The Journal of Emergency Medicine, Vol. -, No. -, pp. 1–10, 2011 Ó 1 Copyright 2011 Elsevier Inc. 69 Printed in the USA. All rights reserved 2 0736-4679/$ - see front matter 70 3 71 4 doi:10.1016/j.jemermed.2011.03.007 72 5 73 6 74 7 75 8 76 9 Administration of 77 10 78 11 Emergency Medicine 79 12 80 13 81 14 82 15 83 16 EXPLORING STRATEGIES TO IMPROVE EMERGENCY DEPARTMENT INTAKE 84 17 85 18 86 19 Shari Welch, MD, FACEP, and Lucy Savitz, PHD, MBA 87 20 88 21 Intermountain Institute for Health Care Delivery Research, Salt Lake City, Utah 89 Reprint Address: 22 Shari Welch MD, FACEP, Intermountain Institute for Health Care Delivery Research, 36 South State Street, 90 23 16th Floor, Salt Lake City, UT 84111 91 24 92 25 , 93 26 Abstract—Background: The emergency department in the United States, according to the Agency for Health- 94 27 (ED) is the point of entry for nearly two-thirds of patients ad- care Research and Quality (AHRQ) (1). Coming through 95 28 mitted to the average US hospital. Due to unacceptable waits, this ‘‘front door’’ to the hospital are over 200 visits every 96 29 3% of patients will leave the ED without being seen by a phy- minute of every day in the United States, or 40 visits to 97 30 sician. Objectives: To study intake processes and identify new the ED per 100 citizens every year. Every 2 1/2 years 98 31 strategies for improving patient intake. Methods: Ayear-long 99 the equivalent of the entire US population passes through 32 learning collaborative was created to study innovations in- 100 33 volving the intake of ED patients. The collaborative focused ED doors. Almost 40% will wait more than an hour to see 101 34 on the collection of successful innovations for ED intake for a physician, and these unacceptable waits will result in an 102 35 103 an ‘‘improvement competition.’’ Using a qualitative scoring average of 3% of patients leaving without being seen by 36 a physician (2,3). Thus, ED inefficiencies and delays 104 37 system, finalists were selected and their innovations were pre- 105 38 sented to the members of the collaborative at an Association translate into almost four million patients walking away 106 39 for Health Research Quality-funded conference. Results: from health care each year. It is no surprise that the 107 40 Thirty-five departments/organizations submitted abstracts Joint Commission has found that the ED is the most 108 41 for consideration involving intake innovations, and 15 were common site for sentinel events in the hospital due to 109 42 selected for presentation at the conference. The innovations waits and delays in care (4). 110 43 111 were presented to ED leaders, researchers, and policymakers. Timeliness of care is an issue that is front and center 44 Innovations were organized into three groups: physical plant 112 for EDs in the United States. Timeliness of care is among 45 changes, technological innovations, and process/flow changes. 113 46 114 Conclusion: The results of the work of a learning collabora- the strongest correlates with patient satisfaction (5).The 47 115 tive focused on ED intake are summarized here as a qualita- time it takes to see a physician (door-to-physician time) 48 has the best correlation of all. By moving patients quickly 116 49 tive review of new intake strategies. Early iterations of these 117 50 new and unpublished innovations, occurring mostly in non- to patient care areas for evaluation, patients perceive that 118 51 academic settings, are presented. Ó 2011 Elsevier Inc. the wait time is acceptable (6–10). As the time from 119 52 arrival to physician evaluation increases, the rate of 120 53 , Keywords—emergency department; intake; triage; pro- patients leaving without being seen increases (11–13). 121 54 cess improvement; door to physician time Finally, a growing number of clinical entities require 122 55 123 56 treatment that is ‘‘on the clock,’’ with outcomes directly 124 57 INTRODUCTION linked to timeliness of care (14–21). 125 58 There is innovation occurring in the area of intake 126 59 The modern-day emergency department (ED) is the point strategies, though much of it is unpublished. Unlike clin- 127 60 ical care processes, improvement in ED operations often 128 61 of entry for two-thirds of patients admitted to the hospital 129 62 130 63 131 64 132 65 133 RECEIVED: 16 April 2010; FINAL SUBMISSION RECEIVED: 18 July 2010; 66 134 ACCEPTED: 16 March 2011 67 135 68 136 1

