- Email: [email protected]
Resident-Initiated Advanced Triage Effect on Emergency Department Patient Flow
The Journal of Emergency Medicine, Vol. 45, No. 5, pp. 746–751, 2013 Published by Elsevier Inc. Printed in the USA 0736-4679/$ - see front matter
http://dx.doi.org/10.1016/j.jemermed.2013.03.019
Administration of Emergency Medicine
RESIDENT-INITIATED ADVANCED TRIAGE EFFECT ON EMERGENCY DEPARTMENT PATIENT FLOW Irina Svirsky, MD,* Lisa R. Stoneking, MD,* Kristi Grall, MD,* Matthew Berkman, MD,* Uwe Stolz, PHD, MPH,* and Farshad Shirazi, MD, PHD*†‡ *EM Residency, PGY II, †University of Arizona College of Medicine, South Campus, Tucson, Arizona, and ‡Arizona Poison and Drug Information Center, Tucson, Arizona Reprint Address: Irina Svirsky, MD, University of Arizona College of Medicine, South Campus, 6790 E Calle La Paz apt 3203, Tucson, AZ 85715
, Abstract—Background: Emergency Department (ED) overcrowding is a national problem. Initiating orders in triage has been shown to decrease length of stay (LOS), however, nurse, physician assistant, and attending physician advanced triage have all been criticized. Study Objectives: Our primary objective was to show that Emergency Medicine resident-initiated advanced triage shortens patient LOS. Our secondary objective was to evaluate whether or not resident triage decreases the number of patients who left prior to medical screening (LPTMS). Methods: This prospective interventional study was performed in a 42-bed, Level III trauma center, academic ED in the United States, with an annual census of approximately 41,000 patients. A junior or senior Emergency Medicine resident initiated orders on 16 weekdays for 6 h daily on patients presenting to triage. Patients evaluated during the 6-h period on other weekdays served as the control. The study was powered to detect a reduction in LOS of 45 min. Multivariable median regression was used to compare length of stay and Fisher’s exact test to compare proportions. Results: There were 1346 patients evaluated in the ED during the intervention time. Regression analysis showed a 37-min decrease in median LOS for patients on intervention days as compared to control days (p = 0.02). The proportion of patients who LPTMS was not statistically different (p = 0.7) for intervention days (96/1346, 7.13%) compared to control days (136/1810, 7.51%). Conclusions: Residentinitiated advanced triage is an effective method to decrease patient LOS, however, our effect size is smaller than
predicted and did not significantly affect the percent of patients leaving before medical screening. Published by Elsevier Inc. , Keywords—advanced triage; physician triage; patient flow; length of stay
INTRODUCTION Many interventions have been done to improve Emergency Department (ED) patient wait times and reduce length of stay (LOS), but some patients still leave before being evaluated. One method to improve ED flow is initiating orders on patients in triage. Various interventions have been proposed, such as having a nurse, physician assistant, or an attending physician in triage (1–4). Most research to date has been done on nurse-initiated triage, including focused radiology studies, administering analgesics and antipyretic medications, and nurse initiated ‘‘order sets’’ based on chief complaint (2,5). Although initiating orders in triage has been shown to decrease LOS by as much as 37–76 min, advanced triage has been criticized for various reasons (2,6). Nurse triage has been mostly criticized for either under-ordering or over-ordering studies. Allerston and Justham’s nurseinitiated triage study found that the final diagnosis of fracture was missed in 4/187 patients with complaint of ankle
RECEIVED: 4 May 2012; FINAL SUBMISSION RECEIVED: 18 November 2012; ACCEPTED: 15 March 2013 746
Resident-Initiated Advanced Triage
pain when X-ray studies were not ordered based on the Ottawa Ankle Rule Protocol that nurses were supposed to have used in triage (7). Alternatively, Lee et al.’s study cited 5.4% of X-ray studies ordered by nurses in triage as unnecessary (8). Thus, this type of advanced triage has been linked to unnecessary radiation. On the other hand, Wiler et al.’s summary of the present advanced triage articles shows that having an attending physician is another approach that might reduce LOS and improve patient satisfaction, but is not always feasible due to the associated increased cost of having an extra attending physician on shift, legal risks, need for increased staff, practice variation, and risk tolerance (9). Furthermore, Rowe’s systemic review in 2011 evaluated 28 eligible studies on physician-advanced triage. Based on 19 studies, physician-initiated triage has been shown to decrease patient LOS by 37 min on average (10). However, all studies except Porter’s study had an attending physician in triage (11). The main disadvantage is that all other methods of triage (including nurse, attending, or physician assistant triage) either circumvented or diminished resident involvement and educational opportunity. A thorough literature search using the Cochrane, MEDLINE, ACP Journal Club, DARE, and PubMed databases from years 1990–2012 for keywords such as: ‘‘resident initiated,’’ ‘‘advanced triage,’’ ‘‘length of stay,’’ ‘‘emergency department flow,’’ and ‘‘patient flow’’ yielded many advanced triage research protocols, but only one publication on resident-initiated advanced triage. Porter’s abstract in Western SAEM (Society for Academic Medicine) in 2009 failed to show that placing a post-graduate year (PGY) 3 resident in triage improves patient satisfaction, LOS, or the number of patients who left without being seen (11). We therefore attempted to demonstrate that resident-initiated advanced triage can be a feasible and effective strategy to improve patient flow in an academic setting. Our objective was to demonstrate that Emergency Medicine resident physicianinitiated triage shortens patient LOS in the ED and decreases the number of patients who left prior to medical screening (LPTMS). MATERIALS AND METHODS This was a prospective intervention study at a community-based, academic, Level III trauma hospital. The intervention consisted of PGY2 or -3 Emergency Medicine resident (triage resident) initiating orders in triage on 16 random weekdays (Wednesday and Friday) for 6 h on patients presenting to triage during the busiest times. This was initially 12:00 pm–6:00 pm, but was changed halfway through the intervention to 2:00 pm– 8:00 pm to capture a higher number of patients. The
747
