation, interaction with the health informatics team, and advocacy for eCPG use. These systemic differences would have influenced patient care regardless of actual eCPG use because they contributed to greater overall familiarity with standard investigations and evidence-based treatments for FN patients. Thus, the systemic development of eCPGs and experience using these tools at the intervention hospital improved the standard of care for patients presenting with FN.

Despite the advances at the intervention hospital, the median door-to-antibiotic time of 3.9 hours still represents a substantial delay in delivering appropriate antibiotic therapy. In similar studies reporting time to antibiotic for FN patients, one study reported a median time of 210 minutes from ED arrival to antibiotic administration and a second study reported an average door-to-antibiotic time of 170 minutes in a sample of 55 patient visits to a US ED [10], [18]. Increased efforts are required to further decrease the time from triage to antibiotic administration at the UAH. The crisis of ED overcrowding experienced in many urban EDs needs to be addressed to mitigate front-end delays before ED physician assessment. This study suggests that ED staff should be further educated to improve the recognition and awareness of the critical need for prompt treatment for FN patients. Efforts are also needed to increase eCPG awareness and utilization at the control hospitals.

This study has several limitations that need to be discussed. First, the study used a retrospective chart review design and there are inherent challenges with missing information and poor documentation. Efforts were made to use accepted chart review methods, to assess outcomes that were automatically recorded in electronic ED information systems, and to examine nursing recorded antibiotic administration [19]. Second, data abstractors were not blinded to the study hypothesis; however, a standard data abstraction form was used and the timing of antibiotic delivery was unbiased [20]. Third, the study did not examine long-term outcomes of the enrolled patients because of transfer to other treatment centers such as the Cross Cancer Institute and the lack of linked follow-up data. Fourth, delays due to overcrowding are now commonplace in Canadian EDs, and differences in time from triage to room placement represent a barrier to achieving timely evidence-based care [21]. Many patients in this cohort experienced delays in room placement and while the eCPG may have helped reduce their time from placement to antibiotic, the overall influence on time from triage to antibiotic was diluted by these placement delays. Lastly, individual ED physicians may tend to use the eCPG either for all FN patients they encounter or not at all [8]. Because eCPG use is not mandatory, there may have been bias introduced based on which physicians used the eCPG and which physicians chose not to.

Back to Article Outline

5. Conclusions 

Successful implementation of eCPGs requires ongoing staff training as well as “champions” advocating for the uptake and use of these tools among their clinical colleagues. This study has shown that systemic hospital-level implementation of an FN eCPG was correlated with a reduction in time from triage to antibiotic administration. There is currently a large disparity in eCPG utilization among the hospitals included in this study, and it appears that additional knowledge dissemination efforts are warranted along with ongoing evaluation of eCPG use. Further research is required to identify and overcome current barriers to implementing all eCPGs, thereby increasing standardization and improving clinical management for patients presenting to EDs here and elsewhere.