FLA 5.1.0 DTD - JEM8728_proof - 11 April 2011 - 6:08 pm - ce 2 S. Welch and L. Savitz 137 205 138 206 139 occurs in the trenches many years before it reaches the lit- acknowledged leader in quality improvement and patient 207 140 erature (22). safety founded by Brent James, MD, MSTAT (35–38).In 208 141 The collaborative model for health care improvement December 2008, EDBA, in collaboration with the 209 142 210 143 has been used by the Institute for Healthcare Improve- Intermountain Institute for Health Care Delivery 211 144 ment (IHI) since 1995. A learning collaborative (also Research, submitted a small conference grant proposal to 212 145 called a learning system) is a short-term (6- to 15-month) AHRQ titled ‘‘Summit Exploring Emergency Department 213 146 organization that brings together a large number of teams Intake Strategies.’’The goals of this learning collaborative 214 147 from hospitals or clinics to seek improvement in a focused and conference are summarized below: 215 148 216 topic area. Since 1995, IHI has sponsored over 50 such 149 1) To provide a venue for sharing the newest ideas on 217 collaborative projects on several dozen topics involving 150 intake 218 151 over 2000 teams from 1000 health care organizations 219 2) To inform health care policymakers of top per- 152 (23). Typically, the work of the health care learning col- 220 forming strategies in this area 153 laborative focuses on operational improvement alone or 221 154 3) To move emergency medicine forward into the 222 incorporates medical care strategies that have already 155 realm of operations management 223 been proven in traditional medical research trials into 156 4) To provide a shared learning experience 224 157 new process models. The ‘‘IHI ED Collaborative: Opera- 225 158 tional and Clinical Improvements for the Emergency A steering committee was formed composed of the 226 159 Department’’ launched a 2-year program in 2006 that Board of Directors of the EDBA and research scientists 227 160 228 161 was a good example of this model (23). Thirty-four geo- from the Intermountain Institute for Health Care Delivery 229 162 graphically diverse EDs marched through time working Research. Our learning collaborative was formed by invi- 230 163 on improvements and sharing ideas. tations from the steering committee. These invitations 231 164 This learning collaborative and conference created were extended to individuals targeted for expertise, inter- 232 165 a forum for sharing new ideas and early trials, but does est, and influence, and for their associations with critical 233 166 234 167 not presume to be presenting the definitive word on these stakeholder organizations. The American College of 235 168 emerging new strategies. Its aim was the diffusion of in- Emergency Physicians maintains a roster of health care 236 169 novation and the dissemination of ideas. Our learning col- leaders and policymakers for their work on quality 237 170 laborative comprised ED leaders, policymakers, and and performance committees. Many of the invited indi- 238 171 front-line workers. It was tasked to identify constraints viduals were contacted through this roster. Our learning 239 172 240 173 to intake and innovative solutions to these constraints. collaborative was organized around three workgroups: 241 174 Aware that much innovative work in this area never rea- 1) Constraints to Intake; 2) Innovations at Intake; and 242 175 ches a forum allowing widespread dissemination, the 3) Definitions, Terminology, and Measures. The work- 243 176 Emergency Department Benchmarking Alliance groups met through conference calls and an eventual 244 177 (EDBA) Board of Directors began an outreach program face-to-face meeting at the collaborative summit meeting 245 178 246 179 looking for new but as yet unpublished innovations rela- and crafted a comprehensive whitepaper on ED intake, 247 180 tive to the intake process. The results of this collaborative which has been presented to AHRQ. One unique feature 248 181 were shared at an AHRQ-sponsored summit in February of this learning collaborative is that it brought together 249 182 of 2010. This is a summary of the findings presented at front-line innovators with health care leaders and policy- 250 183 that meeting. makers to share the work being done by both segments of 251 184 252 185 the collaborative. 253 186 METHODS During early planning of our collaborative, the steer- 254 187 ing committee crafted an informational flyer promoting 255 188 EDBA is a not-for-profit organization comprising 367 an Innovator Competition, with the intent of showcasing 256 189 EDs representing over 14 million ED visits annually. ideas and innovations that resulted in improvements of 257 190 258 191 EDBA was founded in 1997 as an alliance of the intake process into the ED. The steering committee 259 192 performance-driven EDs. It operates as a not-for-profit also developed an electronic abstract submission process 260 193 quality improvement and learning community, sharing for the competition. The only requirement was that mea- 261 194 performance data and operational strategies to identify surable results showing improvement be included in the 262 195 best practices. EDBA has developed a benchmarking da- abstract. Figure 1 is a sample of the simple abstract sub- 263 196 264 197 tabase and educational programs focusing on ED opera- mission form. Upon notification that the grant had been 265 198 tions and performance and disseminates new ideas awarded, the steering committee began an intensive net- 266 199 and innovations through conferences and publications working and outreach effort to solicit abstracts describing 267 200 (24–34). For execution of the AHRQ grant, EDBA successful improvement projects involving intake pro- 268 201 partnered with Intermountain Institute for Health Care cesses. Abstracts were solicited from the following 269 202 270 203 Delivery Research. The Institute is an internationally organizations: 271 204 272

FLA 5.1.0 DTD - JEM8728_proof - 11 April 2011 - 6:08 pm - ce Strategies to Improve ED Intake 3 Q1 273 341 274 342 275 Symposium on New Intake Models for the Emergency Department Abstracts were accepted from April 1, 2009 through 343 276 344 Organization: ______September 30, 2009. Abstracts were scored by members 277 Author(s):______of the steering committed using a four-point scale (1, 2, 3, 345 278 Presenting Author: ______346 279 Email/Phone:______or 4, for fair, good, very good, and excellent, respectively) 347 280 for each of five parameters: 1) creativity, 2) clarity, 3) ap- 348 281 Title: plicability to varying ED types, 4) inclusion of measure- 349 282 able results, and 5) innovative impact. Figure 2 is an 350 283 example of the scoring form used. Although the original 351 284 Summary of New Model: 352 285 plan had been to select the top 10 abstracts, due to scoring 353 286 ties and the inclusion of more than one creative idea in 354 287 the improvement process, 15 abstracts were selected to 355 288 be presented at the ‘‘Summit Exploring New Intake 356 289 Models.’’ Because some innovations were trialed at 357 290 358 291 more than one site, the steering committee identified 13 359 292 Results: meritorious innovations for summary herein. 360 293 361 294 RESULTS 362 295 363 296 364 Submitted abstracts were organized into three thematic 297 Lessons Learned: 365 298 categories: Physical Plant Changes, Technological 366 299 Changes, and Process/Flow Changes. 367 300 368 Figure 1. Abstract submission form. 301 Physical Plant Changes 369 302  370 303 ACEP (The American College of Emergency Physi- 371 304 cians) Physician Cubicles 372 305 - The Quality and Performance Committee Triage Pod 373 306 - The Practice Committee Recliner Intake Area 374 307 - The Quality Improvement and Patient Safety In- Internal Waiting Area 375 308 376 terest Group 309  377 310 SAEM (The Society for Academic Emergency Physician Cubicles. At Arrowhead Regional Medical 378 311 Medicine) Center in California, in response to a census that doubled 379 312 - Patient Safety Interest Group in 5 years to 120,000 visits and left without being seen 380 313  IHI (The Institute for Healthcare Improvement) (LWBS) rates that had reached 20%, the staff trialed 381 314 382 - 315 Operational and Clinical Improvement in the ED a physician-in-triage model made possible by bringing 383 316 Learning Community in furniture modules that created small cubicles in which 384 317  EDPMA (Emergency Department Practice Man- physicians could see patients. Because this change was 385 318 agement Group) implemented by bringing in modular furniture and with- 386 319  EDBA membership out alteration of the physical building structure, the inno- 387 320  388 321 Intermountain Healthcare Intensive Clinical Medi- vation was inexpensive and required no building permits. 389 322 cine Programs The new model involved patients being seen by a provider 390 323  EMP (Emergency Medicine Physicians) first. Using this model, ED staff found that half of patients 391 324  Premier Health Care could be discharged from the cubicles, 30% required 392 325  The Schumacher Group some laboratory or X-ray diagnostics, and only one in 393 326  394 327 The California ED Diversion Project five patients needed bed placement when the provider 395 made the first contact. As a result, more beds became 328 These organizations were asked to consider but not be 396 329 available in the ED and there was an unexpected reduc- 397 limited to the following elements of ED patient intake as 330 tion in nurse staffing. The LWBS dropped from 20% to 398 areas of innovation for the competition: 331 1%, and the time to see a physician was reduced from 399 332 400 333 1) Patient Identification 4 h to 31 min. 401 334 2) Initial Clinical Assessment 402 335 3) IT Support of Intake Workflow Triage Pod.Methodist Hospital in Sacramento was under- 403 336 4) Staffing Models bedded. Over 40,000 visits annually were being managed 404 337 5) Documentation in a 19-bed ED. The staff created a six-bed ‘‘triage pod’’ 405 338 406 339 6) Advanced Triage Protocols area using simple room dividers, for team assessment of 407 340 408

