Placing Physician Orders at Triage: The Effect on Length of Stay

Transcription

1 HEALTH POLICY AND CLINICAL PRACTICE/ORIGINAL RESEARCH Placing Physician Orders at Triage: The Effect on Length of Stay Stephan Russ, MD, MPH, Ian Jones, MD, Dominik Aronsky, MD, PhD, Robert S. Dittus, MD, MPH, Corey M. Slovis, MD From the Department of Emergency Medicine (Russ, Jones, Aronsky, Slovis), the Department of Biomedical Informatics (Aronsky), and the Department of Medicine, Division of General Internal Medicine and Center for Health Services Research (Dittus), Vanderbilt University Medical Center, Nashville, TN; and the Department of Veterans Affairs, National Quality Scholars Fellowship Program, Tennessee Valley Healthcare System, Nashville, TN (Russ, Dittus). Study objective: Emergency department (ED) crowding is a significant problem nationwide, and numerous strategies have been explored to decrease length of stay. Placing a physician in the triage area to rapidly disposition low-acuity patients and begin evaluations on more complex patients is one strategy that can be used to lessen the effect of ED crowding. The goal of this study is to assess the effect of order placement by a triage physician on length of stay for patients ultimately treated in a bed within the ED. Methods: We conducted a pre-experimental study with retrospective data to evaluate patients with and without triage physician orders at a single academic institution. A matched comparison was performed by pairing patients with the same orders and similar propensity scores. Propensity scores were calculated with demographic and triage data, chief complaint, and ED capacity on the patient s arrival. Results: During the 23-month study period, a total of 66,909 patients were sent to the waiting room after triage but still eventually spent time in an ED bed. A quarter of these patients (23%) had physician orders placed at triage. After a matched comparison, patients with triage orders had a 37-minute (95% confidence interval 34 to 40 minutes) median decrease in time spent in an ED bed, with an 11-minute (95% confidence interval 7 to 15 minutes) overall median increase in time until disposition. Conclusion: Our study suggests that early orders placed by a triage physician have an effect on ED operations by reducing the amount of time patients spend occupying an ED bed. [Ann Emerg Med. 2010;56:27-33.] Please see page 28 for the Editor s Capsule Summary of this article. Provide feedback on this article at the journal s Web site, /$-see front matter Copyright 2009 by the American College of Emergency Physicians. doi: /j.annemergmed INTRODUCTION Background Emergency department (ED) crowding is a significant problem within the United States. 1-3 One of the many factors contributing to ED crowding is the practice of boarding patients already admitted to the hospital. 4-7 The result of blocking ED rooms with admitted patients is a constraint on patient flow as physical space and nursing staff become occupied with providing care to admitted patients. Furthermore, the capacity to bring new patients into the ED treatment beds decreases incrementally. The downstream effect of boarding admitted patients in the ED ultimately affects the quality of care and leads to prolonged waiting room times and increased overall length of stay One strategy used to lessen the effect of crowding is to have a physician in the triage area. The goal of this individual is to enhance flow and throughput, as well as to preserve limited physical bed space for the sickest patients. As such, the team triage physician performs multiple tasks, including treating and releasing low-acuity patients without using a traditional ED bed, managing ED flow problems, and initiating early diagnostic evaluations of patients whenever possible. 13 Importance Previous studies have examined the effect of implementing a triage physician program on overall composite length of stay; however, none of the studies have specifically explored the operational outcome of initiating diagnostic testing before placing the patient in an ED bed Determining the effect of early testing on length of stay is essential because such an initiative may result in a number of unintended effects. Using a triage physician with patients destined for treatment by another ED physician adds complexity by involving another provider and an additional handoff in a patient s care. Another concern is that patients in the waiting room can leave the ED unnoticed more easily than can those in a regular ED bed, potentially resulting in additional follow-up requirements because of incomplete testing. If a patient leaves before being Volume 56, NO. 1 : July 2010 Annals of Emergency Medicine 27