choice of weekdays was based on lack of support staff on other weekdays as well as inability to do the intervention during ‘‘teaching days’’ when only attending physicians worked in the ED. The time of intervention changed halfway through the intervention due to the observation that, at these times, empty ED beds and low patient volumes were leading to triage patients bypassing the triage area and being sent directly to an ED bed. The residents were financially compensated, and were paid the same wage no matter how many patients they evaluated. They were blinded to the outcomes being measured. Patients with psychiatric complaints and alcohol intoxication were excluded because in our electronic medical record database there is no documented time for ‘‘medical clearance,’’ and these patients have the longest LOS due to long wait times for social work or psychiatric evaluation. The control group consisted of patients evaluated during the same week and 6-h period as the intervention group but on different weekdays (Monday and Thursday) (Table 1). Per historical data, our ED evaluated the same number of patients on the 2 days we chose as control as on the days we chose to perform the intervention. The population consisted of all patients presenting to the ED during the 6-h period, including: male and female patients, ages 0–99 years, with any medical complaints, presenting by any means of transportation. The excluded population consisted of patients presenting with a psychiatric complaint or ‘‘alcohol intoxication’’ as their primary complaint due to inability to evaluate the time to medical clearance. The intervention was a junior or senior (PGY2 or PGY3) resident (triage resident) who remained in triage with the triage nurse and began their work-up after the triage nurse assessment. Verbal permission to initiate evaluation and treatment was obtained by the triage resident. Upon consent, the patients were then relocated to a separate triage area with recliner chairs and cardiopulmonary monitors where assessment was initiated. After nursing staff completed triage, each patient was temporarily placed in a separate room for the triage resident to collect further history and perform a physical examination, as well as for the nurse to draw laboratory tests and perform electrocardiograms, in compliance with the Health Insurance Portability and Accountability Act. The residents had the ability to order any test or treatment they deemed necessary for patient care, including laboratory tests, radiographic studies, or medications. Separate nursing staff was assigned to this area to collect blood and urine samples and establish intravenous access. Treatment was also initiated in this area. Radiology testing was available, with patients transported from triage to Radiology. Intervention patients were assigned to a room in the original order they were assigned by the triage nurse. Once the patient was assigned a room, the triage resident
748
I. Svirsky et al.
Table 1. Demographics Intervention (n = 1346) Gender Male Female Mean age Years (SD) Triage level 1 2 3 4 5 Disposition Admitted Discharged Transferred LPTMS or eloped (14)
median regression was used to compare LOS and Fisher’s exact test to compare proportions. RESULTS
50.50% 49.50% 44.9 (19.7) 1% 12% 80% 7% 0.50% 30.4% 58.3% 4.3% 7%
LPTMS = left prior to medical screening by a physician; eloped = left the room without Against Medical Advice paperwork, after a medical screening.
communicated with the attending physician and the main ED resident to inform them about the patient course in triage verbally as well as by a quick written note in the ‘‘SOAP’’ format (Subjective, Objective, Assessment, Plan). A triage resident also had the ability to discharge a patient from the triage area if the evaluation and treatment was complete and if the attending physician could evaluate the patient. If the patient was being discharged from triage, the resident in triage wrote an entire history and physical examination note for that patient, evaluated the patient in a separate area of triage to perform a complete physical examination, and the patient was seen by the attending physician before discharge. The ED attending was always available to the triage resident and there were no safety issues or concerns regarding patient care during the intervention period. At the end of the intervention time period, if any patients remained in the triage area, the triage resident transferred patient care to the main ED and verbally communicated with a resident and attending in the main ED as well as via a written SOAP note. The triage area then closed and patients were transferred to the main ED to any open beds. If there were no open beds, staff was assigned to the triage area until patients could be transferred to the main ED. In our study, at the end of the intervention time period, most patients could be transferred to the main ED or were discharged. The study had two outcomes: patient LOS and the percent of patients who LPTMS. We considered the LPTMS patients as those patients who registered but were not evaluated by an attending physician. The study was powered to detect a reduction in LOS of 45 min for all patients being evaluated in the ED. Multivariable
A total of 1346 patients were evaluated during the 6-h intervention period on 16 days from November 2011 to February 2012. Table 1 shows that the demographic data on intervention patients was comparable to control patients in the categories of gender, age, triage level, and the percentage of patients admitted, discharged, or transferred. Further demographic data could not be collected due to limitations in the computer charting software. However, Table 1 shows that gender, age, and Emergency Severity Index levels were similar in the two groups. As shown in Table 2, regression analysis showed a 37-min decrease in median LOS for all patients evaluated in the ED for intervention days compared to control days (95% confidence interval [CI] 4–69) after controlling for age, gender, and triage level. This decrease in LOS was comparable to the mean of other studies using advanced triage with a reduction in LOS of 18–82 min, consistent with a systematic review that evaluated nurse and physician advanced triage from 1966–2008 (9). Further subgroup analysis shows that for all admitted patients, there was no statistically significant difference in ED LOS (p = 0.66), whereas for all discharged patients, there was a trend toward a 34-min decrease in ED LOS (p = 0.07, 95% CI 3–71). The proportion of patients who LPTMS was not statistically different (p = 0.7) for intervention days (96/1346, 7.13%) compared to control days (136/1810, 7.51%) (Table 3). DISCUSSION Our study measured two outcomes: patient LOS and the number of patients who LPTMS. Patient LOS was 37 min less during the intervention 6-h period as compared to control days (p = 0.02, 95% CI 4–69). Our results are consistent with our hypothesis, that having
Table 2. Patient Length of Stay Length of Stay
Coefficient
All patients
Crude coefficient Adjusted (age, sex, triage level) Admitted patients Crude coefficient (age, sex, triage level) Discharged patients Crude coefficient (age, sex, triage level) CI = confidence interval; LOS = length of stay.