FLA 5.1.0 DTD - JEM8728_proof - 11 April 2011 - 6:08 pm - ce 4 S. Welch and L. Savitz 409 477 410 478 411 479 412 480 413 481 414 482 415 483 416 484 417 485 418 486 419 487 420 488 421 489 422 490 423 491 424 492 425 493 426 494 427 495 428 496 429 497 430 498 431 499 432 500 433 501

434 FPO 502 435 = 503 436 504

437 web 4C 505 438 Figure 2. --- Q9 506 439 507 440 508 441 patients and rapid intake. The change in the physical that began with the categorizing patients (also called pa- 509 442 space was married to a process change. The team, com- tient segmentation) by acuity and resources required. The 510 443 posed of one physician, two physician assistants, and so-called START program (Supplemental Triage and 511 444 512 445 four nurses, worked the triage pod with the goal of mov- Rapid Assessment) involved process changes coupled 513 446 ing each patient to an appropriate area in < 15 min. with physical space changes. An important change in 514 447 Patients were then transferred to one of three areas: the the physical plant to support this process involved the 515 448 waiting room, the main ED, or a monitored higher- creation of an internal waiting room called the ‘‘post- 516 449 acuity ED bed. Many patients were sent right to the wait- screening area.’’ The internal waiting room allows less 517 450 518 451 ing room to await discharge or further diagnostics done as acute patients to remain vertical, instead of occupying 519 452 ambulatory patients. The already under-bedded depart- bed space, while awaiting test results. The sum of these 520 453 ment was reduced from 19 beds to 13 beds, but with changes to the physical plant and patient flow resulted 521 454 new processes in place they have smarter bed utilization. in an 8% decrease in length of stay (LOS) and a drop in 522 455 Only the sicker patients occupy ED beds after passing LWBS rate from 4.1% to 2.4%. 523 456 524 457 through the triage pod. Methodist has seen the LWBS 525 458 rates drop from 5% to 1%. Technological Changes 526 459 527 460 Recliner Intake Area. Carolinas Medical Center rede- Self-populating Triage Tool 528 461 signed their intake area putting recliners and supplies Palmar Scanning 529 462 530 463 within reach of the physician and team. Like Arrowhead, Telemedicine Triage 531 464 they found that 45.5% of patients could be discharged by Radio/Communication Devices 532 465 the physician from this intake area. This is an effective 533 466 way to off-load the main department when at overcapac- Self-populating Triage Tool. The ED at the University of 534 467 ity. These data are in line with the findings of other California San Diego (UCSD) has explored ways that 535 468 536 469 departments that put a physician out front. Carolinas technology can facilitate intake. The staff developed an 537 470 has seen improvement in the intake time from 58 min informatics tool that immediately populates the fields of 538 471 to 35 min. the electronic health record in the triage note. For in- 539 472 stance, medications, allergies, and past medical history 540 473 Internal Waiting Room.At Massachusetts General Hospi- are automatically pulled forward to the current health re- 541 474 542 475 tal, a complex new ED flow process was implemented cord if a patient has ever been in the UCSD system before. 543 476 544