2 Placing Physician Orders at Triage Editor s Capsule Summary What is already known on this topic Beginning treatment before emergency department (ED) room assignment is one approach to avoiding delays caused by crowding. What question this study addressed Does placing a physician at triage to write initial orders alter the overall length of stay for patients who eventually are treated in an ED bed? What this study adds to our knowledge In this single-site pre-experimental design, patients who had triage orders authored by a physician spent 11 minutes more overall in their entire ED stay but had less of the care in an ED bed when compared to matched patients absent that benefit. How this might change clinical practice This response to crowding can alter care metrics slightly; however, its effect on overall care and patient outcomes is unknown. brought to a traditional treatment area, the possibility exists for reporting of abnormal laboratory or radiographic results in a patient who is no longer present. The extent to which an obligation exists to track these results and inform the patients remains unclear and can result in considerable expenditure of staff resources. Last, dedicated support staff are necessary to support the physician triage process. The potential advantages of early ordering center on decreasing the time patients spend occupying an ED bed and decreasing the time a patient waits to consult a physician. Nursing care and physical space are the 2 important components of ED flow that are affected by the practice of boarding. Moving patients through the available beds more rapidly could lead to increased productivity despite constrained resources. Having an initial round of diagnostic tests completed or pending when a patient is placed in an ED treatment bed does not, however, necessarily guarantee faster throughput. Time involved in the physician s reevaluating the patient, further diagnostic testing, implementing therapeutic interventions, initiating consultations by other providers, and nursing tasks such as charting may all eliminate the initial potential time savings. Thus, the effect and contribution of the physician triage process on length of stay remain unknown. Goals of This Investigation The goal of our study was to assess the effect of triage physician early order placement on waiting room and ED bed length of stay for patients who are not discharged from the waiting room. Russ et al MATERIALS AND METHODS Study Design We conducted a pre-experimental study using a matched design and data collected retrospectively from an integrated ED information system that includes an electronic whiteboard, 21 a computerized provider order entry system, 22 and an electronic medical record. 23 Patient visits in which a physician initiated orders at triage were matched to similar visits with no physician orders before placement in an ED bed. The study was given exempt status by the institutional review board. Setting The study was conducted at an urban academic tertiary care medical center during a 23-month period. The medical center has an emergency medicine residency program and does not use midlevel providers or have a dedicated fast track area. Triage nurses do not use standing bundled order sets. The duration of the study, from February 1, 2007, to December 31, 2008, was determined by stability in the data sources and no significant changes in staffing, physical resources, or operational procedures within the ED. During the study period, the ED was frequently at or near capacity, with 42% of patient arrivals occurring at times when occupancy rates exceeded 90%. An attending physician, additional patient care technician, and a paramedic or nurse were assigned exclusively to the triage area 5 to 6 days a week, for 12 hours a day (11 AM to 11 PM). This time coincided with hours of peak patient arrival to the ED. During times when the ED had a high number of boarded inpatients but no physician assigned to triage, one or more of the regularly scheduled ED physicians would frequently rotate between the core portion of the ED and the triage area to at least partially fill the triage role. Selection of Participants The potential matching pool included all patient visits during the study period that had a waiting room stay greater than 5 minutes but still had eventual evaluation in a bed within the ED (Figure). For the purposes of the study, an ED hallway bed or other overflow bed was treated as a regular ED bed. Patients who left before being seen by a physician, left against medical advice, or left before formal discharge were excluded and represented only a small portion of the overall patient volume (3.1%). Patients younger than 18 years were excluded because they are transferred to an adjacent dedicated pediatric ED. Interventions Board-certified emergency physicians began working regularly scheduled, dedicated triage shifts 18 months before the onset of the study. The physician s role in triage has been described in more detail elsewhere 16,18 and generally involved the performance of medical screening examinations and discharge for nonurgent patients, treatment and release of low- 28 Annals of Emergency Medicine Volume 56, NO. 1 : July 2010