LOS (min)
95% CI
25 37
56 to 7 59 to 4
12 11 33 34
65 to 41 61 to 40 68 to 2 71 to 3
Resident-Initiated Advanced Triage
749
Table 3. Patients Left Prior to Medical Screening (LPTMS)
LPTMS Did not LPTMS Total % LPTMS (p = 0.730)
Intervention
Control
96 1250 1346 7.13%
136 1674 1810 7.51%
a resident-initiated advanced triage decreased the overall LOS for all patients seen in the ED. One of the goals of this study was to compare the triage resident efficiency to other practitioners. A literature review search published in Academic Emergency Medicine in 2011 reviewed 14,000 ‘‘nurse in triage’’ interventions, which showed that the mean reduction in LOS was 37–51 min in four of the most efficient studies. However, other studies failed to show a statistical difference when interventions were applied to subjects without suspected injury or fracture. When interventions were applied to patients suspected of having a fracture that required radiology, LOS was reduced by 19 min (12). We found no studies that measured LOS for other midlevel practitioners such as physician assistants. We aimed to compare the efficiency of PGY2 and PGY3 residents in our study; however, this was not possible as only three of the 16 shifts were covered by PGY3 level residents. In Canada, in an academic tertiary institution similar to ours, an attending physician was placed in triage for 6 weeks for 9 h daily. A median of 14 patients were seen per shift. Overall, LOS was decreased by 36 min compared with control days (4:21 vs. 4:57; p = 0.001). Patients who left without complete assessment decreased from 6.6% to 5.4% during the intervention (13). Although LOS in the above study was comparable to our intervention, there is clearly an advantage to having a senior physician in triage, including experience and safety regarding patient care. However, in our study we designed a unique intervention to involve residents in learning as well as increase their independence, while still having an attending physician present for back-up. In addition, the residents that were involved in the study had a unique perspective in learning ED flow. Safety was not an issue as all patients were seen by the attending physician either in triage or when they were moved to the main ED bed. In addition, one of the reasons the intervention was done was to identify the sicker patients first by starting their work-up, as opposed to control days, when the patients would have remained in the waiting room. If the triage nurse or resident thought the patient was critically ill, they were moved immediately to an ED bed. Further subgroup analysis demonstrated that there was no statistical difference in ED LOS for admitted patients. However, for the discharged patients, there was a trend
toward a decrease in ED LOS of 34 min, although it was not statistically significant. Failure to show a difference for the subgroups was likely due to the fact that the study was underpowered. Unfortunately, our study could not be carried out to show a statistical difference due to lack of funding. However, in the future, we hope to carry out larger studies to show a statistical difference between the two groups. For patients who LPTMS, there was no statistically significant difference between the two groups: 7.13% for the intervention group compared to 7.51% for the control group (p = 0.7). Using resident-initiated advanced triage to decrease ED LOS, we attempted to more rapidly evaluate a greater number of patients to improve ED flow. Although the ED LOS was decreased, it did not affect the number of patients who LPTMS. Therefore, the number of patients being treated and evaluated early was not large enough to make an impact on the percent of patients leaving before being evaluated by a physician. We realize that at our institution, the number of patients who LPTMS is quite high at 7%. This is a multifactorial problem that may not be fixed by decreasing patient-to-physician time. We attempted, however, to address this complex issue with this intervention. Although in our study there was no statistical difference in the number of patients leaving in the intervention group (7.13%) and the control group (7.51%), in the future we hope to perform a larger study as well as address other issues contributing to this problem. Limitations Our study showed a mean decrease of 37 min for all patients, however, it was underpowered to show a decrease in LOS for the subgroups of admitted and discharged patient. For the discharged patients, there was a trend toward 34-min decrease; however, it was not statistically significant. Our study was carried out for only 16 intervention shifts due to lack of funding; however, we are hoping to increase the power in future studies. For admitted patients, there was a small difference in the LOS that was not statistically significant. In our academic institution, the LOS is measured at the time the resident places a ‘‘bed request’’ in the computer system. This is variable, and a resident could have decided to admit a patient before a bed request was actually placed. Furthermore, admitted patients included patients admitted to our facility as well as transferred to an outside facility. Therefore, a significant amount of time is lost while waiting to get an accepting physician on the line, do a ‘‘physician-to-physician communication,’’ and facilitate transport. The 16 intervention days included in this study were non-consecutive from November 2011 to February