FLA 5.1.0 DTD - JEM8728_proof - 11 April 2011 - 6:08 pm - ce Strategies to Improve ED Intake 5 545 613 546 614 547 This self-populating tool shortened intake time by 20 min facilitated by inexpensive radio communication, the 615 548 and led to improved provider satisfaction. LWBS rate fell from 12% to 1.5%. Staff satisfaction by 616 549 survey is also at an all-time high. 617 550 618 551 Palmar Scanning. At Carolinas Medical Center, the ED 619 552 has streamlined the intake process by using palmar scan- Process/Flow Changes 620 553 ning to create a biomedical identification. Like retinal 621 554 scanning, a palm print is used to generate an immediate Scribe Program 622 555 identifier for a patient and tied to an identification (ID) Low Flow/High Flow Process 623 556 624 557 number. Later the ID can be associated with demographic Physician in Triage 625 558 data. In < 15 s, the patient is identified through this meth- Patient Streaming/Segmentation 626 559 odology, allowing treatment to begin. The device can en- The Philadelphia EMS Admission Rule (PEAR) 627 560 sure that a patient is associated with the right medical 628 561 record number. In addition to preventing identity fraud Scribe Program. At Cortland Regional Medical Center, 629 562 630 563 and mismatched records, the device can quickly identify the emergency physicians were concerned about increas- 631 564 unconscious or ‘‘altered mental status’’ patients who have ing clerical and documentation tasks. Despite increases in 632 565 previously been scanned. Since the implementation of patient acuity, reimbursements and physician satisfaction 633 566 this high-tech patient ID method, the door-to-physician had fallen. The 32,000-visit department began a scribe 634 567 time is now being measured in seconds instead of minutes program and has seen improvement in documentation, 635 568 636 569 and is now 45 s at Carolinas Medical Center. reimbursement, productivity, and patient satisfaction. 637 570 Throughputs have also improved as the scribes have be- 638 571 Telemedicine Triage.The Medical College of Georgia ED gun facilitating data collection, freeing up the physician 639 572 is situated in close proximity to a handful of nursing for other tasks. 640 573 homes and extended-care facilities. On a daily basis, 641 574 642 575 the staff found that they were inundated by low-acuity Low Flow/High Flow Process. At Thomas Jefferson Uni- 643 576 patient transfers that were costly and resulted in high versity, the busy urban teaching hospital dubbed their new 644 577 utilization of Emergency Medical Services (EMS) re- innovation the ‘‘Low Flow/High Flow’’ process model. In 645 578 sources. The Medical College of Georgia has begun trials this model, the intake process varies with the volume of 646 579 using telemedicine to avoid such transfers to the ED. The arrivals to the ED. When the ED is at low census with 647 580 648 581 telemedicine technology allows the physician who is off open beds, the process is a ‘‘pull to full’’ approach, with 649 582 site to see and hear the patients, family members, and immediate bedding of patients and intake processes 650 583 staff. There is a stethoscope that allows the physician to occurring at the bedside. As the ED reaches capacity, it 651 584 hear breath sounds and heart sounds remotely; there is shifts into the ‘‘High Flow’’ process. In this model, a pro- 652 585 also a remote otoscope and ophthalmoscope. A physical cessing area is opened and a team using protocol-guided 653 586 654 587 examination can be carried out remotely with the technol- treatment plans begins the work-up in the processing 655 588 ogy. Using this technology, the ED and nursing home area. The first pilot of the new High Flow model showed 656 589 staff can often address low-acuity medical problems a decrease in LOS from 653 min to 158 min. LWBS rates 657 590 without transport to the hospital. Also, the physician fell from 11% to 6%. Exit surveys of patients involved in 658 591 can often identify acutely ill patients and recommend the pilot showed extremely high patient satisfaction 659 592 660 593 transport with ‘‘lights and sirens’’ while the ED prepares scores: 4.5 on a scale of 5 for satisfaction. 661 594 to treat such a patient in an expeditious manner. Each pa- 662 595 tient not transferred to the ED saves between $404.00 and Physician in Triage (PIT).The most frequently trialed in- 663 596 $662.00 (Basic Life Support non-emergent vs. Advanced novation in our learning collaborative was the placement 664 597 Life Support emergent) in EMS one-way transport of a physician at the front end of the ED visit. Many 665 598 666 599 charges in that community. variations on this theme were trialed by the innovators. 667 600 Memorial Hospital in York, Pennsylvania used a variation 668 601 Radio/Communication Devices. The use of a radio com- of the Physician in Triage (dubbed the ‘‘PIT Process’’). 669 602 munication device has been shown to improve processes York begins the intake process with a podium nurse doing 670 603 at intake. At St. Rose Dominican Siena Campus outside a quick look before the PIT team assesses the patient. For 671 604 672 605 of Las Vegas, the 42,000-visit ED used radios to call 10 h a day the PIT physician makes an initial rapid 673 606 a physician to triage to assess each patient and begin medical assessment of each patient, a process that takes 674 607 the work-up. All Emergency Severity Index (ESI) 1 and < 3 min. Standardized order sets are begun in triage for 675 608 2 patients are immediately placed in a bed. All ESI 3, ESI 2 and 3 patients. ESI 1 patients are immediately 676 609 4, and 5 patients have a physician assessment to direct bedded. ESI 4 and 5 patients are assigned to a physician 677 610 678 611 the work-up and care. With this small process change, assistant in the fast track. York documented a decrease in 679 612 680

FLA 5.1.0 DTD - JEM8728_proof - 11 April 2011 - 6:08 pm - ce 6 S. Welch and L. Savitz 681 749 682 750 683 the LWBS patients from 6% to 0.4% in the first trial of the Spaite et al. found that changes made to the physical 751 684 project. Door-to-physician times were reduced from 65 plant, combined with process changes, resulted in de- 752 685 min to 32 min. The physicians also reported they felt creased wait times and improved patient satisfaction, 753 686 754 687 that bed utilization improved with the new model. though these changes were accomplished at considerable 755 688 expense (39). At the other extreme, Morgan created sub- 756 689 Patient Streaming/Patient Segmentation. At Banner waiting spaces to facilitate changes in flow and did this 757 690 Health System in Arizona, a similar ‘‘quick look’’ of pa- simply by moving banks of chairs into hallways (40). 758 691 tients followed by patient segmentation was employed in The innovators in our learning collaborative used cre- 759 692 760 693 a new intake model. The Banner staff called this process ative means to change the physical space to improve pro- 761 694 D2D SPF (Door to Doc Split Patient Flow). In this model, cesses. Front-line practitioners with a solid understanding 762 695 less sick patients are not undressed or bedded, but rather of operations and processes should and can be able to help 763 696 treated as though they were in a clinic setting. The sickest their departments adapt the physical space to accommo- 764 697 patients are seen in an expedient manner and treatment date workflow. In the improvement projects described, 765 698 766 699 begun. Banner implemented this new process across eight the use of modular furniture to create physician examina- 767 700 different EDs with varying volumes and saw reductions in tion cubicles, recliners to create a processing area, and 768 701 the LWBS rates of 30–60% across the board. Other movable room dividers to create a triage pod are all exam- 769 702 abstracts depicting patient segmentation models were ples of low-cost changes to the physical space to accom- 770 703 submitted by ultra-high-volume EDs, Christiana Care in modate process changes and innovations. Some of the 771 704 772 705 Wilmington, Delaware and Beaumont in Royal Oaks, most dramatic improvements were seen when changes 773 706 Michigan. Abstracts using variations of the patient to the physical plant were married to process changes. 774 707 streaming/patient satisfaction concept were the second Technology has been shown to have a role in improv- 775 708 most frequent process change seen in the learning collab- ing intake and flow. In 2005, Chan et al. used a series of 776 709 orative and competition. process changes involving new technology to improve 777 710 778 711 intake (41). They dubbed these changes the REACT 779 712 The Philadelphia EMS Admission Rule (PEAR).The Uni- project: Rapid Entry and Care at Triage. Their technolog- 780 713 versity of Pennsylvania developed a tool for EMS use to ical innovations included the bar-coding of laboratory 781 714 help predict whether or not a transported patient would be specimens and patient IDs, the ability to access old med- 782 715 admitted from the ED. The PEAR rule uses routine infor- ical records at intake to help create an identifier for the 783 716 784 717 mation obtained at dispatch to predict the likelihood of patient, and information technology (IT) interfaces to Q2 785 718 admission for a patient. An aggregate score between allow access to all available medical information at in- 786 719 0 and 14 is generated based on the presence or absence take. The innovations were associated with significant 787 720 of chest pain, dyspnea, dizziness, or syncope, age over improvements in overall LOS and LWBS (41). 788 721 60 years, diabetes, or cancer. The model was first trialed Innovators in our learning collaborative employed 789 722 790 723 at one site and then repeated at multiple locations. The similar technology, improving on previous work by creat- 791 724 area under the receiver operator curve for the PEAR ing an IT tool with self-populating data fields to expedite 792 725 tool in discriminating between admission and discharge intake. In an even more futuristic model, the Carolinas 793 726 was 0.83 at six hospitals. Patients with a PEAR score of Medical Center is using palmar scanning to create a 794 727 9 or higher had a near 100% chance of being admitted. unique ID for each patient. This allows diagnostic testing 795 728 796 729 and treatments to be ordered before a patient has been for- 797 730 DISCUSSION mally registered. The use of communication devices, in- 798 731 cluding radios, wearable devices, and electronic tracking 799 732 We used the learning collaborative model for health care systems to enhance communication, has been associated 800 733 improvement to convene a group of ED leaders, policy- with mixed results (42–46). However, one innovative 801 734 802 735 makers, and innovators for the purpose of studying team found that by simply using radios to alert the 803 736 strategies for improving ED intake. In a unique imple- physician of a patient arrival in triage, they moved the 804 737 mentation of the collaborative model, we brought innova- physician encounter earlier in the patient’s visit and saw 805 738 tors together with leading health care authorities to dramatic improvement in LWBS rates. Perhaps the 806 739 exchange ideas around this topic. The innovators com- ultimate use of technology was demonstrated by the 807 740 808 741 pared and competed with their ideas in a competition, Medical College of Georgia. Telemedicine enabled 809 742 and a survey of new ideas is presented here as a review. them to do the intake assessment while the patient was 810 743 This is not definitive work, but early iterations of the in another location. 811 744 new models are showcased. Traditional triage typically involves eight or nine re- 812 745 The concept that the physical space in the ED can be dundant steps occurring in series that have limited value 813 746 814 747 transformed to improve work flow is not new. In 2002, to patient care (Figure 3). The use of process change to 815 748 816

