Improving the Efficiency of Hospital-Based Emergency Care
The emergency care system is but one component of the larger health care delivery system and of the even larger social safety net system. As such, it is subject to many forces far beyond its direct control. There is little that emergency care providers and advocates can do to alter such environmental factors as growing use of the emergency department (ED) by the uninsured; the increasing age and number of chronic conditions of patients; staffing shortages in many key areas, especially nurses and on-call specialists; malpractice insurance rates that grew on average more than 50 percent between 2002 and 2003 (AMA, 2003); and declining public and private reimbursements—not to mention disasters, both natural and man-made. There is, however, a great deal that the emergency care system can do to anticipate, prepare for, and manage the effects of these broader trends. This chapter explores strategies for improving the efficiency of hospital-based emergency care within the context of the broader health care delivery system, with a focus on the special issue of patient flow. The chapter also examines approaches to overcoming barriers to improved ED patient flow and operational efficiency. The committee emphasizes the compelling need for regulatory and policy changes to increase accountability and incentivize the efficient management of patient flow throughout the hospital and beyond.
THE ED IN THE CONTEXT OF THE HEALTH CARE DELIVERY SYSTEM
Medical science in the United States is arguably the most advanced in the world, but the organization and delivery of health care lags well behind
many other U.S. industries in terms of innovation, use of information technology, and management practices. Kleinke (1998, p. 6) described medical delivery in the United States as “…a miracle of disorganization, held together through the sheer collective will of overworked professionals tasked with managing tens of millions of patients by memory, pen scrawl, Post-It note, and telephone call.” It is a system that, to quote Berwick (1996, p. i3), “is perfectly designed to achieve exactly the results it gets.” The results, as documented by the Institute of Medicine (IOM) reports To Err Is Human: Building a Safer Health System (IOM, 2000) and Crossing the Quality Chasm: A New Health System for the 21st Century (IOM, 2001), include an estimated 98,000 deaths and more than 1 million injuries each year as a result of health care process and system failures (Starfield, 2000). According to the joint National Academy of Engineering (NAE) and IOM (2005, p. 1) report Building a Better Delivery System: A New Engineering/Health Care Partnership, “an estimated thirty to forty cents of every dollar spent on health care…a half trillion dollars a year…is spent on costs associated with: overuse, underuse, misuse, duplication, system failures…and inefficiency.” While confidence in American medicine remains strong, patients understand that the delivery system is failing. In a survey conducted by the Picker Institute (2000), 75 percent of patients described a system that was fragmented; difficult to navigate; and inconsistent in terms of information, evidence, and treatment.
According to the NAE/IOM report, the U.S. health care system retains a “cottage industry” structure, with physicians and other health care providers operating semiautonomously. As a result, hospitals and other provider organizations lack the hierarchical control of the typical business enterprise, making it difficult to introduce efficiency principles to streamline flows in such areas as production, inputs, and inventory as in other industries. In addition, the prevalent payment structures in health care, which focus on individual encounters and practice settings, tend to reinforce silos, reward inefficient practices, and discourage investment in new technologies and process improvements. As a result, innovations that have swept through other sectors of the economy, including banking, airlines, and manufacturing, have failed to take hold in health care delivery—a sector of the economy that now consumes 16 percent of the nation’s gross domestic product and is growing at twice the rate of inflation. Health care information technology has advanced considerably in the last decade, but mainly in the administrative and financial arenas, as opposed to the core processes of delivering clinical services (NAE and IOM, 2005).
Other industries have made use of a number of tools derived from engineering and operations research, which can be referred to collectively as operations management tools (see Box 4-1). Manufacturers, airlines, banks, the military, and others have adopted systems that employ a number of these
tools. For example, Motorola’s Six Sigma process and the Toyota Production System combine statistical and process controls with worker empowerment and cultural change to minimize defect rates and achieve high levels of quality. Some of these approaches have been promoted and implemented by health care organizations, such as the Joint Commission on Accreditation of Healthcare Organizations (JCAHO), the Veterans Health Administration, Kaiser Permanente, the National Association of Public Hospitals and Health Systems (NAPHHS), the Agency for Healthcare Research and Quality (AHRQ), and several private hospital organizations. But adoption of such approaches has yet to become widespread (Gabow et al., 2005; National Association of Public Hospitals and AHRQ, 2005).
A common thread among these tools is the systems concept, in which the dependence of every component on the others is recognized. To achieve the system’s maximum performance, each unit must not only achieve high individual performance, but also cooperate with interdependent units to optimize system objectives. The tools of operations management facilitate the understanding of complex systems and make it possible for managers to control and improve overall system performance.
Nowhere is the interdependence among individual components more evident and the need for tools to manage complex systems more crucial than in the hospital ED. Taking care of emergency patients involves many discreet components, such as registration, emergency physicians, nurses, laboratory services, imaging, inpatient departments, and on-call specialists. These components are highly interdependent, such that optimizing the performance of any one without considering the broader objectives of the system is unlikely to improve the overall performance of the delivery of emergency care. For example, optimizing care in an inpatient department may slow admissions from the ED, worsen ED crowding, and create a host of associated problems. Indeed, that is what often happens.
UNDERSTANDING PATIENT FLOW THROUGH THE HOSPITAL SYSTEM
Crowding in the nation’s EDs poses a serious threat to the quality, safety, and timeliness of emergency care. While many of the factors contributing to ED crowding are outside the immediate control of the hospital, many more are the result of operational inefficiencies in the management of hospital patient flow. EDs receive an almost steady stream of patients. If an individual arriving by ambulance cannot be transferred quickly to an ED stretcher, efficiently triaged, and then rapidly evaluated, stabilized, and admitted or discharged, ED crowding will quickly develop, and patient care will be compromised.
Operations Management Tools
Many operations management tools could be applied to achieve better management of patient flow:
Quality functional deployment. This iterative process links stakeholder needs to the resources required to meet those needs throughout the organization. Conflicting demands on the organization emerge and are resolved, with all relevant stakeholders examining the trade-offs from a systems perspective. The process has been used in a variety of industrial applications, including integrated circuit and automobile design.
Failure modes and effects analysis (FMEA). FMEA is a formal process for analyzing failures that might occur under varying conditions so they can be avoided through design features. It has been used in manufacturing for more than 30 years and has recently been applied to health care. The Veterans Health Administration encourages its accredited hospitals to use FMEA or hazard analysis tools in a required annual proactive risk assessment of at least one high-risk process each year.
Root-cause analysis is a qualitative, retrospective variation on FMEA that has been widely used to analyze industrial accidents. The Joint Commission on Accreditation of Healthcare Organizations requires accredited hospitals to use the method to evaluate sentinel patient safety events.
Human factors engineering. This set of techniques attempts to integrate human behavior and limitations into process design. Human factors research has been widely used across industries and has had many recent applica
Hospital administrators and policy makers have at their disposal a number of promising options for identifying and resolving the patient flow problems that contribute to ED crowding and its consequences. But these leaders must first be compelled to take action, something that will occur only when the causes of ED crowding are clearly understood, and administrators realize that the strategies required to address the problem go well beyond the ED itself. More than 15 years ago, Lynn and Kellermann (1991) described approaches to improving management of the ED in an overcrowded hospital. Key to their thesis, then as now, was the idea that crowding is an inpatient problem that manifests itself in the ED. Accordingly, measures to address crowding should begin on inpatient units, rather than with diversion of inbound ambulances. Moreover, administrators, policy makers, and the public must have the knowledge, incentives, and regulatory obligations needed to inspire change.
tions in health care, such as medication administration, diagnosis, handoffs of patients between shifts, and telemedicine.
Queuing theory. Queuing theory is used to determine the capacity of services that are subject to variable demand over time. It has been widely used in a number of service industries, such as banking and public transportation. It has had limited use in health care, but has been applied to optimize scheduling and staffing in primary care, operating rooms, nursing homes, radiology departments, and emergency departments (Huang, 1995; Siddharthan et al., 1996; Reinus et al., 2000; Lucas et al., 2001; Gorunescu et al., 2002; Murray and Berwick, 2003; McManus et al., 2004; Green et al., 2006). (See the detailed discussion in Box 4-2.)
Supply-chain management. This set of techniques helps match resources with demand in highly complex production processes. Companies such as Dell, Toyota, and Procter & Gamble represent enormously complex systems that use supply-chain management tools, such as linear integer programs, to optimize performance. Airlines use these models to assign crews to thousands of flights per day across hundreds of cities. The techniques have revolutionized production in many industries but have had very little impact in the hospital environment despite substantial successes. For example, both Vanderbilt University Medical Center and Deaconess Hospital in Evansville, Indiana, have achieved substantial savings using these techniques. It has been estimated that the health care industry could save $11 billion by using supply-chain management (NAE and IOM, 2005).
Statistical process control. This technique involves plotting the outcomes of a process over time to see whether variations fall within an acceptable range or fall outside that range and require corrective action. It is widely used in manufacturing.
From arrival in the ED to hospital admission or discharge, emergency patients receive treatment at multiple points of the care delivery process. Patient flow, defined as the movement of patients through this system, is an important indicator of the timeliness, safety, and quality of the care received. Efficient patient flow ensures maximum throughput (the number of patients treated and discharged from the ED per day), minimizing delays at each point of the delivery process with no decrement in the quality of care. Impaired patient flow, on the other hand, results in bottlenecks that prolong delays for patients already in the system, as well as those awaiting entry.