3 Russ et al Placing Physician Orders at Triage through a computerized provider order entry system that was used to identify and match patients according to individual orders. Data extracted from the computerized provider order entry and ED information system were stored in an Oracle 10.2g database (Oracle Corporation, Redwood Shores, CA). Figure. Selection of patient visits for inclusion in the study. CPOE, Computerized physician order entry. acuity patients, and initiation of diagnostic testing for select patients. A registered nurse still performed the official tasks of triage such as obtaining vital signs, recording medications, and social screening on every patient, regardless of the presence of a physician in the triage area. Only a portion of patients who were ultimately treated in a core ED bed had physician evaluation and order entry at triage because the triage physicians had multiple responsibilities, including treating and discharging low-acuity patients. Patients without initial physician orders at triage served as the control group for the study. Data Collection and Processing We used timestamps from the ED information system to calculate time spent in the waiting room and overall length of stay according to disposition. Waiting room time began at registration, not initial triage. Disposition was defined as discharge from the electronic whiteboard for treat-and-release patients or placement of an electronic bed request for admitted patients. Demographic covariates or timestamps were missing for 58 patient visits that were subsequently excluded from the final data set. Triage vitals, Emergency Severity Index acuity score 24 chief complaint, ED occupancy at arrival, average waiting room times at arrival, and number of boarded patients at arrival were obtained from the ED information system. All orders, including medications, laboratory studies, radiographic studies, requests for consultations, and nursing orders, are processed Primary Data Analysis The primary outcome variables were the differences in waiting room time, time spent in an ED treatment bed, and overall length of stay until disposition between patients with and without physician orders at triage. Patient visits with at least 1 physician order in the computerized provider order entry system before the patient was placed in an ED bed were matched to patient visits with identical overall orders but no orders present before transfer from the waiting room to an ED bed. The possible match results were restricted to patient visits that occurred when the ED was at similar capacity, as defined by an occupancy rate within 20% of one another. The occupancy rate difference was measured as an absolute, rather than a relative percentage, and the value of 20% was a priori chosen arbitrarily. The orders used for matching purposes included orders for all radiology and laboratory tests, the presence of an electronic request for consultations, and extended duration nursing orders such as those for intravenous fluids. Already existing standing nursing protocol orders, including ECGs, pregnancy tests, urinalysis, and non-narcotic analgesics entered before patient placement in an ED bed, were not considered as triage physician orders, even if a physician entered them into the computerized provider order entry system. The initial match by computerized provider order entry and ED occupancy produced clusters of matched patient visits that were then further paired on a one-to-one basis by using a propensity score calculated on the likelihood that a patient was going to receive orders by a physician at triage. The propensity score was calculated with multivariate logistic regression. The components of the propensity score were selected a priori and included age, sex, chief complaint, initial vital signs, Emergency Severity Index triage score, the percentage occupancy of the ED at triage, the number of boarded inpatients at triage, the mean waiting room length of stay at triage, time of day, weekend/ weekday, and the presence or absence of a flag indicating concern for possible acute coronary syndrome by the triage nurse. The nearest-neighbor match on propensity score was used to produce matched pairs for evaluation. The variables chosen for inclusion in the propensity score represent data that are reliably present within the ED information system and have been found to influence the primary outcome variables. Data on ED capacity were used as a continuous variable in the propensity score calculation. The chief complaints in the ED information system were selected from a discrete list of complaints by the triage nurse and were handled as a categorical variable. Arrival time of day was categorized in 3 groups: 8 AM to 4 PM,4PM to midnight, and midnight to 8 AM. Pulse rate was categorized in 4 groups: less than 50, 51 to 100, 101 to 120, and greater than 120 beats/min. Systolic blood pressure was categorized in 4 groups: less Volume 56, NO. 1 : July 2010 Annals of Emergency Medicine 29