FLA 5.1.0 DTD - JEM8728_proof - 11 April 2011 - 6:08 pm - ce Strategies to Improve ED Intake 7 817 885 818 886 819 887 820 888 821 889 822 890 823 891 824 892 825 893 826 894 827 895 828 896 829 897 830 898 831 899 832 900 833 901 834 902 835 903 836 904 837 905 838 906 839 907 840 908 841 909 842 910 843 911 844 912 845 913 846 914 847 915 848 916 849 917 850 918 851 919 852 920 853 921 854 922 855 923 856 924 857 925

858 FPO 926 859 = 927 860 928 861 929 web 4C 862 Figure 3. --- Q10 930 863 931 864 932 865 streamline this intake process is being explored in both placing a physician in triage decreases LOS, decreases 933 866 small and large EDs. In particular, variations on the use LWBS rates, and increases staff satisfaction, and that 934 867 of physicians in triage and team triage are becoming one-third or more of patients can be rapidly discharged 935 868 trialed. There is evidence of success using either model using few or no resources (58–60). 936 869 (47–51). In our learning collaborative, models that moved the 937 870 938 871 There are ample data to support the placement of the physician encounter earlier in the patient intake process, 939 872 highest level of training at intake. Paramedics correctly either as physician triage or team triage, seemed to be the 940 873 predict whether or not patients will need to be admitted most common area of process change. Another idea in- 941 874 from the ED 62% of the time (52). Kosowsky et al. re- corporated into the new models is that of placing patients 942 875 ported that nurses predict a patient’s disposition with into categories based on acuity and anticipated resource 943 876 944 877 slightly better accuracy than paramedics (53). On the utilization, often termed ‘‘patient streams’’ or ‘‘patient 945 878 other hand, there is a growing body of evidence demon- segmentation.’’ This approach has yielded notable im- 946 879 strating that physicians’ assessments of outcome and provement in length of stay, LWBS, and diversion times 947 880 disposition are highly reliable, with 85–95% accuracy (61). It is effectively an expansion of the ‘‘fast track’’ con- 948 881 (54–57). Dedicating a physician to the ED intake has cept, one of the most successful strategies used in emer- 949 882 950 883 a number of advantages. Studies have shown that gency medicine to improve patient flow (62–66).The 951 884 952