The input/throughput/output (I/T/O) model of patient care, based on engineering principles from queuing theory and compartmental models of flow, applies operations management concepts to patient flow within the acute care system (see Figure 4-1). The I/T/O model defines the acute care system as including unscheduled ambulatory care, urgent care, ED care and
its ancillary services, inpatient care for those admitted through the ED, and out-of-hospital emergency medical services (EMS) care. In this way, the I/T/O model allows for the identification of all components of the health care system that contribute to or are affected by ED crowding (Asplin et al., 2003; Solberg et al., 2003).
Under the I/T/O model, ED input, or demand, comprises three distinct categories of care: emergency care (treatment of seriously ill or injured patients), unscheduled urgent care (treatment of patients unable to receive needed care in a timely manner from other components of the acute care system), and safety net care (treatment of patients who experience substantial barriers to accessing unscheduled care from other components of the health care system). Variations in the demand for each of these types of care, both patient- and systems-driven, determine the input fluctuations in the ED. That is, ED input levels depend on both the volume of critically ill and injured patients and the ability of the overall health care system to care for nonemergent and safety net patients (Asplin et al., 2003; Solberg et al., 2003).
The throughput component of the I/T/O model represents a patient’s length of stay in the ED and comprises two key phases: (1) triage, room placement, and medical evaluation, and (2) diagnostic testing and ED treatment. ED boarding is also included in the throughput component as it extends ED lengths of stay. The output component of the model represents the disposition of ED patients. It includes hospital admission, transfer to another facility, and patient discharge. It also includes the ability of the ambulatory health care system to provide timely and appropriate postdischarge care (Asplin et al., 2005).
As designed, the structure of the I/T/O model allows hospitals to systematically identify and resolve impediments to patient flow across a spectrum of acute care settings. It also provides direction for researchers, policy makers, and hospital administrators seeking to understand and alleviate ED crowding as a way to improve access to and quality of care (Asplin et al., 2003; Solberg et al., 2003; Wilson et al., 2005).
IMPEDIMENTS TO EFFICIENT PATIENT FLOW IN THE ED
While hospitals are unable to control forces outside the facility that contribute to high levels of demand, they can understand the impact of those forces on operations and structure their organization for optimal response. At the same time, hospitals have direct control over a number of variables that affect operational efficiency, including such factors as inpatient bed capacity, ancillary service delays, the scheduling of surgeries and support staff, and provision of adequate physical space in the ED to permit evaluation and treatment (GAO, 2003). By applying variability methodology, queuing
theory, and the I/T/O model, hospitals can identify and eliminate many of the impediments to patient flow caused by operational inefficiencies (Litvak and Long, 2000; Litvak, 2005).
One of the most important factors currently outside the control of most hospitals is the regional flow of patients (see Chapter 3). Short of the need to go on diversion, there is typically little information sharing between hospitals and EMS regarding overloaded EDs and trauma centers and the availability of ED beds, operating suites, equipment, trauma surgeons, and critical specialists. Such information is needed to balance the patient load among EDs and trauma centers in a region, which requires that many elements within the regional system—community hospitals, trauma centers, and particularly prehospital EMS—effectively coordinate the regional flow of patients. In addition to improving patient care, coordinating the regional flow of patients is a critical tool for reducing overcrowding in EDs. Unfortunately, only a handful of systems around the country coordinate transport effectively throughout their region. Some examples were described in Chapter 3.
Inpatient Admissions Bottlenecks
The most commonly cited contributor to ED crowding is the inability to move admitted patients from the ED into inpatient hospital beds, in particular intensive care unit (ICU) beds. This lack of inpatient beds has the immediate effect of forcing ED staff to “board” admitted patients until an inpatient ICU or medical-surgical bed is available (see Chapter 1). Placing ED patients who require hospital admission in hallways or examination spaces temporarily until an inpatient bed becomes available is a poor substitute for inpatient care. EDs are not designed to provide privacy to hallway boarders, and staff are often too busy to meet an admitted patient’s needs in a timely manner. Moreover, boarding is the primary cause of ambulance diversion, a practice that delays access to emergency care and can send inbound patients to a hospital where the medical staff does not know them and has no access to their medical records. Ambulance diversion also contributes to reduced EMS capacity as ambulances seeking to offload patients are forced to find an open ED and once there, to wait until the ED staff are able to find an empty stretcher (Gallagher and Lynn, 1990; Thorpe, 1990; Andrulis et al., 1991; Derlet and Richards, 2000; Epstein and Slate, 2001; Derlet et al., 2001; Henry, 2001; Viccellio, 2001; The Lewin Group, 2002; McManus et al., 2003; Asplin et al., 2003; GAO, 2003; Schull et al., 2003; Solberg et al., 2003; Weissert et al., 2003; Eckstein and Chan, 2004; JCAHO, 2004; Kennedy et al., 2004; see also Chapter 1). By failing to manage patient flow effectively, hospitals allow the most time-critical access point in the facility—the ED—to become blocked and ultimately inaccessible.
In addition to contributing to an overall shortage of bed space, the current reimbursement structure discourages hospitals from making provision of inpatient beds to ED admissions a management priority. Within the hospital, ED patients compete for beds, staff, and services with patients who have been scheduled for elective admission, particularly elective surgical patients and those being admitted for invasive diagnostic or therapeutic procedures. When beds are scarce, elective admissions generally prevail because they pay better margins and promote loyalty among admitting physicians. ED admissions typically generate less revenue for the hospital, and may even cost the hospital money. Furthermore, since these patients are already in the system, they are unlikely to leave, whereas an elective admission can choose to go to another hospital. Finally, because hospitals benefit financially from increased volume (up to a point), there is a financial disincentive to hold vacant beds open for ED admissions.
Delays in Ancillary Services
Enhanced standards of care and improved medical technology mean that today’s ED patients routinely receive a number of complex diagnostic and screening services (McCaig and Burt, 2005). Whether complex or routine, the timely administration of these ancillary services and the prompt availability of test results are imperative for smooth hospital operations and efficient patient flow. Data suggest, however, that delays in diagnostic and screening tests for ED patients are both common and strongly associated with prolonged lengths of stay in the ED. In fact, nearly one-half of all ED service delays were related to wait times for radiology and laboratory results according to one survey conducted by the Emergency Nurses Association (Derlet and Richards, 2000; Weissert et al., 2003; JCAHO, 2004). Housekeeping also is frequently a problem, as most ED admissions occur in the late afternoon to early morning hours, while housekeeping staffs are usually reduced after 5:00 PM.
Overuse of ED Services
Physicians treating patients in the ED have access to a wide range of complex medical screening and evaluation tools, all within the confines of a single physical space—the hospital. This means that ED patients often also have access to the best technology in the community, as hospitals are frequently more able than local providers or smaller health clinics to purchase and operate expensive medical equipment. These factors have resulted in the dual effect of some patients opting to seek care in the ED and some primary or specialty care providers referring their patients to the
ED as a means of streamlining the medical testing process. In short, the ED is assuming, by default, another new role—that of “one-stop shop” for complex medical workups, a phenomenon that improves the efficiency of office-based practitioners, but contributes to ED crowding and hinders the safety and timeliness of true emergency care. Also, because EDs often have limited access to patient records, redundant workups and diagnostic tests are often performed.
The rise in the number and severity of medical malpractice claims, especially in high-risk fields such emergency medicine, has led to an increasingly defensive approach to providing care in the ED. Because emergency physicians have such a range of tests and diagnostic technologies at their fingertips, they are more likely to be blamed if they fail to use them and ultimately miss a diagnosis. For example, missed myocardial infarction has been the leading cause of malpractice claims in emergency medicine, yet definitively excluding the possibility of a myocardial infarction or acute coronary syndrome requires a minimum of 6–12 hours of evaluation and diagnostic tests costing more than a thousand dollars. Fearing potential litigation, ED physicians and on-call specialists may order additional tests or prolong monitoring periods, slowing patient flow and contributing to service delays. ED staff may also hospitalize patients in borderline condition rather than running the risk that a discharged patient will have an adverse outcome. This is even more likely to happen when the physician is concerned that the patient may not be able to secure outpatient follow-up care in a timely manner (Asplin et al., 2005). It should be noted, however, that it is difficult to quantify the increment over and above appropriate evaluation in emergency care that constitutes “defensive medicine.”
In contrast to the strict nurse-to-patient ratios on many inpatient units and ICUs, most hospitals have declined to adopt nurse-to-patient ratios for the ED. As a result, an inpatient unit that has vacant beds but has reached its maximum ratio of nurses to patients may block admissions from an ED that may be caring for two or even three times as many patients per nurse. The merits of staffing ratios in general are discussed in Chapter 6.
Inadequate Physical Space
Unlike most high-risk enterprises, health care has been slow to embrace principles of ergonomics or human factors engineering in the design and
maintenance of its various workplaces. As a result, ED providers often face limitations on the amount of space available in which to provide care, and they routinely encounter user-unfriendly spatial layouts and equipment placement and design. In many hospitals, for example, computed tomography (CT) scans, operating rooms, or ICUs are located a significant distance from the ED, requiring the staffed transport of patients across multiple hospital divisions or floors. Similarly, desktop-only registration, whiteboard tracking, and land-line phone paging systems routinely pull physicians and other staff away from the bedside, extending patients’ lengths of stay and leading to disruptions in the course of care. Fortunately, many of these design failures can be addressed through the adoption of new information technology tools (McKay, 1999; Chisholm et al., 2000; Derlet and Richards, 2000; Wears and Perry, 2002). Additional discussion of these tools is provided in Chapter 6.