4 Placing Physician Orders at Triage Russ et al Table 1. Descriptive statistics for the unmatched and matched cohorts. Descriptive Statistics Absent (n 51,627) Before Matching (95% CI) Present (n 15,282) Absent (n 10,901) After Matching (95% CI) Present (n 10,901) ED conditions at patient registration Median number of patients registered 30 (30 30) 36 (35 37) 34 (34 34) 35 (34 35) in the previous 4 h Median percentage occupancy of ED 85 (85 85) 95 (95_95) 94 (94 94) 95 (95 95) beds Median waiting room count 3 (3 3) 8 (8 8) 8 (8 8) 8 (8 8) Median number of boarded patients 11 (11 11) 15 (15 15) 14 (14 14) 15 (15 15) Patient characteristics Percentage of patients arriving on a 35.3 ( ) 8.6 ( ) 13.2 ( ) 10.3 ( ) weekend Median age, y 40 (40 40) 40 (39 40) 37 (37 38) 36 (36 37) Percentage of patients with HR ( ) 12.7 ( ) 11.5 ( ) 10.9 ( ) beats per min at triage Percentage of patients with a 26.4 ( ) 25.8 ( ) 21.0 ( ) 21.0 ( ) consulting service Percentage of patients with a billable 7.9 ( ) 5.6 ( ) 6.5 ( ) 6.7 ( ) ED procedure Percentage of patients admitted 26.7 ( ) 28.3 ( ) 21.9 ( ) 21.9 ( ) Percentage of patients with chief complaint of Abdominal pain 13.0 ( ) 17.8 ( ) 17.1 ( ) 18.6 ( ) Chest pain 8.6 ( ) 11.0 ( ) 8.5 ( ) 8.4 ( ) Shortness of breath 4.7 ( ) 4.3 ( ) 4.3 ( ) 4.0 ( ) Generalized weakness 7.0 ( ) 7.4 ( ) 6.8 ( ) 6.0 ( ) Percentage of patients with orders for Quantitative d-dimer 3.3 ( ) 4.4 ( ) 2.2 ( ) 2.2 ( ) Troponin 15.6 ( ) 15.7 ( ) 11.0 ( ) 11.0 ( ) Abdominal/pelvic CT with contrast 10.9 ( ) 13.7 ( ) 11.3 ( ) 11.3 ( ) Chest radiograph (PA/Lat) 30.6 ( ) 34.8 ( ) 30.8 ( ) 30.8 ( ) Intravenous fluid bolus 28.8 ( ) 30.2 ( ) 26.9 ( ) 26.9 ( ) than 90, 91 to 110, 111 to 180, and greater than 180 mm Hg. Respiratory rate was categorized in 3 groups: less than 11, 12 to 20, and greater than 20 breaths/min. Temperature was categorized in 2 groups: less than 38 C (100.4 C) and greater than 38 C (100.4 C) and oxygen saturation in 2 groups as well: less than 93% and greater than 93%. Vital signs were handled as categorical rather than continuous linear variables because of the wide normal ranges and the inherent clinical differences present between low (eg, hypotension) versus high (eg, hypertension) outlying values. Patients were matched by propensity score because it is a common technique for reducing selection bias in observational studies. 25,26 The potential for selection bias in our study occurred because patients were not randomly assigned to the exposure or control groups. The propensity score is the conditional probability of assignment to a particular exposure (in this case, triage physician orders), given a number of observed covariates. Matching patients with similar propensity scores helps to balance these covariates between the exposed and control groups. All data reporting and hypothesis testing were performed with nonparametric methods. The unit of analysis was the patient visit. We compared unmatched outcome variables with a 2-sample Hodges-Lehmann estimation method to calculate the median of the differences and 95% confidence intervals (CIs). 27 CIs for median differences in paired outcome variables were estimated with bootstrapping, with 1,000 iterations. 28 Statistical analysis was performed with Stata (version 10.0; StataCorp, College Station, TX). RESULTS During the 23-month study period, 104,386 patients were treated and formally discharged or admitted. A total of 66,909 patient visits (Figure) met the study criteria of being sent to the waiting room after triage but eventually spending time in an ED bed. In 23% of patient visits, the triage physician entered orders while the patient was in the waiting room. Of the initial 15,282 eligible patient visits with physician orders at triage, 10,901 had matched patients with identical computerized provider order entry orders in the group that had no orders placed by a triage physician. Selected descriptive statistics for the matched and unmatched cohorts are presented in Table 1. Patients with physician orders placed while the patient was in the waiting room arrived during 30 Annals of Emergency Medicine Volume 56, NO. 1 : July 2010