FLA 5.1.0 DTD - JEM8728_proof - 11 April 2011 - 6:08 pm - ce 8 S. Welch and L. Savitz 953 1021 954 1022 955 third concept woven into many of the abstract models in Available at: http://www.pressganey.com/galleries/default-file/2008 1023 956 the collaborative competition is that of ‘‘team care,’’ pulse report. Accessed October 18, 2009. 1024 957 6. Matters Urgent. Engaging staff to reduce ED wait times to 30 min- 1025 a concept that has a history of effectiveness in hospital- utes: recognition, rewards and review of data energizes Chicago’s 958 1026 Mount Sinai. Patient flow e-newsletter 2006;3(4). Available at: 959 based care (67–69). 1027 http://urgentmatters.org/346834/318749/318750. Q5 960 7. Busy ED keeps promise of ‘door to doc’ in 31 minutes. Hosp Case 1028 961 Limitations Manag 2008;16:87–9. 1029 962 8. Bursch B, Beezy J, Shaw R. Emergency department satisfaction: 1030 963 The results presented by the innovators were not statisti- what matters most? Ann Emerg Med 1993;22:586–91. 1031 9. Thompson DA, Yarnold PP, Williams DR, et al. Effects of actual 964 cally tested. The abstracts were presented as self reports 1032 965 waiting time, perceived waiting time, information delivery and ex- 1033 966 without oversight or review. Even the definitions of the pressive quality on patient satisfaction in the emergency depart- 1034 ment. Ann Emerg Med 1996;28:657–65. 967 performance metrics used were not standardized. The in- 1035 10. Boudreaux ED, D’Autremont S, Wood K, et al. Predictors of emer- 968 novations presented here may not be reproducible else- gency department patient satisfaction: stability over 17 months. 1036 969 where, and may not hold up under intense statistical Acad Emerg Med 2004;11:51–8. 1037 970 analysis, but they represent the early trials of a wide vari- 11. Patel PB, Vinson DR. Team assignment system: expediting emer- 1038 971 gency department care efficiency. Ann Emerg Med 2005;46: 1039 972 ety of institutions struggling to improve on a process that 499–506. 1040 973 is recognizably faulty. 12. Kyriacou DN, Ricketts V, Dyne PL, McCollough MD, Talan DA. A 1041 974 5-year time study analysis of emergency department patient care 1042 975 efficiency. Ann Emerg Med 1999;34:326–35. 1043 CONCLUSIONS 13. Goodacre S, Webster A. Who waits longest in the emergency de- 976 1044 partment and who leaves without being seen? Emerg Med J 2005; 977 1045 Although there may be little Class 1 evidence to direct ED 22:93–6. 978 14. Khot UN, Johnson ML, Ramsey C, et al. Emergency department 1046 979 leaders and practitioners regarding front-end operations, physician activation of the catheterization laboratory and immediate 1047 980 the innovation continues with energy and creativity transfer to an immediately available catheterization laboratory re- 1048 981 (70). Each ED faces similar problems regarding capacity, duce door to balloon time in ST-elevation myocardial infarction. 1049 Circulation 2007;116:67–76. 982 limited resources in an era of growing public expecta- 1050 983 15. Diericks DB, Roe MT, Chen AY,et al. Prolonged emergency depart- 1051 984 tions, boarding, ancillary service delays, and staffing ment stays of non-ST segment elevation myocardial infarction pa- 1052 tients are associated with worse adherence to the American 985 and space constraints. 1053 Cardiology/American Heart Association guidelines for manage- 986 Innovations involving intake and the front end are oc- ment and increased adverse events. Ann Emerg Med 2007;50: 1054 987 curring in three main areas: changes to the physical space, 489–96. 1055 988 integration of technology to improve process, and process/ 16. Semplicine A, Benetton V, Macchini E, et al. Intravenous thrombol- 1056 989 ysis in the emergency department for the treatment of acute ische- 1057 990 flow re-design. Whether or not such innovation needs to be mic stroke. Emerg Med J 2008;25:403–6. 1058 991 vetted using the traditional medical research model is 17. Wardlaw JM, Zoppo G, Yamaguchi T, et al. Thrombolysis for acute 1059 992 a question worth considering. Our work supports the learn- ischaemic stroke. Cochrane Database Syst Rev 2009;(4): 1060 993 CD000213. 1061 ing collaborative as a model for sharing new innovations 18. Dean NC, Bateman KA, Donnelly SM, et al. Improved clinical out- 994 1062 and disseminating successful ideas in their earliest itera- comes with utilization of a community acquired pneumonia guide- 995 1063 tions. line. Chest 2006;13:794–9. 996 19. Fee C, Weber EJ, Maak CA, et al. Effect of emergency department 1064 997 crowding on time to antibiotics in patients admitted with commu- 1065 998 Acknowledgments—The Emergency Department Benchmarking nity acquired pneumonia. Ann Emerg Med 2007;50:501–9. 1066 999 Alliance (EDBA) and the Intermountain Institute for Health 20. Pines JM, Hollander JE. Emergency department crowding is associ- 1067 1000 Care Delivery Research applied for and were awarded a grant ated with poor care for patients with severe pain. Ann Emerg Med 1068 2008;51:1–5. 1001 from AHRQ (grant No. 1R13HS018126(01) to study intake 1069 1002 21. Bernstein SL, Aronsky D, Duseja R, et al. The effect of emergency 1070 and hold a summit for the dissemination of new ideas. department crowding on clinically oriented outcomes. Acad Emerg 1003 1071 Med 2008;16:1–10. 1004 22. Welch S. Be Thomas Edison: invent solutions for the ED. EMNews 1072 1005 REFERENCES 2010;32:16. 1073 1006 23. Institute for Healthcare Improvement (IHI). The breakthrough se- 1074 1007 1. Clancy CM. Emergency departments in crisis: implications for ries: IHI’s collaborative model for achieving breakthrough improve- 1075 1008 quality and safety. Am J Med Qual 2007;22:59–62. ment. IHI Innovations Series 2003. Cambridge, MA: Institute for 1076 1009 2. Pitts SR, Niska RW, Xu J, Burt CW. National Hospital Ambulatory Healthcare Improvement; 2003. Available at: http://www.ihi.org/ 1077 1010 Medical Care Survey: 2006 emergency department summary. Natl NR/rdonlyres/BCA88D8F-35EE-4251-BB93-E2252619A06D/0/ 1078 1011 Health Stat Report 2008;(7):1–38. Available at: http://www.cdc. BreakthroughSeriesWhitePaper2003:Accessed Sept. 30, 2009. 1079 24. Michalke JA, Patel SG, Siler Fisher A, et al. Emergency department 1012 gov/nchs/ahcd/emergency_department_visitshtm; 2008:Accessed 1080 April 15, 2010. size determines the demographics of emergency department pa- 1013 1081 3. VHA ED Process Database and VHAOn-Line Survey 11/2006, Sol- tients [abstract]. Ann Emerg Med 2005;46(Suppl):39. 1014 nd 1082 Q3 ucient’s Action 0–1 Database, 2 quarter 2006. 25. Siler Fisher A, Hoxhaj S, Patel SG, et al. Predicting patient volume 1015 4. The Joint Commission. Sentinel event alert. Issue 26: Delays in per hour. Ann Emerg Med 2005;46(Suppl):6–7. 1083 1016 Q4 treatment. June 17, 2002. 26. Welch SJ, Jones SS, Allen TA. Mapping the 24-hour emergency de- 1084 1017 5. Emergency department pulse report 2008: patient perspectives on partment cycle to improve patient flow. Jt Comm J Qual Patient Saf 1085 1018 American healthcare. South Bend, IN: Press Ganey Associates. 2007;33:247–55. Q6 1086 1019 1087 1020 1088