STRATEGIES FOR OPTIMIZING EFFICIENCY
A number of initiatives now under way are aimed at improving patient flow in order to reduce ED crowding and its related effects. These include Urgent Matters, a $6.4 million, 10-hospital campaign supported by The Robert Wood Johnson Foundation that aims to eliminate ED crowding and improve public understanding of challenges facing the health care safety net; the IHI IMPACT Network, which, through its Improving Flow Learning and Innovation Community, seeks to increase patient throughput and minimize delays while ensuring that high performance in flow is not achieved at the expense of quality; and the University HealthSystem Consortium (UHC) Patient Flow Benchmarking Project, which targets in-hospital factors that impede or impair efficient patient flow. Recognizing the importance of managing patient flow to addressing ED crowding, JCAHO published a new standard for accredited hospitals: “LD.3.11.” “The leaders develop and implement plans to identify and mitigate impediments to efficient patient flow throughout the hospital” (JCAHO, 2004).
Based on the above efforts, a wide range of tools have been developed and tested to address patient flow issues, generally with good success. While controlled studies have yet to be conducted, a growing body of anecdotal evidence suggests that by smoothing the peaks and valleys of patient flow (the movement of patients into and between various hospital areas for care), hospitals can reduce crowding while improving quality and reducing cost (JCAHO, 2004; Wilson and Nguyen, 2004). Boston Medical Center and St. John’s Regional Health Center in Springfield, Missouri, for example, reduced crowding by adjusting elective surgery schedules so they did not conflict with predictable peaks in emergency surgeries (Litvak and Long, 2000; Crute, 2005).
Techniques That Address Bottlenecks in Patient Flow
The effective management of patient flow in the ED and between the ED and hospital inpatient units is essential to the quality and safety of patient care (Begley et al., 2004). By smoothing the inherent peaks and valleys of patient flow and eliminating the artificial variabilities that unnecessarily impair that flow, hospitals can minimize the occurrence of queues and improve safety and quality while simultaneously reducing hospital waste and costs (Litvak and Long, 2000). For inherent, patient-driven peaks and valleys, the necessary ED capacity (number of beds, nurses, ancillary services) can be determined by applying queuing theory (see Box 4-2). This approach leads to greater predictability and control and ultimately to improved quality, safety, and timeliness of care (Litvak and Long, 2000; NAE and IOM, 2005).
A number of additional techniques have been tested for improving the
Queuing theory applies analytical expressions to problems involving waiting times, or queues, that develop because of limited resources. Its purpose is to understand and achieve a balance between fixed capacity and the random demands of customer services. Queuing models have long been used in a number of industries, including telecommunications, the Internet, commercial banking, sales, and public transportation. They are increasingly being recognized as a tool that can help identify and manage the variabilities in patient flow that contribute to crowding in the emergency department (ED) (Litvak, 2005; NAE and IOM, 2005).
Many basic queuing models comprise three variables: arrival rate, service time, and number of servers. In the ED setting, the arrival rate is the frequency of patient arrivals, while service time is the average time spent caring for a particular type of patient at a specific point of care in the ED and its related sites. The number of servers can be the number of stations, beds, nurses, or work areas providing similar services to all patients who enter those areas (NAE and IOM, 2005). The problem, however, is that service time frequently has two components: the average time spent caring for patients and boarding time. Since boarding time is frequently a result of artificial variability in hospital patient flow (artificial peaks in inpatient bed census), basic queuing models cannot be applied to determining adequate ED resources. Thus to determine true (versus inflated) resources needed, one must exclude boarding time from the service time (length of stay in the ED).
Coordinated Surgery Schedule
The two most common routes to hospital admission today are through the ED (e.g., 50 percent) and through scheduled elective surgery in the operating room (OR) (e.g., 35 percent). Variability in admissions is well documented and leads to substantial fluctuations in inpatient capacity. For many hospitals, periods of limited capacity are often followed by periods of excess capacity, and managing this variability has the potential to improve patient flow and ED crowding significantly (DeLia, 2006). While the natural variability associated with emergency care might lead one to assume the ED is responsible for most of the fluctuations in inpatient traffic, data demonstrate that scheduled elective surgery in the OR, when adjusted for patient volume, is in fact the more variable of the two admission routes, thereby creating a significant artificial component of the variability in case volume (Litvak and Long, 2000; Litvak, 2005). Coordinating surgery times for scheduled and unscheduled admissions therefore not only adds organization to the rate and flow of scheduled elective OR admissions, but also allows hospitals to smooth out variabilities in ED and OR patient flow—an effect that serves to alleviate ED crowding (Litvak and Long, 2000; Cedars-Sinai Learns, 2004; Wilson and Nguyen, 2004; Litvak, 2005).
Many of the hospitals participating in the Urgent Matters, IHI, and UHC patient flow initiatives have undertaken systematic reviews and revamping of OR scheduling as a way of improving patient flow; enhancing the quality, safety, and timeliness of emergency care; reducing unnecessary costs; and increasing surgical revenue. Two related tactics are among those employed most frequently by these hospitals: (1) setting aside one OR for unscheduled surgical cases admitted through the ED and (2) smoothing the elective surgery schedule by distributing surgery times more evenly across the entire week (Litvak and Long, 2000). Both techniques have significantly reduced waiting times for surgical cases, especially among ED patients. This in turn has reduced the amount of time ED patients must wait for an inpatient bed, easing ED crowding and its effects. Improved coordination of surgery schedules also has been associated with increased revenue for surgeons, an important compensation for the disruption to the surgeons’ schedules (Litvak, 2005; Crute, 2005).
Coordinated Bed Management
One strategy that has been successful in smoothing patient flow and alleviating ED crowding is the creation of “bed czars” or “bed teams” charged
Case Study: Boston Medical Center, Boston, Massachusetts
Boston Medical Center (BMC) is a private, nonprofit academic medical center that serves as the primary teaching affiliate for the Boston University School of Medicine. It has nearly 500 licensed beds and is the largest safety net hospital in New England, with an annual operating budget of $1 billion. BMC offers an array of medical services, including a level I trauma center, full-service acute care, pediatric care, and cardiothoracic surgery. Its emergency department (ED), staffed by 26 full-time physicians, treats over 120,000 patients annually.
As recently as 2003, BMC experienced significant ED crowding and ambulance diversion and high rates of patients leaving without being seen. To alleviate these conditions, BMC initiated a comprehensive project to identify and address inefficiencies in hospital operations, particularly those that inhibited patient flow. Before embarking on the initiative, BMC chief executive officer (CEO) Elaine Ullian established a project stakeholders group that included, among others, hospital leadership, the chiefs of surgery and anesthesiology, and key nursing staff. Ullian also convened several issue-focused teams, including an inpatient team, an ED team, and a surgery schedule smoothing team.
BMC employed a rapid cycle change (RCC) model, in which small changes are rapidly implemented and evaluated by staff. The study team first identified a specific aim or goal intended to improve patient flow. It then
with various aspects of bed management. Typically a nurse manager, a bed czar has primary responsibility for accounting for inpatient beds and working with housekeeping to ensure rapid bed turnaround. To fulfill this responsibility, bed czars are given authority to notify staff of impending bed shortages, make decisions regarding inpatient bed transfers, cancel elective procedures, and initiate hospital diversion. Bed teams, on the other hand, usually consist of nurses from multiple units, each of whom has access to real-time hospital census data. Working collaboratively, these teams meet throughout the day to discuss the types of ED patients waiting for inpatient beds and the types of beds expected to become available, making flow changes as necessary (JCAHO, 2004; Wilson and Nguyen, 2004; Wilson et al., 2005).
Among its many advantages, the bed czar or bed team approach offers a consistent, timely mechanism through which hospital staff can be notified about bed status; a centralized patient placement process; and improved ability to anticipate bed needs across multiple settings. Use of coordinated bed management techniques has been associated with significant reductions
developed, implemented, and evaluated strategies on a small scale, modifying or rejecting the approach based on the results obtained. For example, one goal of the BMC team was to reduce ED throughput time. In response to suggestions from the nursing staff and nurse manager, the team decided to test a “zone nursing” approach in which nurses were assigned to patients in a particular area of the ED. Historically, ED nurses at BMC had been assigned to patients randomly, meaning that each nurse typically was responsible for a number of patients located throughout the ED. After a week-long, small-scale trial, the zone approach was associated with a 70-minute reduction in ED throughput time. In response to this success, the BMC team subsequently decided to extend the zone approach to the entire ED.
Another BMC project goal was to smooth surgery schedule variations in order to improve operating room (OR) and ED throughput. The team worked with the Cardiothoracic Surgery Department and Vascular Surgery Section to reduce peaks in elective surgical case volume; place a daily cap on the number of elective surgeries; switch surgeons’ clinic and surgery days; and dedicate one of the hospital’s eight ORs to emergent cases, with the other seven being open for block scheduling. The resulting improvements in patient flow through the ORs were significant; the number of “bumped” surgical cases, for example, fell from 337 between April and September 2003 to 3 between April and September 2004. At the same time, BMC ambulance diversion rates declined by 40 percent and overall ED throughput times by 17 percent.
SOURCE: Wilson et al., 2005.
in bed turnaround times at a number of EDs nationwide. The Regional Medical Center in Memphis, Tennessee, for example, reduced its average bed turnaround time by nearly 70 percent, cutting wait times from 150 to 47 minutes (JCAHO, 2004; Wilson and Nguyen, 2004; Wilson et al., 2005).
Efficiencies can also be achieved by use of a transfer center to coordinate referrals to a tertiary center. Such a center can reduce delays for transfer patients in the ED, ensure the availability of timely resources needed by such patients, and help coordinate transfers between facilities (Southard et al., 2005).