Back to Article Outline

Appendix A: Data Collection Form 

• 
  • View Large Image
  • Download to PowerPoint

• 
  • View Large Image
  • Download to PowerPoint

• 
  • View Large Image
  • Download to PowerPoint

• 
  • View Large Image
  • Download to PowerPoint

• 
  • View Large Image
  • Download to PowerPoint

Back to Article Outline

References 

1. Klastersky J. Management of fever in neutropenic patients with different risks of complications. Clin Infect Dis. 2004;39(Suppl 1):S32–S37
   • View In Article
   • CrossRef
2. Courtney DM, Aldeen AZ, Gorman SM, Handler JA, Trifilio SM, Parada JP, et al. Cancer-associated neutropenic fever: clinical outcome and economic costs of emergency department care. Oncologist. 2007;12(8):1019–1026
   • View In Article
   • CrossRef
3. Baltic T, Schlosser E, Bedell MK. Neutropenic fever: one institution's quality improvement project to decrease time from patient arrival to initiation of antibiotic therapy. Clin J Oncol Nurs. 2002;6(6):337–340
   • View In Article
   • MEDLINE
4. Paul M, Yahav D, Fraser A, Leibovici L. Empirical antibiotic monotherapy for febrile neutropenia: systematic review and meta-analysis of randomized controlled trials. J Antimicrob Chemother. 2006;57(2):176–189
   • View In Article
   • MEDLINE
   • CrossRef
5. Marti FM, Cullen MH, Roila F ESMO Guidelines Working Group. Management of febrile neutropenia: ESMO clinical recommendations. Ann Oncol. 2009;20(Suppl 4):166–169
   • View In Article
   • CrossRef
6. Shiffman RN, Michel G, Essaihi A, Thornquist E. Bridging the guideline implementation gap: a systematic, document-centered approach to guideline implementation. J Am Med Inform Assoc. 2004;11(5):418–426
   • View In Article
   • MEDLINE
   • CrossRef
7. Bates DW, Kuperman GJ, Wang S, Gandhi T, Kittler A, Volk L, et al. Ten commandments for effective clinical decision support: making the practice of evidence-based medicine a reality. J Am Med Inform Assoc. 2003;10(6):523–530
   • View In Article
   • MEDLINE
   • CrossRef
8. Graham TA, Bullard MJ, Kushniruk AW, Holroyd BR, Rowe BH. Assessing the sensibility of two clinical decision support systems. J Med Syst. 2008;32(5):361–368
   • View In Article
   • CrossRef
9. Graham TA, Kushniruk AW, Bullard MJ, Holroyd BR, Meurer DP, Rowe BH. How usability of a web-based clinical decision support system has the potential to contribute to adverse medical events. AMIA Annu Symp Proc. 2008;257–261
   • View In Article
10. Perrone J, Hollander JE, Datner EM. Emergency department evaluation of patients with fever and chemotherapy-induced neutropenia. J Emerg Med. 2004;27(2):115–119
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (70 KB)
    • CrossRef
11. Hughes WT, Armstrong D, Bodey GP, Bow EJ, Brown AE, Calandra T, et al. 2002 guidelines for the use of antimicrobial agents in neutropenic patients with cancer. Clin Infect Dis. 2002;34(6):730–751
    • View In Article
    • CrossRef
12. Holroyd BR, Bullard MJ, Graham TA, Rowe BH. Decision support technology in knowledge translation. Acad Emerg Med. 2007;14(11):942–948
    • View In Article
    • CrossRef
13. Garg AX, Adhikari NK, McDonald H, Rosas-Arellano MP, Devereaux PJ, Beyene J, et al. Effects of computerized clinical decision support systems on practitioner performance and patient outcomes: a systematic review. JAMA. 2005;293(10):1223–1238
    • View In Article
    • CrossRef
14. Natsch S, Kullberg BJ, van der Meer JW, Meis JF. Delay in administering the first dose of antibiotics in patients admitted to hospital with serious infections. Eur J Clin Microbiol Infect Dis. 1998;17(10):681–684
    • View In Article
    • MEDLINE
    • CrossRef
15. Natsch S, Kullberg BJ, Meis JF, van der Meer JW. Earlier initiation of antibiotic treatment for severe infections after interventions to improve the organization and specific guidelines in the emergency department. Arch Intern Med. 2000;160(9):1317–1320
    • View In Article
    • MEDLINE
    • CrossRef
16. Corey AL, Snyder S. Antibiotics in 30 minutes or less for febrile neutropenic patients: a quality control measure in a new hospital. J Pediatr Oncol Nurs. 2008;25(4):208–212
    • View In Article
    • CrossRef
17. Paul M, Soares-Weiser K, Grozinsky S, Leibovici L. Beta-lactam versus beta-lactam–aminoglycoside combination therapy in cancer patients with neutropaenia. Cochrane Database Syst Rev. 2003;(3):CD003038
    • View In Article
18. Nirenberg A, Mulhearn L, Lin S, Larson E. Emergency department waiting times for patients with cancer with febrile neutropenia: a pilot study. Oncol Nurs Forum. 2004;31(4):711–715
    • View In Article
19. Worster A, Haines T. Advanced statistics: understanding medical record review (MRR) studies. Acad Emerg Med. 2004;11(2):187–192
    • View In Article
    • MEDLINE
    • CrossRef
20. Randolph AG, Haynes RB, Wyatt JC, Cook DJ, Guyatt GH. Users' guides to the medical literature: XVIII. How to use an article evaluating the clinical impact of a computer-based clinical decision support system. JAMA. 1999;282(1):67–74
    • View In Article
    • MEDLINE
    • CrossRef
21. Bond K, Ospina MB, Blitz S, Afilalo M, Campbell SG, Bullard M, et al. Frequency, determinants and impact of overcrowding in emergency departments in canada: a national survey. Healthc Q. 2007;10(4):32–40
    • View In Article