5 Russ et al Placing Physician Orders at Triage Table 2. Unmatched length of stay for patients sent to the waiting room after triage and eventually treated in an ED bed. Length of Stay Absent (n 51,627) Present (n 15,282) Median Difference (95% CI) Waiting room time, min, median (55 to 57) Time in ED bed, min, median ( 16 to 11) Total length of stay until disposition, min, median (48 to 54) Table 3. Length of stay for patients matched by order set and propensity score. Length of Stay Absent (n 10,901) Present (n 10,901) Median Difference (95% CI) Waiting room time, min, median (39 to 43) Time in ED bed, min, median ( 40 to 34) Total length of stay until disposition, min, median (7 to 15) Table 4. Length of stay for patients matched by order set and propensity score, stratified by disposition. Length of Stay Absent (n 10,901) Present (n 10,901) Median Difference With 95% CIs Discharged Waiting room time, min, median (41 to 45) Time in ED bed, min, median ( 35 to 41) Total length of stay until disposition, min, median (5 to 13) Admitted Waiting room time, min, median (32 to 39) Time in ED bed, min, median ( 19 to 37) Total length of stay until disposition, min, median (13 to 32) busier periods of the day and were less likely to be treated on weekends. Patients with triage physician orders were also more likely to have a chief complaint of chest pain or abdominal pain, as well as to receive a computed tomography (CT) scan of the abdomen/pelvis or chest radiograph. The matching procedure resulted in a reduction in these differences. An unmatched comparison of length of stay for patient visits with and without physician orders is presented in Table 2. Final matched pairs were created with propensity scores, with the resulting differences in length of stay displayed in Table 3. After matching, the median time in an ED bed changed from 13 fewer minutes in the triage physician order group to 37 fewer minutes. The difference in waiting room times decreased from 56 additional minutes for the triage physician order group before matching to 41 additional minutes after matching. The overall length of stay until disposition for the triage physician order group was reduced from 51 minutes to 11 minutes. The 37 fewer median minutes spent in an ED bed by the triage physician order group represents an 18% decrease in ED bed length of stay compared with the 212 median minutes spent in an ED bed by the control group. A comparison of length of stay for the matched pairs stratified by disposition is displayed in Table 4. Patients ultimately discharged had a longer waiting room length of stay than admitted patients in the physician order group but spent even less time in an ED bed, which resulted in a significant difference in overall length of stay until disposition between the admitted and discharged groups. LIMITATIONS Several limitations need to be considered. The study was performed at a single academic institution, which limits the generalizability of the findings. The patient visits were categorized by the presence of physician orders at triage, but this does not take into account whether those orders were subsequently acted on or completed before a patient s placement in an ED bed. Orders that were placed in triage but not acted on before moving the patient to an ED bed would bias the study toward the null hypothesis and decrease the observed difference between the 2 groups. The study was retrospective and therefore has the potential for greater confounding and bias than a prospective study. Only a portion of patients had physician evaluation and order entry at triage because the triage physicians had multiple responsibilities, including treating and discharging of low-acuity patients. Patients who were treated by physicians in triage were not randomly selected, thus introducing a potential selection bias. We attempted to mitigate this bias by exactly matching patients on computerized provider order entry orders and the requirement that matched pairs arrive when the ED had similar occupancy. We attempted to correct for occupancy-dependent performance variation by the physicians and other elements of the ED system by setting the maximum allowable difference in ED occupancy between matched patients to 20%. The goal of computing a propensity score was to further strengthen the matching procedure and reduce the effect of a potential Volume 56, NO. 1 : July 2010 Annals of Emergency Medicine 31

6 Placing Physician Orders at Triage selection bias. The study does not address the implications of the triage physician intervention on quality of care other than length of stay. DISCUSSION Our study found that the use of a physician providing and initiating care in the triage area resulted in an overall decrease in time patients spent in an ED bed, a scarce resource in crowded EDs. When including time spent in the waiting room, patients with triage physician orders had a marginal increase ( 5%) in overall time until disposition in a matched comparison. The effect of physician orders at triage was greater on patients ultimately discharged, rather than admitted. Optimal care for arriving ED patients traditionally entails being treated rapidly by a physician in a private ED bed to which they are assigned throughout their stay. Unfortunately, because many EDs have a high census of admitted patients waiting for an inpatient bed, innovative processes must be explored to provide the best care possible with the few remaining beds