Clinical Decision Units (CDUs), or Observation Units
CDUs, or observation units, are separate areas that allow for the observation of patients to determine whether admission is necessary. Originally, these units were developed to provide a method for monitoring patients with chest pain who had a low to intermediate probability of acute myocardial
Case Study: Grady Health Systems, Atlanta, Georgia
Grady Health Systems, comprising Grady Memorial Hospital, Hughes Spalding Children’s Hospital, and 10 regional health centers, is one of the largest public hospitals in the southeastern United States. Licensed for more than 1,000 beds, Grady Memorial Hospital (Grady) houses the only level I trauma center within a 100-mile radius, the state’s only poison control center, and the city of Atlanta’s emergency medical services (EMS) ambulance fleet. Grady also serves as the teaching hospital for both Emory and Morehouse schools of medicine. More than 100,000 visits are made to the Grady emergency department (ED) each year.
In 2002, as ED patient satisfaction levels fell to historic lows, Grady found itself experiencing a number of significant ED crowding–related challenges. Average ED throughput times, for example, frequently exceeded 7 hours, with fast-track throughput times reaching 10 hours. Rates of patients leaving the ED without being seen were estimated at 2.4 percent, or 200 patients per month. And by 2003, Grady’s ED was operating under diversionary status more than 20 percent of the time.
Attempting to turn the tide on these trends, Grady used the input/through-put/output (I/T/O) model to identify major bottlenecks in patient flow. Under the direction of a project steering committee, led jointly by the hospital chief executive officer and chief operating officer, the Grady team developed and implemented a number of approaches involving a wide range of staff. For example, Grady instituted a new diagnostic test ordering process whereby
infarction (AMI) (Zwiche et al., 1982; Fineberg et al., 1984; Talbot-Stern et al., 1986; Vallee et al., 1988; de Leon et al., 1989; Henneman et al., 1989; Mikhail et al., 1997; Rydman et al., 1998; Graff et al., 2000). By observing patients for up to 23 hours, ED staff were able to rule out many patients at risk of AMI while using fewer resources than if these same patients were admitted to the ICU or an inpatient telemetry unit (Graff et al., 1997). Today, observation units are used most frequently for the efficient management of patients with complaints of chest pain, abdominal pain, back pain, dehydration, congestive heart failure, asthma, and shortness of breath (Hostetler et al., 2002; Ross et al., 2003). They are typically overseen full time by a nurse practitioner with assistance from an attending physician and other nursing staff. CDUs have been shown to reduce costs associated with inpatient admissions (Mikhail et al.,1997; Rydman et al., 1998; Graff et al., 2000), although the net impact on hospital costs is unclear (Sinclair and Green, 1998). One recent study found that approximately 30 percent
requests were handled by the unit clerk rather than the charge nurse; under the new process, wait times for test results were reduced by as much as 95 minutes during periods of ED crowding. In addition, Grady improved staff coordination and training in its fast-track unit; these changes were associated with signification reductions in the average time from ED arrival to bed placement (from 219 to 94 minutes) and average ED throughput time (from 340 to 211 minutes) for fast-track patients.
Finally, Grady implemented a care management unit (CMU), consisting of seven beds staffed by four CMU nurses and four case managers, to which patients diagnosed with one of four conditions—chest pain, heart failure, asthma, or hyperglycemia—are assigned. This dual CMU–ED structure allows for faster treatment and longer-term observation of nonemergent patients. Following their CMU stay, which lasts an average of 19 hours, 85 percent of patients are discharged, while 15 percent are admitted as inpatients. Prior to hospital discharge, CMU patients are assigned a case manager who provides disease-specific education, coordinates primary care follow-up (defined as occurring within 48–72 hours of discharge), and directs follow-up via telephone, as well as performing various data management chores. Among other benefits, the establishment of the Grady CMU has resulted in decreases in the number of short-stay admissions, admissions to telemetry beds, and patient relapse rates. The CMU also has resulted in improved patient satisfaction with Grady’s ED services.
SOURCES: Grady Health System, 2005; Wilson et al., 2005.
of hospitals and two-thirds of teaching hospitals had opened or planned to open a CDU (Mace et al., 2003).
CDUs offer the potential to alleviate crowding in EDs and add elements of continuity to patient care. These units care for patients who would otherwise be admitted for inpatient stays two to three times as long. This frees up beds for other patients who would otherwise be boarded in the ED (Schneider et al., 2001), which in turn leads to a reduction in diversion hours (Dick et al., 2005). Use of CDUs may also allow ED staff to downgrade the type of bed required for those who still need admission after a CDU stay—instead of a telemetry or stepdown bed, admitting the patient to a regular medical-surgical bed.
Some units combine the concept of a CDU with the concept of case management. Such units employ case managers to focus on patients with exacerbations of chronic diseases that are known as “ambulatory care sensitive conditions” (e.g., asthma, diabetes, congestive heart failure). The
Case Study: St. John’s Regional Health Center, Springfield, Missouri
St. John’s Regional Health Center is an 866-bed, not-for-profit hospital and trauma center that serves as the dominant health care center in southwestern Missouri and parts of northwestern Arkansas. There were 74,000 visits to St. John’s emergency department (ED) in fiscal year 2005, with approximately 22 percent of all ED patients requiring hospital admission. During the same time, ED-based admissions accounted for roughly 20 percent of the hospital’s total surgical load.
In 2002, hospital leaders faced two significant patient flow–related problems. First, an inflexible process for scheduling elective surgeries had resulted in unpredictable and excessive use of overtime. Second, midweek peaks in surgery demands had resulted in admissions backups that were causing patients to be placed in beds on the wrong floors, jeopardizing the safe delivery of appropriate postsurgical care.
Seeking to resolve these issues, St. John’s set aside a single operating room (OR) for elective and unplanned surgery overflow. This required the hospital’s trauma surgeons to give up an OR that historically had been set aside in case they decided to schedule a surgery the day after their on-call period. The surgeons agreed on the condition that if no noticeable improvements were achieved during a 30-day trial period, the OR would be returned for their use. At the conclusion of the trial period, St. John’s was able to increase the number of elective and unplanned surgeries by 5.1 percent, the number of OR rooms required after 3:00 PM dropped by 45 percent, and hospital trauma surgeons experienced a 4.6 percent increase in revenue. Based on this success, the OR change was made permanent.
assumption is that a diabetic with a blood sugar level of 700 mg/dL needs not only CDU care with the goal of avoiding hospitalization, but also case management, because the episode of hyperglycemia is a sentinel event for the failure of ambulatory care. While patients are getting hydrated or receiving an infusion of insulin, they are also being taught self-care skills and being reconnected with a primary care provider for close outpatient follow-up. Case managers can follow up with patients after discharge to make sure they keep their appointments. The goal is not only to prevent an expensive hospitalization, but also to reduce relapse rates and repeat visits to the ED due to another hyperglycemia/asthma/congestive heart failure episode by reconnecting the patient to primary care. In this way, the CDU aids the hospital in managing patient flow and reducing crowding while at the same time contributing to the smooth functioning of the ambulatory care system.
A second trial modified the elective surgery schedule, booking elective orthopedic surgeries evenly throughout the week. Although many surgeons initially objected to the plan and the physical therapy staff were required to adjust their work schedules, the change resulted in a 13 percent increase in the number of patients able to move to the appropriate floor for recovery. It also provided a number of surgical specialties, including orthopedics, with additional hours of OR block time. As with the OR change, modifications to the surgery schedule were made permanent following the successful trial period.
As a result of the above changes, the hospital was able to increase its surgical case volume by 7–11 percent annually with no capital investment over 3 years. The hospital administration attributes a number of recent operational, financial, and quality improvements to the continued success of the smoothing of elective surgeries for all surgical subspecialties (Personal communication, C. Dempsey, March 21, 2006):
SOURCE: Crute, 2005.
Hospitals can receive reimbursement for CDUs for three conditions: chest pain, asthma, and congestive heart failure. For other conditions, reimbursement for observation care is packaged or bundled into other ambulatory payment classification (APC) rates and not listed separately. Many groups, including the Society for Academic Emergency Medicine (SAEM) and the American College of Emergency Physicians (ACEP), have encouraged the Centers for Medicare and Medicaid Services (CMS) to expand separate payments for observation services beyond the three conditions currently allowed, claiming that the literature supports the effectiveness of observation services for many other conditions. Further, an APC advisory panel appointed by CMS unanimously recommended removing restrictions on diagnoses and conditions eligible for separate payment of observation; however, CMS has not enacted this change (Personal communication, M.B.
McClellan, July 8, 2005). While Medicare CDU payments would increase with the addition of eligible conditions, total costs of care should decline because of the reduction in the number of admissions. For Medicare, the change would be cost-saving.
On the basis of the foregoing evidence, the committee concludes that CDUs reduce the need for boarding and diversion, avoid expensive hospitalizations, and appear to contribute to improved management of common ambulatory care sensitive conditions. The committee believes CDU payments should be available for all clinical conditions for which observation is indicated, and therefore recommends that the Centers for Medicare and Medicaid Services remove the current restrictions on the medical conditions that are eligible for separate clinical decision unit payment (4.1).