750
I. Svirsky et al.
2012. The study included junior and senior level residents that have different expertise levels, and those subgroups were not further analyzed due to the few shifts that the senior level residents worked. The biggest limitation was that the reduced LOS did not affect the number of patients leaving before medical screening. We hypothesize that this was due to the fact that our triage area can only encompass 5–7 patients at a time due to the limited equipment and nursing staff. We found that the highest number of patients who left without being seen was after the intervention time ended and all of the ED beds were occupied. Additional limitations included not being able to measure other outcomes such as patient satisfaction or the number of patients who left at any time during the visit. This was due to the newness of the computer software in our ED and the inability to measure this type of data. The ED had also expanded in the number of beds and was moved to a new and larger geographical area, twice the size of the old ED 3 months before the intervention. CONCLUSION Preliminary results indicate that resident-initiated advanced triage is an effective and feasible method to decrease ED LOS; however, our effect size is smaller than predicted. Our results failed to show a statistical significance among the numbers of patients leaving before medical screening. Likely, this was due to a limitation in space, equipment, nursing staff, and the study being underpowered. More research needs to be done to determine whether resident-initiated advanced triage can improve the flow of the emergency department. Acknowledgments—Jackie DeBeche, RN, MBA/HCA, Nursing Director, UAMC-South Campus, University of Arizona Emergency Medicine Residency Program, and the Residents
and Staff of University of Arizona, Emergency Residency Program.
REFERENCES 1. Rosmulder RW, Krabbendam JJ, Kerkhoff AH, Schinkel ER, Beenen LF, Luitse JS. ‘Advanced triage’ improves patient flow in the emergency department without affecting the quality of care [Dutch]. Ned Tijdschr Geneeskd 2010;154:A1109. 2. Cheung WWH, Heeney L, Pound JL. An advanced triage system. Accid Emerg Nurs 2002;10:10–6. 3. Love RA, Murphy JA, Lietz TE, Jordan KS. The effectiveness of a provider in triage in the emergency department: a quality improvement initiative to improve patient flow. Adv Emerg Nurs J 2012;34: 65–74. 4. Travers JP, Lee FC. Avoiding prolonged waiting time during busy periods in the emergency department: is there a role for the senior emergency physician in triage? Eur J Emerg Med 2006;13: 342–8. 5. Tambimuttu J, Hawley R, Marshall A. Nurse-initiated x-ray of isolated limb fractures in the emergency department: research outcomes and future directions. Aust Crit Care 2002;15:119–22. 6. Lindley-Jones M, Finlayson BJ. Triage nurse requested x rays—are they worthwhile? J Accid Emerg Med 2000;17:103–7. 7. Allerston J, Justham D. Nurse practitioners and Ottawa Ankle Rules: comparisons with medical staff in requesting X rays for ankle injured patients. Accid Emerg Nurs 2000;8:110–5. 8. Lee KM, Wong TW, Chan R, Lau CC, Fu YK, Fung KH. Accuracy and efficiency of X-ray requests initiated by triage nurses in an accident and emergency department. Accid Emerg Nurs 1996;4:179– 81. Ugeskr Laeger 2009;171:1747–51. 9. Wiler JL, Gentle C, Halfpenny JM, et al. Optimizing emergency department front-end operations. Ann Emerg Med 2010;55:142–60. 10. Rowe BH, Guo X, Villa-Roel C, et al. The role of triage liaison physicians on mitigating overcrowding in emergency departments: asystematic review. Acad Emerg Med 2011;18:111–20. 11. Porter J, Brennan D, Parrish G, Papa L, Nickolenko P, Bullard T. Resident triage-impact on patient satisfaction [abstract]. Acad Emerg Med 2009;16(4 Suppl 1):A272. 12. Rowe BH, Villa-Roel C, Guo X, et al. The role of triage nurse ordering on mitigating overcrowding in emergency departments: a systematic review. Acad Emerg Med 2011;18:1349–57. 13. 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: 702–8.
Resident-Initiated Advanced Triage
ARTICLE SUMMARY 1. Why is this topic important? Crowding in Emergency Departments (EDs) is a national problem that begins with long lengths of stay in triage. Various interventions have been proposed to improve patient length of stay, but no uniform intervention has been adapted as the universal solution. 2. What does this study attempt to show? This study evaluates whether placing a junior or a senior resident in triage is a feasible approach to decrease patients’ average length of stay. The objective is to show that an Emergency Medicine-trained resident in triage of an academic ED decreases length of stay and thereby decreases the number of patients leaving before medical screening. 3. What are the key findings? The intervention shows that resident advanced triage decreases the total length of stay for all patients evaluated in the ED during the intervention by 37 min after controlling for age, gender, and triage level. Further subgroup analysis shows that for all admitted patients, there was no statistically significant difference, whereas for all discharged patients, there was a trend toward a 34-min decrease in the length of stay. However, the proportion of patients who left before medical screening was not impacted. 4. How is patient care impacted? The goal of the study is to find a solution to decrease patient length of stay in an attempt to prevent patients from leaving the ED before medical screening. The hypothesis is that having a resident in triage helps to identify ‘‘sicker’’ patients faster, and by decreasing the amount of time an ED bed is occupied, more patients can be evaluated in the same amount of time. Furthermore, decreased time patients spend in the department, as well as shorter time to see a physician, likely has a positive impact on patient satisfaction, which is an objective we hope to study in the future.