available to ED patients. Time spent occupying an ED bed is an important throughput measure during conditions of high ED census. Our finding of a decrease in ED bed length of stay of roughly 20% for patients with triage physician orders can be thought of as having creating one transient virtual bed for every 5 patients with orders placed. The increased waiting room time for patients with triage physician orders in this study is likely a result of the common practice among some of the physicians to hold patients in the waiting room until diagnostic testing is well under way, thereby allowing more rapid access to ED beds for patients who have not yet been evaluated. This practice benefits other patients and overall combined length of stay in a manner not captured by our study. Our study differs from other published literature because it focuses on the effect of early physician ordering, rather than the physician triage concept as a whole. Previous studies focused on changes in the overall length of stay, without reporting changes to actual time in an ED bed, which is an important throughput aspect during crowding The studies included patients who were treated and released from the waiting room in the lengthof-stay analysis. Many of the treat-and-release patients receive no orders, and their rapid turnover makes the effects of physician triage orders challenging to quantify. Our study also differs from previous reports because we used a longer period of 23 months, which may be less subject to temporal trends. We examined the physician triage concept beginning 18 months after implementation, which allowed the process to stabilize and be less subject to operational adjustments and the Hawthorne effect, which may have occurred during the early course of the intervention. In our study, the triage physicians did not represent additional staffing in comparison to the control group, and the staffing pattern was constant throughout the course of the 23-month study period. Overall, the study s findings suggest that early orders placed by a triage physician can significantly influence ED bed utilization. Utilization of physician orders at triage allows for more patients to be treated in conditions with large numbers of boarded patients by limiting the amount of time patients spend occupying an ED bed. During times of increased boarding of admitted patients, available beds become a valuable resource that must be used more efficiently to maintain throughput. Further efforts are needed to identify which patients and ED conditions are affected the most from this type of process change to optimize its benefit, given ever-present limitations in resources. Supervising editor: Donald M. Yealy, MD Author contributions: SR and RSD conceived and designed the study. SR performed the data analysis and statistics and drafted the initial article. IJ and DA created the data collection systems. IJ, DA, and CMS contributed substantially to data collection systems revision. IJ and CMS implemented the described intervention. SR takes responsibility for the paper as a whole. Funding and support: By Annals policy, all authors are required to disclose any and all commercial, financial, and other relationships in any way related to the subject of this article that might create any potential conflict of interest. See the Manuscript Submission Agreement in this issue for examples of specific conflicts covered by this statement. Dr. Russ was supported by the Department of Veterans Affairs Office of Academic Affiliations, with resources of the VA Tennessee Valley Healthcare System, Nashville, TN. Publication dates: Received for publication October 30, Revisions received December 30, 2009, and January 28, Accepted for publication February 3, Available online March 16, Reprints not available from the authors. Russ et al Address for correspondence: Stephan Russ, MD, MPH, Department of Emergency Medicine, Vanderbilt University Medical Center, 703 Oxford House, Nashville, TN ; , fax ; stephan.russ@ vanderbilt.edu. REFERENCES 1. Moskop JC, Sklar DP, Geiderman JM, et al. Emergency department crowding, part 1: concept, causes, and moral consequences. Ann Emerg Med. 2009;53: Trzeciak S, Rivers EP. Emergency department overcrowding in the United States: an emerging threat to patient safety and public health. Emerg Med J. 2003;20: Hoot NR, Aronsky D. Systematic review of emergency department crowding: causes, effects, and solutions. Ann Emerg Med. 2008; 52: Derlet RW. Overcrowding in emergency departments: increased demand and decreased capacity. Ann Emerg Med. 2002;39: Moskop JC, Sklar DP, Geiderman JM, et al. Emergency department crowding, part 2: barriers to reform and strategies to overcome them. Ann Emerg Med. 2009;53: Annals of Emergency Medicine Volume 56, NO. 1 : July 2010