Unit Assessment Tools
Unit assessment tools, based on the traffic light concept, can be used to determine and monitor the capacity of various units throughout the hospital system. The tool comprises graded, color-coded indicators that note the “workload tolerances” of each unit, based on a preset range of numerical scores. Under the system, green (go) indicates the unit is working at 85 percent of maximum capacity and therefore open for admissions; yellow (early caution) indicates the unit is working at >85 percent capacity and, though it is still able to accept admissions, alerts other units of current resource limitations; orange (late caution) indicates the unit is working immediately below its maximum capacity and suggests that capacity could be reached unless additional resources are made available; and red (stop) indicates the unit is at full capacity and cannot accept additional admissions without risking patient safety and staff burnout (JCAHO, 2004). Routine updates of the color grid allow staff to reallocate resources in response to status changes. This is accomplished most easily by a “resource czar,” typically the nurse supervisor, with the authority to redirect staff or cap a unit as necessary. Using the unit assessment tool model, Luther Midelfort, a Mayo Health Systems hospital in Eau Claire, Wisconsin, saw steady declines in the number of red codes and steady increases in the number of green codes during a recent 6-month trial period (JCAHO, 2004).
Coordinated Patient Discharge
One of the most widely recognized bottlenecks in patient flow is the discharge process. By expediting discharge in a coordinated way, hospitals can better prepare patients for discharge, improve turnaround of vacant beds, and align vacancies with bed demands more accurately—all of which
help alleviate crowding in the ED. Hospitals can alleviate discharge-related patient flow impediments through the creation of “discharge coordinator” positions and “discharge resource rooms.” Much like a bed czar, a discharge coordinator can monitor charts to determine which patients are ready for discharge and work to expedite the disposition process. This coordinator, usually a nurse, also can provide or facilitate case management services. A discharge resource room is an area of the hospital where staff help patients prepare for their home care after discharge in a comfortable, central location. Upon arrival at the discharge room, patients are considered discharged from the hospital, making their bed available for rapid turnaround (JCAHO, 2004; Wilson and Nguyen, 2004; Wilson et al., 2005).
A number of hospitals have been able to reduce discharge delays and alleviate related ED crowding following establishment of a discharge coordinator or discharge resource room. For example, one Chicago-area facility was able to reduce the average length of stay for some patients from 5.7 to 4.3 days, with concurrent reductions in ED crowding rates (JCAHO, 2004). Short of adopting coordinated discharge approaches, simply requiring physicians to write discharge orders earlier in the day can also result in a substantial improvement in patient flow.
Techniques That Address Care of Patients in the ED
An ED fast track is a dedicated area in or next to the ED that is specifically designed and designated for patients with minor illnesses or injuries. It is typically staffed by midlevel providers, such as physician assistants and nurse practitioners working under the supervision of an emergency physician. Fast tracks can operate during regular business hours or during the ED’s busiest times (e.g., evenings and weekends). Currently, fast tracks are in place at roughly 30 percent of all EDs, with approximately 30 percent of presenting patients being routed to these areas for care (JCAHO, 2004; Wilson and Nguyen, 2004). Identifying nonurgent patients and routing them to the fast track allows the ED to treat them more quickly. It also frees non–fast track ED resources to care for the most seriously ill and injured patients, moving them quickly into appropriate inpatient units. In this way, fast tracks can reduce delays in care for both urgent and nonurgent patients, thereby improving patient flow across the ED.
One example of the throughput time reductions associated with fast tracks is Grady Health Systems in Atlanta, Georgia (see Box 4-4). Using the fast-track approach, Grady was able to reduce the time from arrival to bed placement for nonurgent patients from 219 to 94 minutes, a 57 percent
decrease (JCAHO, 2004; Wilson and Nguyen, 2004). It is important to note that fast-track capacity may vary widely for different hospitals and should be determined according to the specific circumstances of each ED, such as volume, patient mix, and severity-of-illness levels.
Based on the engineering concept of collocation, zone nursing ensures that all of a nurse’s patients are located in one area, thereby eliminating the need for nurses to traverse a unit to provide care (JCAHO, 2004; Wilson and Nguyen, 2004). Explored by a number of hospitals nationwide, the zone approach has been found to reduce ED crowding. For example, as part of a pilot project at Boston Medical Center during which just one nurse received zone-approach assignments, the average patient throughput time was reduced by 70 minutes. Based on the success of the pilot, the approach was extended to the entire ED. A new version of the concept was recently initiated, involving zone collocation of both ED residents and nursing staff. Evaluation of this team approach is still under way (Wilson and Nguyen, 2004).
Bedside registration can help reduce long stays in the waiting room. Patients are quickly triaged in the reception area and immediately moved to a bed in the treatment area, where they can be seen immediately by a physician. In the treatment area, a computer on wheels allows staff to register the patient and collect insurance and other administrative information at the bedside, even after treatment has begun.
ED triage is typically performed by experienced emergency nurses, and sometimes by physicians. In crowded conditions, staff can feel pressure to perform triage quickly, creating opportunities for error. Some EDs are divided into separate areas—for example, pediatrics, obstetrics, and psychiatry—and triage is used to direct patients to the appropriate setting. Computer-enhanced triage is also being adopted by some hospitals to improve the reliability of triage decisions and expedite patient flow. These approaches are discussed in the next chapter.
Full-capacity protocols are plans put in place by hospitals to improve the treatment of patients and patient flow in conditions of extreme crowding due
to full inpatient units. Rather than keeping patients in the ED, perhaps in hallways and unsafe areas, full-capacity protocols allocate patients to inpatient beds in alternative units on a temporary basis. The approach recognizes the systemwide nature of ED crowding and requires that all departments share the responsibility for addressing crowding. Allocating patients to several different departments greatly improves conditions in already understaffed EDs, while the addition of one or two hall patients to several inpatient units has a minimal impact on those units’ staffing ratios. For example, adding two patients to a 30-bed unit with a 6- to 30-nurse staffing requirement yields a staffing ratio of 6.4—less than half a nurse below full staffing.
The State University of New York Stony Brook Hospital instituted this practice and found that a large percentage of patients never actually stayed in the hallway of the inpatient unit because staff were motivated to make beds available more quickly. Other patients spent less time in the inpatient unit than they would have had to spend in the ED. Early results showed that the average length of stay of ED hallway patients was 6.2 hours while that for unit hallway patients was 5.4 hours, and that patient satisfaction increased. A number of other institutions have adopted the practice, which is currently promoted by the New York State Department of Health.
An admission/discharge unit separate from the ED area has the potential to improve coordination of emergency patients and enhance patient flow. Such a unit provides several advantages to an ED. It can respond rapidly to the needs of the ED since it will always have the physical capacity to add a patient to the expandable ward, and it is not dependent upon the location of a patient’s physician for the writing of patient discharge orders. In addition, recently discharged patients remaining in the hospital are often poorly monitored and represent a liability exposure. Having a separate discharge unit greatly reduces this risk. Further, patients being staged for admission are conveniently located in one place for the staff to do their workups without taking resources (e.g., nursing, staff, space) from the ED.
Use of Information Technology
A number of approaches involving information technology can greatly enhance quality and efficiency in the ED. These include adoption of electronic health records with embedded error detection, patient tracking throughout the hospital system, “look-up” displays in critical care bays, health system–wide scheduling directly from the ED, and enhanced use of point-of-care testing and imaging. While many of these techniques are well established, others are in the early stages of development, and although they
show promise, their effectiveness is unproven in many cases. Implementation of such techniques should be supported and informed by a robust clinical and health services research agenda.
Timely Support for Consults and Procedures
Just as patients often wait for laboratory results and pharmaceutical deliveries, excessive amounts of time can be required for staff physicians to arrive for consults or minor (nonoperative) procedures. There are myriad reasons for these delays, but the general cause in many cases is a simple lack of planning and coordination, for example, failure to anticipate and staff for periods of high demand. Arrangements for specialists who provide on-call services are also critical. Lack of adequate on-call coverage can cause serious delays and compromise patient care. This issue is dealt with extensively in Chapter 6.
Given the wide range of tools available to improve the efficiency of hospitals, their potential benefit in alleviating emergency and trauma care crowding and enhancing quality, and their limited application in these settings to date, the committee believes adoption of such tools is crucial to improving the delivery of emergency care services. The committee therefore recommends that hospital chief executive officers adopt enterprisewide operations management and related strategies to improve the quality and efficiency of emergency care (4.2).
OVERCOMING BARRIERS TO ENHANCED EFFICIENCY
Although a growing body of evidence supports a range of strategies for improving patient flow and efficiency of operations while reducing ED crowding, a number of barriers exist to the adoption and implementation of these strategies within the hospital setting. The challenges to improving the efficiency of hospital-based emergency care are multiple, and the demands on physicians and administrators should not be taken lightly, particularly in light of the many other demands they face—for example, interdepartmental battles for resources, cost and revenue management, community relations, and a bewildering assortment of potential threats and opportunities. Despite the best intentions, hospitals face an uphill battle to focus sufficient attention on emergency care in the face of these other demands. Some of the specific challenges are discussed below.
Hospital Leadership Issues
Hospitals are extremely complex, highly political environments that present numerous leadership challenges for chief executive officers (CEOs)
and other executives. In many facilities, the clinical staff consists largely of independent agents working outside the traditional full-time staff structure (NAE and IOM, 2005). As a result, the vast majority of U.S. hospitals rely on clinicians who essentially serve as independent agents with distinct, and often disparate, agendas. Changes in the health care marketplace have resulted in many hospitals facing budget shortfalls, and the current reimbursement system offers little incentive for wholesale change. Added to these factors is the tenuous nature of most CEO appointments—data suggest the average hospital CEO tenure is just 6 years (American College of Healthcare Executives, 2004; Garman and Tyler, 2004)—and it is not surprising that many hospitals lack the leadership or support needed to embark on the systemwide analysis, innovation, and change necessary for improvements in patient flow.