751
http://dx.doi.org/10.1016/j.jemermed.2013.03.019
Administration of Emergency Medicine
RESIDENT-INITIATED ADVANCED TRIAGE EFFECT ON EMERGENCY DEPARTMENT PATIENT FLOW Irina Svirsky, MD,* Lisa R. Stoneking, MD,* Kristi Grall, MD,* Matthew Berkman, MD,* Uwe Stolz, PHD, MPH,* and Farshad Shirazi, MD, PHD*†‡ *EM Residency, PGY II, †University of Arizona College of Medicine, South Campus, Tucson, Arizona, and ‡Arizona Poison and Drug Information Center, Tucson, Arizona Reprint Address: Irina Svirsky, MD, University of Arizona College of Medicine, South Campus, 6790 E Calle La Paz apt 3203, Tucson, AZ 85715
, Abstract—Background: Emergency Department (ED) overcrowding is a national problem. Initiating orders in triage has been shown to decrease length of stay (LOS), however, nurse, physician assistant, and attending physician advanced triage have all been criticized. Study Objectives: Our primary objective was to show that Emergency Medicine resident-initiated advanced triage shortens patient LOS. Our secondary objective was to evaluate whether or not resident triage decreases the number of patients who left prior to medical screening (LPTMS). Methods: This prospective interventional study was performed in a 42-bed, Level III trauma center, academic ED in the United States, with an annual census of approximately 41,000 patients. A junior or senior Emergency Medicine resident initiated orders on 16 weekdays for 6 h daily on patients presenting to triage. Patients evaluated during the 6-h period on other weekdays served as the control. The study was powered to detect a reduction in LOS of 45 min. Multivariable median regression was used to compare length of stay and Fisher’s exact test to compare proportions. Results: There were 1346 patients evaluated in the ED during the intervention time. Regression analysis showed a 37-min decrease in median LOS for patients on intervention days as compared to control days (p = 0.02). The proportion of patients who LPTMS was not statistically different (p = 0.7) for intervention days (96/1346, 7.13%) compared to control days (136/1810, 7.51%). Conclusions: Residentinitiated advanced triage is an effective method to decrease patient LOS, however, our effect size is smaller than
predicted and did not significantly affect the percent of patients leaving before medical screening. Published by Elsevier Inc. , Keywords—advanced triage; physician triage; patient flow; length of stay
INTRODUCTION Many interventions have been done to improve Emergency Department (ED) patient wait times and reduce length of stay (LOS), but some patients still leave before being evaluated. One method to improve ED flow is initiating orders on patients in triage. Various interventions have been proposed, such as having a nurse, physician assistant, or an attending physician in triage (1–4). Most research to date has been done on nurse-initiated triage, including focused radiology studies, administering analgesics and antipyretic medications, and nurse initiated ‘‘order sets’’ based on chief complaint (2,5). Although initiating orders in triage has been shown to decrease LOS by as much as 37–76 min, advanced triage has been criticized for various reasons (2,6). Nurse triage has been mostly criticized for either under-ordering or over-ordering studies. Allerston and Justham’s nurseinitiated triage study found that the final diagnosis of fracture was missed in 4/187 patients with complaint of ankle
RECEIVED: 4 May 2012; FINAL SUBMISSION RECEIVED: 18 November 2012; ACCEPTED: 15 March 2013 746
Resident-Initiated Advanced Triage
pain when X-ray studies were not ordered based on the Ottawa Ankle Rule Protocol that nurses were supposed to have used in triage (7). Alternatively, Lee et al.’s study cited 5.4% of X-ray studies ordered by nurses in triage as unnecessary (8). Thus, this type of advanced triage has been linked to unnecessary radiation. On the other hand, Wiler et al.’s summary of the present advanced triage articles shows that having an attending physician is another approach that might reduce LOS and improve patient satisfaction, but is not always feasible due to the associated increased cost of having an extra attending physician on shift, legal risks, need for increased staff, practice variation, and risk tolerance (9). Furthermore, Rowe’s systemic review in 2011 evaluated 28 eligible studies on physician-advanced triage. Based on 19 studies, physician-initiated triage has been shown to decrease patient LOS by 37 min on average (10). However, all studies except Porter’s study had an attending physician in triage (11). The main disadvantage is that all other methods of triage (including nurse, attending, or physician assistant triage) either circumvented or diminished resident involvement and educational opportunity. A thorough literature search using the Cochrane, MEDLINE, ACP Journal Club, DARE, and PubMed databases from years 1990–2012 for keywords such as: ‘‘resident initiated,’’ ‘‘advanced triage,’’ ‘‘length of stay,’’ ‘‘emergency department flow,’’ and ‘‘patient flow’’ yielded many advanced triage research protocols, but only one publication on resident-initiated advanced triage. Porter’s abstract in Western SAEM (Society for Academic Medicine) in 2009 failed to show that placing a post-graduate year (PGY) 3 resident in triage improves patient satisfaction, LOS, or the number of patients who left without being seen (11). We therefore attempted to demonstrate that resident-initiated advanced triage can be a feasible and effective strategy to improve patient flow in an academic setting. Our objective was to demonstrate that Emergency Medicine resident physicianinitiated triage shortens patient LOS in the ED and decreases the number of patients who left prior to medical screening (LPTMS). MATERIALS AND METHODS This was a prospective intervention study at a community-based, academic, Level III trauma hospital. The intervention consisted of PGY2 or -3 Emergency Medicine resident (triage resident) initiating orders in triage on 16 random weekdays (Wednesday and Friday) for 6 h on patients presenting to triage during the busiest times. This was initially 12:00 pm–6:00 pm, but was changed halfway through the intervention to 2:00 pm– 8:00 pm to capture a higher number of patients. The
747