FLA 5.1.0 DTD - JEM8728_proof - 11 April 2011 - 6:08 pm - ce Strategies to Improve ED Intake 9 1089 1157 1090 1158 1091 27. Hoxhaj S, Jones LL, Fisher AS, et al. Nurse staffing levels affect the 49. Choi YF, Vong TW, Lau CC. Triage rapid assessment by doctor 1159 1092 number of emergency department patients that leave without treat- (TRIAD) improves waiting time and processing time of the emer- 1160 1093 ment. Acad Emerg Med 2004;11:459–63. gency department. Emerg Med J 2006;23:262–5. 1161 28. Jensen K, Welch S, Mayer T, Haraden C. Leadership for smooth pa- 50. Mayer T. Team, triage, and treatment (T3). Unpublished quality im- 1094 1162 tient flow (ACHE Management Series). Chicago: Health Adminis- provement data, Fairfax Inova Hospital. Presented at ED Bench- 1095 1163 Q7 tration Press; 2006. marks 2005 Conference, Orlando, FL, March 4, 2005. 1096 29. Jones S, Allen TA, Welch S, et al. An independent evaluation of four 51. Richardson JR, Braitberg G, Yeoh MJ. Multidisciplinary assess- 1164 1097 quantitative emergency department crowding scales. Acad Emerg ment at triage. Emerg Med Australas 2004;16:41–6. 1165 1098 Med 2006;13:1204–11. 52. Levine SD, Colwell CB, Pons PT, et al. How well do paramedics 1166 1099 30. Jones S, Welch S, Allen T, et al. Forecasting daily patient volumes in predict admission to the hospital? Emerg Med J 2006;31:1–5. 1167 110 0 the emergency department. Acad Emerg Med 2008;15:159–71. 53. Kosowsky JM, Schindel S, Liu T, et al. Can emergency department 1168 1101 31. Welch SJ, Augustine J, Camargo C, et al. Performance measures nurses predict patients’ dispositions? Am J Emerg Med 2001;19: 1169 1102 and benchmarking summit. Acad Emerg Med 2006;13:1074–80. 10–4. 1170 32. Welch SJ, Davidson SJ. Exploring new intake models into the emer- 54. Sinuff T, Adhikari NK, Cook DJ, et al. Mortality predictions in the 1103 1171 gency department. Am J Med Qual 2010;25:172–80. intensive care unit: comparing physicians with scoring systems. Crit 1104 33. Welch SJ, Asplin BR, Stone-Griffith S, Davidson SJ, Augustine J, Care Med 2006;34:878–85. 1172 1105 Schuur J. Emergency department metrics, measures, and defini- 55. Rocker G, Sjokvist P, Weaver B, et al. Clinician predictions of inten- 1173 1106 tions: results of the Second Performance Measures and Benchmark- sive care unit mortality. Crit Care Med 2004;32:1149–54. 1174 1107 ing Summit. Ann Emerg Med 2010 Nov 8 [Epub ahead of print]. 56. Rodriguez RM, Wang NE, Pearl RG. Prediction of poor outcome of 1175 1108 34. Welch SJ, Augustine J, Savitz L, et al. Size matters: census and intensive care unit patients admitted from the emergency depart- 1176 1109 acuity predict emergency department performance. JAMA 2010. ment. Crit Care Med 1997;25:1801–6. 1177 1110 Q8 Under review. 57. Dent AW, Weiland TJ, Vallender L, et al. Can medical admission 1178 1111 35. Swensen SJ, Meyer GS, Nelson EC, et al. Cottage industry to post- and length of stay be accurately predicted by emergency staff, 1179 industrial care: the revolution in health care delivery. N Engl J Med patients or relatives? Aust Health Rev 2007;31:633–41. 1112 1180 2010;362:e12. 58. Taylor D, Bennett DM, Cameron PA. A paradigm shift in the nature 1113 36. James BC. Every defect a treasure. Med J Aust 1997;166:484–7. of care provision in emergency departments. Emerg Med J 2004;21: 1181 1114 37. James BC. Breaks in the outcomes measurement chain. Hosp Health 681–4. 1182 1115 Netw 1994;68:60. 59. Besson K. Care initiation area yields dramatic results. ED Manag 1183 1116 38. Doolan DF, Bates DW, James BC. The use of computers for clinical 2009;21:28–30. 1184 1117 care: a case series of advanced us sites. J Am Med Inform Assoc 60. Vega V, McGuire SJ. Speeding up the emergency department: the 1185 1118 2003;10:94–107. RADIT emergency program at St. Joseph Hospital of Orange. 1186 1119 39. Spaite DW, Bartholomeaux F, Guisto J, et al. Rapid process redesign Hosp Top 2007;85:17–24. 1187 1120 in a university based emergency department: decreasing wait times 61. King DL, Ben-Tovim DI, Basham J. Redesigning emergency de- 1188 and improving patient satisfaction. Ann Emerg Med 2002;39:168–77. partment patient flows: application of Lean Thinking to health 1121 1189 40. Morgan R. Turning around the turnarounds: improving ED through- care. Emerg Med Australas 2006;18:391–7. 1122 put processes. J Emerg Nurs 2007;33:530–6. 62. Sanchez M, Smalley AJ, Grant RJ, et al. Effects of a fast track area 1190 1123 41. Chan TC, Killeen JP, Kelly D, et al. Impact of rapid entry and accel- on emergency department performance. J Emerg Med 2006;31: 1191 1124 erated care at triage on reducing emergency department wait times, 117–20. 1192 1125 length of stay and rate of left without being seen. Ann Emerg Med 63. Rodi SW, Grau MV, Orsini CM. Evaluation of a fast track unit: 1193 1126 2005;46:491–7. alignment of resources and demand results in improved satisfaction 1194 1127 42. Le MM, Zwemmer FL, Dickerson VJ, et al. Providing mobile and decreased length of stay for emergency department patients. 1195 1128 phones to emergency medicine residents; perceived effects on phy- Q Manage Health Care 2006;15:163–70. 1196 1129 sician communication and work. Ann Emerg Med 2004;44:S28. 64. Simon HK, McLario D, Daily R, et al. ‘‘Fasttracking’’patients in an 1197 43. Walsh B, Yamarick WK. Beam me up Scotty. A new emergency de- urban pediatric emergency department. Am J Emerg Med 1996;14: 1130 1198 partment in Ohio goes live with a wearable push-button communi- 242–4. 1131 cation system on opening day, reducing noise, improving staff 65. Darrab AA, Fan J, Fernandes CM. How does fast track affect quality 1199 1132 communication and increasing patient privacy. Health Manag Tech- of care in the emergency department? Eur J Emerg Med 2006;13: 1200 1133 nol 2005;26:24–6. 32–5. 1201 1134 44. Aronosky D. Supporting patient care in the emergency department 66. Kwa P, Blake D. Fast Track: has it changed patient care in the emer- 1202 1135 with a computerized whiteboard system. J Am Med Inform Assoc gency department? Emerg Med Auatralas 2008;20:10–5. 1203 1136 2008;15:184–94. 67. Risser DT, Rice MM, Salisbury ML, et al. The potential for im- 1204 1137 45. Welch SJ, Dalto J. Improving door-to-physician times in 2 commu- proved teamwork to reduce medical errors in the emergency depart- 1205 1138 nity hospital emergency departments. Am J Med Qual 2011;26: ment. Ann Emerg Med 1999;34:373–83. 1206 138–44. 68. Sexton JB, Thomas J, Heimrich RL. Error, stress and teamwork in 1139 1207 46. Boger E. Electronic tracking board reduces ED patient length of medicine and aviation: cross-sectional surveys. BMJ 2000;320: 1140 stay at Indiana hospital. J Emerg Nurs 2003;29:39–43. 745–9. 1208 1141 47. Terris J, Leman P, O’Connor N, et al. Making an IMPACT on emer- 69. Morey JC, Simon R, Jay GD, et al. Error reduction and performance 1209 1142 gency department flow: improving patient processing assisted by improvement in the emergency department through formal team- 1210 1143 consultant at triage. Emerg Med J 2004;21:537–41. work training: evaluation results of the Med Teams Project. Health 1211 1144 48. Travers JP, Lee FC. Avoiding prolonged waiting time during busy Serv Res 2002;37:1553–81. 1212 1145 periods in the emergency department: is there a role for the senior 70. Wiler JL, Fite DL, Gentle C, et al. Optimizing emergency depart- 1213 1146 emergency physician in triage? Eur J Emerg Med 2006;13:342–8. ment front-end operations. Ann Emerg Med 2010;55:142–60. 1214 1147 1215 1148 1216 1149 1217 1150 1218 1151 1219 1152 1220 1153 1221 1154 1222 1155 1223 1156 1224