Despite these challenges, it is clear that hospital leaders must be willing to lead if efforts to reduce ED crowding through improved patient flow and efficiency of operations are to succeed. Specifically, hospital leaders must recognize that ED crowding is a systemwide issue that must be addressed across hospital settings and is not limited to the ED itself. They must be willing to send a strong, consistent message that improving patient flow is a hospital priority. And they must back up those words with specific, demonstrable actions, including personal involvement in the development, implementation, and evaluation of patient flow improvement strategies.
Hospital executives, including both CEOs and midlevel managers, have an opportunity to provide visionary leadership in promoting patient flow and operations management approaches to improve hospital efficiency. The traditional paradigm of the ED as a safety valve for hospitalwide bottlenecks and inefficiencies is rapidly giving way to a modern view of the ED as an integrated component of a highly interconnected, organic system. Hospital leaders should be open to learning from the experiences of industries outside of health care and be bold and creative in applying these and other new ideas. The early evidence from The Robert Wood Johnson Foundation’s Urgent Matters project, IHI, and other such efforts suggests that not only does this view make sense for patients and providers, but it also makes sense for the bottom line.
To foster the development of hospital leadership in improving hospital efficiency, the committee recommends that training in operations management and related approaches be promoted by professional associations; accrediting organizations, such as the Joint Commission on Accreditation of Healthcare Organizations and the National Committee for Quality Assurance; and educational institutions that provide training in clinical, health care management, and public health disciplines (4.3).
Hospital clinicians, including those in the ED, tend to be conservative in nature and reluctant to embrace systemic change; efforts to identify and resolve barriers to patient flow through such strategies as those noted above are not likely to succeed without the early and strong support of hospital leaders, clinicians, and other staff. The recent failure of Cedars-Sinai Medical Center to implement a computerized provider order entry (CPOE) system demonstrates the magnitude and significance of this resistance. In November 2002, Cedars-Sinai began a 14-week, department-by-department rollout of its newly installed CPOE system. The rollout was called off and the system removed less than 2 months later following what has been characterized as a “staff revolt” (Chin, 2003; Cedars-Sinai Learns, 2004; Connolly, 2005).
The selection of a well-respected, highly regarded individual to serve as a champion for improved patient flow is an important step in ensuring the success of flow improvement strategies. Among other responsibilities, this individual can help sell the necessary changes to medical staff and executive managers. He/she can also help exert the constant pressure needed to reshape the policies, processes, relationships, and cultural norms that have historically impeded patient flow throughout the hospital.
The collection and analysis of reliable, comprehensive data concerning all aspects of patient flow is imperative if hospitals are to understand and resolve the factors contributing to crowding in their EDs. Currently, however, most hospital data systems do not adequately monitor or measure patient flow. For example, few systems distinguish between when a patient is ready to move to an ancillary location for care and when that move actually takes place—a limitation that prevents the capture and analysis of data on ED boarding, as well as other ED throughput delays.
Rigorous data collection and analysis is essential to the success of any patient flow improvement strategy. Using the I/T/O model, hospitals can identify key performance indicators for evaluating patient flow performance. Examples of such indicators used successfully by hospitals participating in the Urgent Matters initiative are time from inpatient bed assignment to bed placement, inpatient bed turnaround time, total ED throughput time, and time to thrombolysis for cardiac patients (Wilson et al., 2005). Other key performance indicators identified by the Government Accountability Office (GAO) as measures of ED crowding include the number of hours an ED is on ambulance diversion, the percentage of patients who board in the ED and for how many hours, and the number of patients
who leave the ED after triage but before a medical evaluation as a percentage of ED visits (GAO, 2003).
Research has shown that while the causes of ED crowding, boarding, and diversion are complex, the principal factors involved lie not in the ED itself but in inpatient departments to which ED patients are referred (Asplin et al., 2003). As a result, as noted earlier, it is increasingly understood that ED crowding is a systemwide issue that must be addressed across multiple hospital and acute care settings (Richardson et al., 2002; Asplin et al., 2003; Schafermeyer and Asplin, 2003; GAO, 2003; Magid et al., 2004). Thus it is not surprising that a key characteristic of successful patient flow improvement is the adoption of a systemwide approach to change. Such an approach includes, among other features, the development of a multidisciplinary, hospitalwide team that can work collaboratively to identify problems, propose solutions, and oversee the implementation and evaluation of various improvement strategies. (An example of a hospital team is shown in Figure 4-2.) Such an approach also includes timely collection and analysis of data at multiple points across several hospital settings to enable
evaluation of patient flow and assess changes in operations. Results of these analyses and outcome measures should be shared within and outside the hospital setting. Such transparency increases ownership and accountability among hospital leaders and staff; it also improves patient understanding of the complex, multidisciplinary nature of emergency care.
Alignment of Incentives
The degree of crowding and boarding that occurs in the ED would not be tolerated in inpatient departments. The strategies discussed above have the potential to improve patient flow significantly; enhance the quality, safety, and timeliness of emergency care; and produce related cost savings. Yet history has demonstrated that little progress will be made toward achieving these goals unless hospitals are held accountable through regulatory and incentive-based policies. Without such policies, hospitals will continue to marginalize patient flow matters, relegating most of the related consequences to EDs and their patients through crowding, prolonged periods of boarding, and ambulance diversions. There are a number of steps that can be taken by hospital leaders to address these issues, as well as policy initiatives that should be considered to align payment incentives with the goals of enhanced efficiency and quality of care.
No major change in health care can take place without strong financial incentives, and today hospitals have almost no incentives to address the myriad problems associated with inefficient patient flow or ED crowding. Indeed, as detailed below, hospitals have a number of financial incentives to continue the practices that lead to these problems.
Financial incentives must be instituted to ensure that hospitals act aggressively to eliminate ED crowding, boarding, and ambulance diversions. Rewarding hospitals that demonstrate efficient delivery practices that appropriately manage patient flow should be a consideration in reimbursement. All payers, including Medicare, Medicaid, and private insurers, should develop contracts that reward hospitals for efficient ED operations and penalize them for delays in hospital admission, for ED crowding, and for ambulance diversions. Through its purchaser and regulatory power, CMS has the ability to drive hospitals to address and manage patient flow and ensure timely access to quality care for its clients. Current CMS payment policies should be revised to reward hospitals that manage patient flow appropriately; conversely, hospitals that fail to do so should be subject to penalties.
Finally, CMS should evaluate the potential effect of existing diagnosis-
related group (DRG) payments on the relative priority assigned to elective patients and emergency admissions. Patients admitted from the ED are more likely to have a higher severity of illness, to be uninsured, or to have lower rates of reimbursement. Results of research undertaken at a small group of hospitals indicate that patients transferred to inpatient units and ICUs from the ED are more costly than elective patients for selected surgical DRGs (Munoz et al., 1985; Henry et al., 2003). A similar study found that patients transferred acutely to tertiary surgical ICUs were significantly more costly than elective admissions (Borlase et al., 1991). A disincentive to admit ED or transferred patients over elective patients may contribute to crowding and boarding in the ED. If such a disincentive exists, CMS should identify alternative payment methodologies to eliminate it.
Hospitals face virtually no reimbursement-related disincentives for operating a crowded ED. Indeed, they may benefit financially if this situation reduces Emergency Medical Treatment and Active Labor Act (EMTALA)-mandated admissions and preserves their capacity to admit elective patients. In 2004, JCAHO instituted new guidelines that would require accredited hospitals to take serious steps to reduce crowding, boarding, and diversion. This action followed a July 2002 alert that linked treatment delays to more than 50 deaths. Under pressure from the hospital industry, however, these requirements were withdrawn (Morrissey, 2004). They were replaced in January 2005 with a patient flow standard—Managing Patient Flow—that applies to the entire hospital. Among other things, this standard requires that hospitals develop plans and implement ways to monitor and manage patient flow that will reduce ED overcrowding and its consequences and ensure acceptable quality of care. Joint Commission Resources, an arm of JCAHO, has published a document aimed at educating hospital leadership about the new standard and providing guidance on how to comply with its provisions (JCAHO, 2004). While the new standard correctly acknowledges that patient flow is a system-level issue that must be addressed on a hospitalwide basis, it allows hospitals to continue using the ED as a holding area. Therefore, the committee recommends that the Joint Commission on Accreditation of Healthcare Organizations reinstate strong standards designed to sharply reduce and ultimately eliminate emergency department crowding, boarding, and diversion (4.4).
Not only do hospitals face no financial penalties for crowding and boarding, but there are several financial incentives that promote the practices that lead to these problems. First, a hospital benefits financially from increased volume (up to a point). Operating at high capacity is risky for any business because it means there is limited capacity available to deal with
spikes in demand. But the ED provides a convenient escape valve for hospitals operating at or near capacity. During periods of peak demand, patients can be cared for in the ED in relative safety because of the highly skilled and interdisciplinary staff that are available to deal with any exigency, staff that are used to a high-volume, high-pressure environment.
Second, according to a recent GAO report, one reason patients back up in the ED is that, as noted earlier, elective admissions for surgery or other procedures tend to be more profitable than emergency admissions through the ED (GAO, 2003). While many hospitals may not intentionally favor scheduled over ED admissions, which would potentially constitute an EMTALA violation, the GAO report found that only a minority of hospitals that diverted ambulances took other measures, such as postponing or canceling elective admissions.
Third, as discussed previously, patients admitted through the ED are more likely to be uninsured—indeed in many private hospitals, the only way an uninsured patient can be admitted is through the ED—and ED crowding has the effect of slowing the influx of uninsured and underinsured patients admitted through the ED.