choice of weekdays was based on lack of support staff on other weekdays as well as inability to do the intervention during ‘‘teaching days’’ when only attending physicians worked in the ED. The time of intervention changed halfway through the intervention due to the observation that, at these times, empty ED beds and low patient volumes were leading to triage patients bypassing the triage area and being sent directly to an ED bed. The residents were financially compensated, and were paid the same wage no matter how many patients they evaluated. They were blinded to the outcomes being measured. Patients with psychiatric complaints and alcohol intoxication were excluded because in our electronic medical record database there is no documented time for ‘‘medical clearance,’’ and these patients have the longest LOS due to long wait times for social work or psychiatric evaluation. The control group consisted of patients evaluated during the same week and 6-h period as the intervention group but on different weekdays (Monday and Thursday) (Table 1). Per historical data, our ED evaluated the same number of patients on the 2 days we chose as control as on the days we chose to perform the intervention. The population consisted of all patients presenting to the ED during the 6-h period, including: male and female patients, ages 0–99 years, with any medical complaints, presenting by any means of transportation. The excluded population consisted of patients presenting with a psychiatric complaint or ‘‘alcohol intoxication’’ as their primary complaint due to inability to evaluate the time to medical clearance. The intervention was a junior or senior (PGY2 or PGY3) resident (triage resident) who remained in triage with the triage nurse and began their work-up after the triage nurse assessment. Verbal permission to initiate evaluation and treatment was obtained by the triage resident. Upon consent, the patients were then relocated to a separate triage area with recliner chairs and cardiopulmonary monitors where assessment was initiated. After nursing staff completed triage, each patient was temporarily placed in a separate room for the triage resident to collect further history and perform a physical examination, as well as for the nurse to draw laboratory tests and perform electrocardiograms, in compliance with the Health Insurance Portability and Accountability Act. The residents had the ability to order any test or treatment they deemed necessary for patient care, including laboratory tests, radiographic studies, or medications. Separate nursing staff was assigned to this area to collect blood and urine samples and establish intravenous access. Treatment was also initiated in this area. Radiology testing was available, with patients transported from triage to Radiology. Intervention patients were assigned to a room in the original order they were assigned by the triage nurse. Once the patient was assigned a room, the triage resident
748
I. Svirsky et al.
Table 1. Demographics Intervention (n = 1346) Gender Male Female Mean age Years (SD) Triage level 1 2 3 4 5 Disposition Admitted Discharged Transferred LPTMS or eloped (14)
median regression was used to compare LOS and Fisher’s exact test to compare proportions. RESULTS
50.50% 49.50% 44.9 (19.7) 1% 12% 80% 7% 0.50% 30.4% 58.3% 4.3% 7%
LPTMS = left prior to medical screening by a physician; eloped = left the room without Against Medical Advice paperwork, after a medical screening.
communicated with the attending physician and the main ED resident to inform them about the patient course in triage verbally as well as by a quick written note in the ‘‘SOAP’’ format (Subjective, Objective, Assessment, Plan). A triage resident also had the ability to discharge a patient from the triage area if the evaluation and treatment was complete and if the attending physician could evaluate the patient. If the patient was being discharged from triage, the resident in triage wrote an entire history and physical examination note for that patient, evaluated the patient in a separate area of triage to perform a complete physical examination, and the patient was seen by the attending physician before discharge. The ED attending was always available to the triage resident and there were no safety issues or concerns regarding patient care during the intervention period. At the end of the intervention time period, if any patients remained in the triage area, the triage resident transferred patient care to the main ED and verbally communicated with a resident and attending in the main ED as well as via a written SOAP note. The triage area then closed and patients were transferred to the main ED to any open beds. If there were no open beds, staff was assigned to the triage area until patients could be transferred to the main ED. In our study, at the end of the intervention time period, most patients could be transferred to the main ED or were discharged. The study had two outcomes: patient LOS and the percent of patients who LPTMS. We considered the LPTMS patients as those patients who registered but were not evaluated by an attending physician. The study was powered to detect a reduction in LOS of 45 min for all patients being evaluated in the ED. Multivariable
A total of 1346 patients were evaluated during the 6-h intervention period on 16 days from November 2011 to February 2012. Table 1 shows that the demographic data on intervention patients was comparable to control patients in the categories of gender, age, triage level, and the percentage of patients admitted, discharged, or transferred. Further demographic data could not be collected due to limitations in the computer charting software. However, Table 1 shows that gender, age, and Emergency Severity Index levels were similar in the two groups. As shown in Table 2, regression analysis showed a 37-min decrease in median LOS for all patients evaluated in the ED for intervention days compared to control days (95% confidence interval [CI] 4–69) after controlling for age, gender, and triage level. This decrease in LOS was comparable to the mean of other studies using advanced triage with a reduction in LOS of 18–82 min, consistent with a systematic review that evaluated nurse and physician advanced triage from 1966–2008 (9). Further subgroup analysis shows that for all admitted patients, there was no statistically significant difference in ED LOS (p = 0.66), whereas for all discharged patients, there was a trend toward a 34-min decrease in ED LOS (p = 0.07, 95% CI 3–71). The proportion of patients who LPTMS was not statistically different (p = 0.7) for intervention days (96/1346, 7.13%) compared to control days (136/1810, 7.51%) (Table 3). DISCUSSION Our study measured two outcomes: patient LOS and the number of patients who LPTMS. Patient LOS was 37 min less during the intervention 6-h period as compared to control days (p = 0.02, 95% CI 4–69). Our results are consistent with our hypothesis, that having
Table 2. Patient Length of Stay Length of Stay
Coefficient
All patients
Crude coefficient Adjusted (age, sex, triage level) Admitted patients Crude coefficient (age, sex, triage level) Discharged patients Crude coefficient (age, sex, triage level) CI = confidence interval; LOS = length of stay.