FLA 5.1.0 DTD - JEM8728_proof - 11 April 2011 - 6:08 pm - ce 10 S. Welch and L. Savitz 1225 1293 1226 1294 1227 1295 1228 ARTICLE SUMMARY 1296 1229 1297 1230 1. Why is this topic important? 1298 1231 More than three-quarters of emergency departments 1299 1232 spend part of everyday over capacity (more patients 1300 1233 than treatment spaces). When this happens they need 1301 1234 strategies for seeing the backlog of patients. 1302 1235 1303 1236 2. What does this study attempt to show? 1304 1237 Unlike clinical progress, operational progress is occur- 1305 1238 ring at the front lines, in non-academic settings. These in- 1306 1239 novations are being trialed around the country and the 1307 1240 ideas are being aired before they reach the literature. 1308 1241 1309 1242 3. What are the key findings? 1310 1243 The waits and delays at intake may be addressed 1311 1244 through three main approaches: Changing the physical 1312 1245 plant, changing the process, or changing the technology. 1313 1246 The best results were seen with marriages of the three 1314 1247 1315 1248 types of innovation. 1316 1249 4. How is patient care impacted? 1317 1250 Moving patients more quickly to the encounter with a 1318 1251 provider improves the ability to treat the time-dependent 1319 1252 clinical entities appropriately, improves patient outcomes, 1320 1253 1321 1254 and decreases patients leaving without being seen. 1322 1255 1323 1256 1324 1257 1325 1258 1326 1259 1327 1260 1328 1261 1329 1262 1330 1263 1331 1264 1332 1265 1333 1266 1334 1267 1335 1268 1336 1269 1337 1270 1338 1271 1339 1272 1340 1273 1341 1274 1342 1275 1343 1276 1344 1277 1345 1278 1346 1279 1347 1280 1348 1281 1349 1282 1350 1283 1351 1284 1352 1285 1353 1286 1354 1287 1355 1288 1356 1289 1357 1290 1358 1291 1359 1292 1360

FLA 5.1.0 DTD - JEM8728_proof - 11 April 2011 - 6:08 pm - ce Our reference: JEM 8728 P-authorquery-v9

AUTHOR QUERY FORM

Journal: JEM Please e-mail or fax your responses and any corrections to:

E-mail: [email protected]

Article Number: 8728 Fax: 1-845-896-0870

Dear Author,

Please check your proof carefully and mark all corrections at the appropriate place in the proof (e.g., by using on-screen annotation in the PDF file) or compile them in a separate list. To ensure fast publication of your paper please return your corrections within 48 hours.

For correction or revision of any artwork, please consult http://www.elsevier.com/artworkinstructions.

Any queries or remarks that have arisen during the processing of your manuscript are listed below and highlighted by flags in the proof.

Location Query / Remark: Click on the Q link to find the query’s location in text in article Please insert your reply or correction at the corresponding line in the proof Q1 Is this okay as a running title? If not, please provide one. Q2 Is “information technology” the correct definition of “IT”? Q3 Please provide more information for ref. #3, e.g., publisher info or URL where source can be accessed. Q4 If ref. #4 is available online, please provide URL. Q5 Because ref. #6 is an electronic source, a URL has been added: please provide most recent date URL was accessed. Q6 The article title for ref. #26 has been changed to match that shown in PubMed; please check. Q7 Ref. #28 in the manuscript was a duplicate of #26, so it has been deleted and subsequent references re- numbered. Q8 Please update ref. #34 if published. Q9 Please provide a descriptive legend for Figure 2. Q10 Please provide a descriptive legend for Figure 3.

Thank you for your assistance.