Fourth, when hospitals hold emergency admits in the ED and instead give an available bed to the next elective patient, they essentially receive two inpatient reimbursements for the price of one because ED staff (a fixed cost) provide inpatient care at no additional cost to the hospital, while the elective patient gets the bed. Giving the ED admission priority over the elective one forfeits that advantage. Also, if the elective admission does not get the bed, the patient’s admitting physician may look to another hospital for admission. By contrast, ED admissions are “captive” in that they are already inside the facility and are too sick or injured to go elsewhere except in extreme circumstances.
Finally, when EDs are crowded in a community, especially if ambulances are being diverted and patients are walking away from the local public hospital or nonprofit equivalent, it can be financially perilous under EMTALA to have a “wide open” ED because uninsured and low-reimbursement patients are likely to flood the available ED. Although there is a paucity of data on the practice, some hospitals have been known to adopt “defensive diversion” to shield themselves from receiving diverted ambulance patients from the local public hospital. Further, some hospitals divert on a case-by-case basis—meaning they accept ambulances if the patient’s doctor is on the medical staff and refuse otherwise. While this practice constitutes an EMTALA violation, it is difficult to identify and pursue. In the absence of external regulatory mechanisms, monitoring of diversion status, and independent verification of how crowded the ED and hospital really are, it is impossible to limit this sort of practice.
The committee would like to see improved monitoring of hospital admission patterns by CMS to ensure that hospitals are not regularly using diversion while continuing to accept elective admissions. Such a practice would be in violation of EMTALA, and its prohibition should be strictly enforced (Medical Advisory Committee and Pennsylvania Emergency Health Services Council, 2004). Furthermore, the committee concludes that the practices of boarding and diversion are so antithetical to quality medical care that the strongest possible measures must be taken to eliminate them. Therefore, the committee recommends that hospitals end the practices of boarding patients in the emergency department and ambulance diversion, except in the most extreme cases, such as a community mass casualty event. The Centers for Medicare and Medicaid Services should convene a working group that includes experts in emergency care, inpatient critical care, hospital operations management, nursing, and other relevant disciplines to develop boarding and diversion standards, as well as guidelines, measures, and incentives for implementation, monitoring, and enforcement of these standards (4.5).
A final step in implementing the changes recommended by the committee is to make the public understand what is going on; appreciate the seriousness of the situation; know what questions to ask; and realize that the problem affects each individual, rich or poor, old or young, black or white, urban or rural. In short, the public needs to know what good performance is and understand who does and does not experience it. Hospitals should be required to measure key indicators of ED crowding and make those measures available to policy makers and the public. This could be accomplished through a variety of mechanisms, including patient flow performance report cards, public notices regarding diversion, and educational efforts focused on the unique and critical role served by safety net hospitals. For example, a community could provide “diversion alerts,” similar to storm alerts, to inform the public about EDs unable to accept new patients.
The reliance of EDs on other hospital units to eliminate ED crowding and its consequences through the effective management of patient flow demands a systemwide approach supported by hospital leaders and staff, policy makers, and the American public. Without immediate intervention, the quality, safety, and timelines of emergency care will continue to experience strain under the pressures of ED crowding, boarding, and diversion. Eliminating these pressures is no longer just a matter of convenience; it is a matter of life and death.
SUMMARY OF RECOMMENDATIONS
4.1: The Centers for Medicare and Medicaid Services should remove the current restrictions on the medical conditions that are eligible for separate clinical decision unit payment.
4.2: Hospital chief executive officers should adopt enterprisewide operations management and related strategies to improve the quality and efficiency of emergency care.
4.3: Training in operations management and related approaches should be promoted by professional associations; accrediting organizations, such as the Joint Commission on Accreditation of Healthcare Organizations and the National Committee for Quality Assurance; and educational institutions that provide training in clinical, health care management, and public health disciplines.
4.4: The Joint Commission on Accreditation of Healthcare Organizations should reinstate strong standards designed to sharply reduce and ultimately eliminate emergency department crowding, boarding, and diversion.
4.5: Hospitals should end the practices of boarding patients in the emergency department and ambulance diversion, except in the most extreme cases, such as a community mass casualty event. The Centers for Medicare and Medicaid Services should convene a working group that includes experts in emergency care, inpatient critical care, hospital operations management, nursing, and other relevant disciplines to develop boarding and diversion standards, as well as guidelines, measures, and incentives for implementation, monitoring, and enforcement of these standards.
AMA (American Medical Association). 2003. National Physician Survey of Professional Medical Liability. Chicago, IL: AMA.
American College of Healthcare Executives. 2004. Hospital CEO Turnover: 1981–2004. Chicago, IL: Health Administration Press.
Andrulis DP, Kellermann A, Hintz EA, Hackman BB, Weslowski VB. 1991. Emergency departments and crowding in United States teaching hospitals. Annals of Emergency Medicine 20(9):980–986.
Asplin BR, Magid DJ, Rhodes KV, Solberg LI, Lurie N, Camargo CA Jr. 2003. A conceptual model of emergency department crowding. Annals of Emergency Medicine 42(2):173–180.
Asplin BR, Rhodes KV, Levy H, Lurie N, Crain AL, Carlin BP, Kellermann AL. 2005. Insurance status and access to urgent ambulatory care follow-up appointments. Journal of the American Medical Association 294(10):1248–1254.
Begley CE, Chang YWRC, Weltge A. 2004. Emergency department diversion and trauma mortality: Evidence from Houston, Texas. The Journal of Trauma Injury, Infection, and Critical Care 57(6):1260–1265.
Berwick DM. 1996. A primer on leading the improvement of systems. British Medical Journal 312(7031):619–622.
Borlase BC, Baxter JK, Kenney PR, Forse RA, Benotti PN, Blackburn GL. 1991. Elective intrahospital admissions versus acute interhospital transfers to a surgical intensive care unit: Cost and outcome prediction. Journal of Trauma-Injury Infection & Critical Care 31(7):915–918; discussion 918–919.
Cedars-Sinai Learns from its CPOE Mistakes to Improve Workflow. 2004. [Online]. Available: http://www.bio-itworld.com/newsletters/healthit/2004/09/09/20040909_10115/ [accessed August 1, 2005].
Chin T. 2003, February 17. Doctors pull plug on paperless system. AMNews. [Online]. Available: http://www.ama-assn.org/amednews/2003-02/17/bil20217.htm [accessed May 20, 2006].
Chisholm CD, Collison EK, Nelson DR, Cordell WH. 2000. Emergency department workplace interruptions: Are emergency physicians “interrupt-driven” and “multitasking”? Academic Emergency Medicine 7(11):1239–1243.
Connolly C. 2005, March 21. Cedars-Sinai doctors cling to pen and paper. The Washington Post. P. A.01.
Crute S. 2005. Quality Matters, Case Study: Flow Management at St. John’s Regional Health Center. New York: The Commonwealth Fund.
de Leon AC Jr, Farmer CA, King G, Manternach J, Ritter D. 1989. Chest pain evaluation unit: A cost-effective approach for ruling out acute myocardial infarction. South Medical Journal 82(9):1083–1089.
DeLia D. 2006, in press. Annual bed statistics give a misleading picture of hospital surge capacity. Annals of Emergency Medicine.
Derlet R, Richards J. 2000. Overcrowding in the nation’s emergency departments: Complex causes and disturbing effects. Annals of Emergency Medicine 35(1):63–68.
Derlet R, Richards J, Kravitz R. 2001. Frequent overcrowding in U.S. emergency departments. Academic Emergency Medicine 8(2):151–155.
Dick RS, Schneider SM, Macdonald I. 2005. A cure for crowding: The impact of an emergency department observation unit on ambulance diversionary hours. Academic Emergency Medicine 12(5 Suppl. 1):10-a.
Eckstein M, Chan LS. 2004. The effect of emergency department crowding on paramedic ambulance availability. Annals of Emergency Medicine 43(1):100–105.
Epstein SK, Slate D. 2001. The Massachusetts College of Emergency Physicians ambulance diversion study. Academic Emergency Medicine 8(5):526–527.
Fineberg HV, Scadden D, Goldman L. 1984. Care of patients with a low probability of acute myocardial infarction. Cost effectiveness of alternatives to coronary-care-unit admission. New England Journal of Medicine 310(20):1301–1307.
Gabow P, Eisert S, Karkhanis A, Knight A, Dickson P. 2005. A Toolkit for Redesign in Health Care, Final Report. Rockville, MD: Agency for Healthcare Research and Quality.
Gallagher EJ, Lynn SG. 1990. The etiology of medical gridlock: Causes of emergency department overcrowding in New York City. Journal of Emergency Medicine 8(6):785–790.
GAO (U.S. Government Accountability Office). 2003. Hospital Emergency Departments: Crowded Conditions Vary among Hospitals and Communities. Washington, DC: GAO.
Garman AN, Tyler JL. 2004. CEO Succession Planning in Freestanding U.S. Hospitals: Final Report. Chicago, IL: Health Administration Press.
Gorunescu F, McClean SI, Millard PH. 2002. Using a queuing model to help plan bed allocation in a department of geriatric medicine. Health Care Management Science 5(4):307–312.
Grady Health System. 2005. About Grady. [Online]. Available: http://www.gradyhealthsystem.org/ [accessed July 1, 2005].
Graff LG, Dallara J, Ross MA, Joseph AJ, Itzcovitz J, Andelman RP, Emerman C, Turbiner S, Espinosa JA, Severance H. 1997. Impact on the Care of the Emergency Department Chest Pain Patient from the Chest Pain Evaluation Registry (CHEPER) Study. American Journal of Cardiology 80(5):563–568.