LOS (min)
95% CI
25 37
56 to 7 59 to 4
12 11 33 34
65 to 41 61 to 40 68 to 2 71 to 3
Resident-Initiated Advanced Triage
749
Table 3. Patients Left Prior to Medical Screening (LPTMS)
LPTMS Did not LPTMS Total % LPTMS (p = 0.730)
Intervention
Control
96 1250 1346 7.13%
136 1674 1810 7.51%
a resident-initiated advanced triage decreased the overall LOS for all patients seen in the ED. One of the goals of this study was to compare the triage resident efficiency to other practitioners. A literature review search published in Academic Emergency Medicine in 2011 reviewed 14,000 ‘‘nurse in triage’’ interventions, which showed that the mean reduction in LOS was 37–51 min in four of the most efficient studies. However, other studies failed to show a statistical difference when interventions were applied to subjects without suspected injury or fracture. When interventions were applied to patients suspected of having a fracture that required radiology, LOS was reduced by 19 min (12). We found no studies that measured LOS for other midlevel practitioners such as physician assistants. We aimed to compare the efficiency of PGY2 and PGY3 residents in our study; however, this was not possible as only three of the 16 shifts were covered by PGY3 level residents. In Canada, in an academic tertiary institution similar to ours, an attending physician was placed in triage for 6 weeks for 9 h daily. A median of 14 patients were seen per shift. Overall, LOS was decreased by 36 min compared with control days (4:21 vs. 4:57; p = 0.001). Patients who left without complete assessment decreased from 6.6% to 5.4% during the intervention (13). Although LOS in the above study was comparable to our intervention, there is clearly an advantage to having a senior physician in triage, including experience and safety regarding patient care. However, in our study we designed a unique intervention to involve residents in learning as well as increase their independence, while still having an attending physician present for back-up. In addition, the residents that were involved in the study had a unique perspective in learning ED flow. Safety was not an issue as all patients were seen by the attending physician either in triage or when they were moved to the main ED bed. In addition, one of the reasons the intervention was done was to identify the sicker patients first by starting their work-up, as opposed to control days, when the patients would have remained in the waiting room. If the triage nurse or resident thought the patient was critically ill, they were moved immediately to an ED bed. Further subgroup analysis demonstrated that there was no statistical difference in ED LOS for admitted patients. However, for the discharged patients, there was a trend
toward a decrease in ED LOS of 34 min, although it was not statistically significant. Failure to show a difference for the subgroups was likely due to the fact that the study was underpowered. Unfortunately, our study could not be carried out to show a statistical difference due to lack of funding. However, in the future, we hope to carry out larger studies to show a statistical difference between the two groups. For patients who LPTMS, there was no statistically significant difference between the two groups: 7.13% for the intervention group compared to 7.51% for the control group (p = 0.7). Using resident-initiated advanced triage to decrease ED LOS, we attempted to more rapidly evaluate a greater number of patients to improve ED flow. Although the ED LOS was decreased, it did not affect the number of patients who LPTMS. Therefore, the number of patients being treated and evaluated early was not large enough to make an impact on the percent of patients leaving before being evaluated by a physician. We realize that at our institution, the number of patients who LPTMS is quite high at 7%. This is a multifactorial problem that may not be fixed by decreasing patient-to-physician time. We attempted, however, to address this complex issue with this intervention. Although in our study there was no statistical difference in the number of patients leaving in the intervention group (7.13%) and the control group (7.51%), in the future we hope to perform a larger study as well as address other issues contributing to this problem. Limitations Our study showed a mean decrease of 37 min for all patients, however, it was underpowered to show a decrease in LOS for the subgroups of admitted and discharged patient. For the discharged patients, there was a trend toward 34-min decrease; however, it was not statistically significant. Our study was carried out for only 16 intervention shifts due to lack of funding; however, we are hoping to increase the power in future studies. For admitted patients, there was a small difference in the LOS that was not statistically significant. In our academic institution, the LOS is measured at the time the resident places a ‘‘bed request’’ in the computer system. This is variable, and a resident could have decided to admit a patient before a bed request was actually placed. Furthermore, admitted patients included patients admitted to our facility as well as transferred to an outside facility. Therefore, a significant amount of time is lost while waiting to get an accepting physician on the line, do a ‘‘physician-to-physician communication,’’ and facilitate transport. The 16 intervention days included in this study were non-consecutive from November 2011 to February