7 Russ et al Placing Physician Orders at Triage 6. Carr BG, Hollander JE, Baxt WG, et al. Trends in boarding of admitted patients in US emergency departments J Emerg Med. In press. 7. Olshaker JS, Rathlev NK. Emergency department overcrowding and ambulance diversion: the impact and potential solutions of extended boarding of admitted patients in the Emergency Department. J Emerg Med. 2006;30: Bernstein SL, Aronsky D, Duseja R, et al. The effect of emergency department crowding on clinically oriented outcomes. Acad Emerg Med. 2009;16: McCarthy ML, Zeger SL, Ding R, et al. Crowding delays treatment and lengthens emergency department length of stay, even among high-acuity patients. Ann Emerg Med. In press. 10. Diercks DB, Roe MT, Chen AY, et al. Prolonged emergency department stays of non-st-segment-elevation myocardial infarction patients are associated with worse adherence to the American College of Cardiology/American Heart Association guidelines for management and increased adverse events. Ann Emerg Med. 2007;50: Pines JM, Hollander JE. Emergency department crowding is associated with poor care for patients with severe pain. Ann Emerg Med. 2008;51: Pines JM, Hollander JE, Localio AR, et al. The association between emergency department crowding and hospital performance on antibiotic timing for pneumonia and percutaneous intervention for myocardial infarction. Acad Emerg Med. 2006;13: Wiler JL, Gentle C, Halfpenny JM, et al. Optimizing emergency department front-end operations. Ann Emerg Med. In press. 14. Chan TC, Killeen JP, Kelly D, et al. Impact of rapid entry and accelerated care at triage on reducing emergency department patient wait times, lengths of stay, and rate of left without being seen. Ann Emerg Med. 2005;46: Choi YF, Wong TW, Lau CC. Triage rapid initial assessment by doctor (TRIAD) improves waiting time and processing time of the emergency department. Emerg Med J. 2006;23: ; discussion Han JH, France DJ, Levin SR, et al. The effect of physician triage on emergency department length of stay. J Emerg Med. In press. 17. Holroyd BR, Bullard MJ, Latoszek K, et al. Impact of a triage liaison physician on emergency department overcrowding and throughput: a randomized controlled trial. Acad Emerg Med. 2007;14: Partovi SN, Nelson BK, Bryan ED, et al. Faculty triage shortens emergency department length of stay. Acad Emerg Med. 2001;8: Subash F, Dunn F, McNicholl B, et al. Team triage improves emergency department efficiency. Emerg Med J. 2004;21: Terris J, Leman P, O Connor N, et al. Making an IMPACT on emergency department flow: improving patient processing assisted by consultant at triage. Emerg Med J. 2004;21: Aronsky D, Jones I, Lanaghan K, et al. Supporting patient care in the emergency department with a computerized whiteboard system. J Am Med Inform Assoc. 2008;15: Aronsky D, Johnston PE, Jenkins G, et al. The effect of implementing computerized provider order entry on medication prescribing errors in an emergency department. AMIA Annu Symp Proc. 2007; Giuse DA. Supporting communication in an integrated patient record system. AMIA Annu Symp Proc. 2003; Tanabe P, Gimbel R, Yarnold PR, et al. Reliability and validity of scores on the Emergency Severity Index version 3. Acad Emerg Med. 2004;11: Rosenbaum PR, Rubin DB. The central role of the propensity score in observational studies for causal effects. Biometrika. 1983;70: Glynn RJ, Schneeweiss S, Sturmer T. Indications for propensity scores and review of their use in pharmacoepidemiology. Basic Clin Pharmacol Toxicol. 2006;98: Hodges JL Jr, Lehmann EL. Estimates of location based on rank tests. Ann Math Stat. 1963;34: Efron B. Nonparametric standard errors and confidence intervals. Can J Stat. 1981;9: Request for Abstracts for ACEP s Research Forum (non-moderated) Researchers have a unique opportunity to showcase emergency medicine research published or presented in other specialties journals or meetings in the past year. This is an excellent venue to share outstanding emergency medicine research from competing meetings or journals. ACEP s Research Forum is providing emergency medicine researchers with another opportunity this year to present scientific emergency medicine research at the 2010 conference, which will be held September 28-29, 2010 in conjunction with Scientific Assembly in Las Vegas, NV. Abstracts from emergency physicians who have presented or published in non-emergency medicine specialty meetings or journals within the past 12 months will be considered. Case reports or subject reviews are not considered original research. These abstracts will be accepted on a space-available basis as non-moderated posters. If accepted the presenter is obligated to be available to discuss their poster(s) with Research Forum attendees. Please submit your research abstract(s) to the academic affairs department by Friday, August 20, 2010 at academicaffairs@acep.org, or by fax at Notifications will be sent by August 31, For questions, please call , ext Volume 56, NO. 1 : July 2010 Annals of Emergency Medicine 33