Graff LG, Prete M, Werdmann M, Monico E, Smothers K, Krivenko C, Maag R, Joseph A. 2000. Implementing emergency department observation units within a multihospital network. Joint Commission Journal on Quality Improvement 26(7):421–427.
Green LV, Soares J, Giglio JF, Green RA. 2006. Using queuing theory to increase the effectiveness of emergency department provider staffing. Academic Emergency Medicine 13(1):61–68.
Henneman PL, Marx JA, Cantrill SC, Mitchell M. 1989. The use of an emergency department observation unit in the management of abdominal trauma. Annals of Emergency Medicine 18(6):647–650.
Henry MC. 2001. Overcrowding in America’s emergency departments: Inpatient wards replace emergency care. Academic Emergency Medicine 8(2):188–189.
Henry MC, Thode HCJ, Havasy SP. 2003. Financial effects of emergency department admissions compared to electives within surgical diagnosis related groups (DRGs) at 11 hospitals in Suffolk County, NY. Academic Emergency Medicine 10(5):532-a.
Hostetler B, Leikin JB, Timmons JA, Hanashiro PK, Kissane K. 2002. Patterns of use of an emergency department-based observation unit. American Journal of Therapeutics 9(6):499–502.
Huang XM. 1995. A planning model for requirement of emergency beds. IMA Journal of Mathematics Applied in Medicine & Biology 12(3–4):345–353.
IOM (Institute of Medicine). 2000. To Err Is Human: Building a Safer Health System. Washington, DC: National Academy Press.
IOM. 2001. Crossing the Quality Chasm: A New Health System for the 21st Century. Washington, DC: National Academy of Sciences.
JCAHO (Joint Commission on Accreditation of Healthcare Organizations). 2004. Managing Patient Flow: Strategies and Solutions for Addressing Hospital Overcrowding. Washington, DC: Joint Commission Resources, Inc.
Kennedy J, Rhodes K, Walls CA, Asplin BR. 2004. Access to emergency care: Restricted by long waiting times and cost and coverage concerns. Annals of Emergency Medicine 43(5):567–573.
Kleinke JD. 1998. Bleeding Edge: The Business of Health Care in the New Century. Gaithersburg, MD: Aspen Publishers, Inc.
The Lewin Group. 2002. Emergency Department Overload: A Growing Crisis, the Results of the AHA Survey of Emergency Department (ED) and Hospital Capacity. Washington, DC: American Hospital Association.
Litvak E. 2005. Optimizing patient flow by managing its variability. In: JCAHO, From Front Office to Front Line: Essential Issues for Health Care Leaders. Oakbrook Terrace, IL: Joint Commission Resources, Inc.
Litvak E, Long MC. 2000. Cost and quality under managed care: Irreconcilable differences? American Journal of Managed Care 6(3):305–312.
Lucas CE, Buechter KJ, Coscia RL, Hurst JM, Meredith JW, Middleton JD, Rinker CR, Tuggle D, Vlahos AL, Wilberger J. 2001. Mathematical modeling to define optimum operating room staffing needs for trauma centers. Journal of the American College of Surgeons 192(5):559–565.
Lynn SG, Kellermann AL. 1991. Critical decision making: Managing the emergency department in an overcrowded hospital. Annals of Emergency Medicine 20:287–292.
Mace SE, Graff L, Mikhail M, Ross M. 2003. A national survey of observation units in the U.S. American Journal of Emergency Medicine 21(7):529–533.
Magid DJ, Asplin BR, Wears RL. 2004. The quality gap: Searching for the consequences of emergency department crowding. Annals of Emergency Medicine 44(6):586–588.
McCaig LF, Burt CW. 2005. National Hospital Ambulatory Medical Care Survey: 2003 Emergency Department Summary. Hyattsville, MD: National Center for Health Statistics.
McKay JI. 1999. The emergency department of the future: The challenge is in changing how we operate! Journal of Emergency Nursing 25(6):480–488.
McManus ML, Long MC, Cooper A, Mandell J, Berwick DM, Pagano M, Litvak E. 2003. Variability in surgical caseload and access to intensive care services. Anesthesiology 98(6):1491–1496.
McManus ML, Long MC, Cooper A, Litvak E. 2004. Queuing theory accurately models the need for critical care resources. Anesthesiology 100(5):1271–1276.
Medical Advisory Committee, Pennsylvania Emergency Health Services Council. 2004. Joint Position Statement: Guidelines for Hospital Ambulance-Diversion Policies. Mechanicsburg, PA: Pennsylvania Emergency Health Services Council.
Mikhail MG, Smith FA, Gray M, Britton C, Frederiksen SM. 1997. Cost-effectiveness of mandatory stress testing in chest pain center patients. Annals of Emergency Medicine 29(1):88–98.
Morrissey J. 2004. Going with the (patient) flow. JCAHO’s watered down’ ER patient-management standard relieves hospital executives, disappoints docs. Modern Healthcare 34(6):1, 6–7.
Munoz E, Regan DM, Margolis IB, Wise L. 1985. Surgonomics: The identifier concept: Hospital charges in general surgery and surgical specialties under prospective payment systems. Annals of Surgery 202(1):119–125.
Murray M, Berwick DM. 2003. Advanced access: Reducing waiting and delays in primary care. Journal of the American Medical Association 289(8):1035–1040.
NAE, IOM (National Academy of Engineering, Institute of Medicine). 2005. Building a Better Delivery System: A New Engineering/Health Care Partnership. Washington, DC: The National Academies Press.
National Association of Public Hospitals, AHRQ. 2005. Presentation at the meeting of the Getting LEAN: Health care’s challenge, Denver, CO.
Picker Institute. 2000. Eye on Patients. A Report by the Picker Institute for the American Hospital Association. Boston, MA: Picker Institute.
Reinus WR, Enyan A, Flanagan P, Pim B, Sallee DS, Segrist J. 2000. A proposed scheduling model to improve use of computed tomography facilities. Journal of Medical Systems 24(2):61–76.
Richardson LD, Asplin BR, Lowe RA. 2002. Emergency department crowding as a health policy issue: Past development, future directions. Annals of Emergency Medicine 40(4):388–393.
Ross MA, Compton S, Richardson D, Jones R, Nittis T, Wilson A. 2003. The use and effectiveness of an emergency department observation unit for elderly patients. Annals of Emergency Medicine 41(5):668–677.
Rydman RJ, Isola ML, Roberts RR, Zalenski RJ, McDermott MF, Murphy DG, McCarren MM, Kampe LM. 1998. Emergency department observation unit versus hospital inpatient care for a chronic asthmatic population: A randomized trial of health status outcome and cost. Medical Care 36(4):599–609.
Schafermeyer RW, Asplin BR. 2003. Hospital and emergency department crowding in the United States. Emergency Medicine (Fremantle) 15(1):22–27.
Schneider S, Zwemer F, Doniger A, Dick R, Czapranski T, Davis E. 2001. Rochester, New York: A decade of emergency department overcrowding. Academic Emergency Medicine 8(11):1044–1050.
Schull MJ, Lazier K, Vermeulen M, Mawhinney S, Morrison LJ. 2003. Emergency department contributors to ambulance diversion: A quantitative analysis. Annals of Emergency Medicine 41(4):467–476.
Siddharthan K, Jones WJ, Johnson JA. 1996. A priority queuing model to reduce waiting times in emergency care. International Journal of Health Care Quality Assurance 9(5):10–16.
Sinclair D, Green R. 1998. Emergency department observation unit: Can it be funded through reduced inpatient admission? Annals of Emergency Medicine 32(6):670–675.
Solberg LI, Asplin BR, Weinick RM, Magid DJ. 2003. Emergency department crowding: Consensus development of potential measures. Annals of Emergency Medicine 42(6):824–834.
Southard PA, Hedges JR, Hunter JG, Ungerleider RM. 2005. Impact of a transfer center on interhospital referrals and transfers to a tertiary care center. Academic Emergency Medicine 12(7):653–657.
Starfield B. 2000. Is U.S. health really the best in the world? Journal of the American Medical Association 284(4):483–485.
Talbot-Stern J, Richardson H, Tomlanovich MC, Obeid F, Nowak RM. 1986. Catheter aspiration for simple pneumothorax. Journal of Emergency Medicine 4(6):437–442.
Thorpe KE. 1990. The current hospital crisis in New York City and policy options for resolving it. New York State Journal of Medicine 90(5):247–252.
Vallee P, Sullivan M, Richardson H, Bivins B, Tomlanovich M. 1988. Sequential treatment of a simple pneumothorax. Annals of Emergency Medicine 17(9):936–942.
Viccellio P. 2001. Emergency department overcrowding: An action plan. Academic Emergency Medicine 8(2):185–187.
Wears RL, Perry SJ. 2002. Human factors and ergonomics in the emergency department. Annals of Emergency Medicine 40(2):206–212.
Weissert W, Chernew M, Hirth R. 2003. Titrating versus targeting home care services to frail elderly clients: An application of agency theory and cost–benefit analysis to home care policy. Journal of Aging & Health 15(1):99–123.
Wilson JW, Oyen LJ, Ou NN, McMahon MM, Thompson RL, Manahan JM, Graner KK, Lovely JK, Estes LL. 2005. Hospital rules-based system: The next generation of medical informatics for patient safety. American Journal of Health-System Pharmacy 62:499–504.
Wilson MJ, Nguyen K. 2004. Bursting at the Seams: Improving Patient Flow to Help America’s Emergency Departments. Washington, DC: The George Washington University Medical Center.
Zwiche DL, Donohue JF, Wagner EH. 1982. Use of the emergency department observation unit in the treatment of acute asthma. Annals of Emergency Medicine 11:77–83.