750
I. Svirsky et al.
2012. The study included junior and senior level residents that have different expertise levels, and those subgroups were not further analyzed due to the few shifts that the senior level residents worked. The biggest limitation was that the reduced LOS did not affect the number of patients leaving before medical screening. We hypothesize that this was due to the fact that our triage area can only encompass 5–7 patients at a time due to the limited equipment and nursing staff. We found that the highest number of patients who left without being seen was after the intervention time ended and all of the ED beds were occupied. Additional limitations included not being able to measure other outcomes such as patient satisfaction or the number of patients who left at any time during the visit. This was due to the newness of the computer software in our ED and the inability to measure this type of data. The ED had also expanded in the number of beds and was moved to a new and larger geographical area, twice the size of the old ED 3 months before the intervention. CONCLUSION Preliminary results indicate that resident-initiated advanced triage is an effective and feasible method to decrease ED LOS; however, our effect size is smaller than predicted. Our results failed to show a statistical significance among the numbers of patients leaving before medical screening. Likely, this was due to a limitation in space, equipment, nursing staff, and the study being underpowered. More research needs to be done to determine whether resident-initiated advanced triage can improve the flow of the emergency department. Acknowledgments—Jackie DeBeche, RN, MBA/HCA, Nursing Director, UAMC-South Campus, University of Arizona Emergency Medicine Residency Program, and the Residents
and Staff of University of Arizona, Emergency Residency Program.
REFERENCES 1. Rosmulder RW, Krabbendam JJ, Kerkhoff AH, Schinkel ER, Beenen LF, Luitse JS. ‘Advanced triage’ improves patient flow in the emergency department without affecting the quality of care [Dutch]. Ned Tijdschr Geneeskd 2010;154:A1109. 2. Cheung WWH, Heeney L, Pound JL. An advanced triage system. Accid Emerg Nurs 2002;10:10–6. 3. Love RA, Murphy JA, Lietz TE, Jordan KS. The effectiveness of a provider in triage in the emergency department: a quality improvement initiative to improve patient flow. Adv Emerg Nurs J 2012;34: 65–74. 4. Travers JP, Lee FC. Avoiding prolonged waiting time during busy periods in the emergency department: is there a role for the senior emergency physician in triage? Eur J Emerg Med 2006;13: 342–8. 5. Tambimuttu J, Hawley R, Marshall A. Nurse-initiated x-ray of isolated limb fractures in the emergency department: research outcomes and future directions. Aust Crit Care 2002;15:119–22. 6. Lindley-Jones M, Finlayson BJ. Triage nurse requested x rays—are they worthwhile? J Accid Emerg Med 2000;17:103–7. 7. Allerston J, Justham D. Nurse practitioners and Ottawa Ankle Rules: comparisons with medical staff in requesting X rays for ankle injured patients. Accid Emerg Nurs 2000;8:110–5. 8. Lee KM, Wong TW, Chan R, Lau CC, Fu YK, Fung KH. Accuracy and efficiency of X-ray requests initiated by triage nurses in an accident and emergency department. Accid Emerg Nurs 1996;4:179– 81. Ugeskr Laeger 2009;171:1747–51. 9. Wiler JL, Gentle C, Halfpenny JM, et al. Optimizing emergency department front-end operations. Ann Emerg Med 2010;55:142–60. 10. Rowe BH, Guo X, Villa-Roel C, et al. The role of triage liaison physicians on mitigating overcrowding in emergency departments: asystematic review. Acad Emerg Med 2011;18:111–20. 11. Porter J, Brennan D, Parrish G, Papa L, Nickolenko P, Bullard T. Resident triage-impact on patient satisfaction [abstract]. Acad Emerg Med 2009;16(4 Suppl 1):A272. 12. Rowe BH, Villa-Roel C, Guo X, et al. The role of triage nurse ordering on mitigating overcrowding in emergency departments: a systematic review. Acad Emerg Med 2011;18:1349–57. 13. 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: 702–8.
Resident-Initiated Advanced Triage
ARTICLE SUMMARY 1. Why is this topic important? Crowding in Emergency Departments (EDs) is a national problem that begins with long lengths of stay in triage. Various interventions have been proposed to improve patient length of stay, but no uniform intervention has been adapted as the universal solution. 2. What does this study attempt to show? This study evaluates whether placing a junior or a senior resident in triage is a feasible approach to decrease patients’ average length of stay. The objective is to show that an Emergency Medicine-trained resident in triage of an academic ED decreases length of stay and thereby decreases the number of patients leaving before medical screening. 3. What are the key findings? The intervention shows that resident advanced triage decreases the total length of stay for all patients evaluated in the ED during the intervention by 37 min after controlling for age, gender, and triage level. Further subgroup analysis shows that for all admitted patients, there was no statistically significant difference, whereas for all discharged patients, there was a trend toward a 34-min decrease in the length of stay. However, the proportion of patients who left before medical screening was not impacted. 4. How is patient care impacted? The goal of the study is to find a solution to decrease patient length of stay in an attempt to prevent patients from leaving the ED before medical screening. The hypothesis is that having a resident in triage helps to identify ‘‘sicker’’ patients faster, and by decreasing the amount of time an ED bed is occupied, more patients can be evaluated in the same amount of time. Furthermore, decreased time patients spend in the department, as well as shorter time to see a physician, likely has a positive impact on patient satisfaction, which is an objective we hope to study in the future.
751