Quality Matters offers reports on emerging models and trends in health care quality improvement and interviews with leaders in the field.
Summary: Smoothing the flow of patients in and out of hospitals and other health care settings can help to reduce overcrowding, prevent poor handoffs, and avoid delays, all of which may worsen as more people gain access to insurance coverage and care. A number of hospitals and health systems are pursuing strategies to improve patient flow such as orchestrating the arrival and discharge of patients undergoing elective procedures and transferring the oversight of patients waiting to be admitted from emergency departments to other hospital units.
By Martha Hostetter and Sarah Klein
As far back as 2003, the Institute for Healthcare Improvement (IHI) and some other quality improvement groups were making the case that improving patient flow—the process by which patients move through a hospital or other health care setting—could improve the safety and quality of care.1 Smoothing the flow of patients in and out institutions, they argued, could help to reduce wide fluctuations in occupancy rates and prevent surges in patient visits that lead to overcrowding, poor handoffs, and delays in care.
Since then, evidence has continued to mount demonstrating that one of the key manifestations of poorly managed flow—overcrowding in emergency departments, intensive care units, and other hospital floors—leads to adverse outcomes, most likely because support staff in services such as laboratory and radiology are unable to keep up and physicians and nurses have less time to focus on individual patients.2 One study found that for each additional patient with heart failure, pneumonia, or myocardial infarction assigned to a nurse, the odds of readmission increased between 6 percent and 9 percent.3 Another found that when patients are discharged from an ICU because of overcrowding, they are at much greater risk of being readmitted to the ICU.4 Mortality rates have also been shown to increase when the ratio of surgical nurses per patient decreases, while patients held in the emergency department (ED) until inpatient beds are freed up have higher rates of morbidity and mortality.5
Despite this evidence, overcrowding in hospitals and other health care settings remains widespread and persistent because, experts in operations management say, health care administrators have tended to treat the symptoms rather than the causes of the problem. They have built new facilities and hired additional staff—approaches that may no longer be viable for institutions facing capital constraints and workforce shortages. More effective and lasting solutions, they say, involve orchestrating the arrival and discharge of patients undergoing elective procedures, the timing of which is often under institutional control; making admission and discharge processes more efficient; and moving oversight of patients waiting to be admitted from emergency departments to other hospital units.
Putting an End to Long Stays in the ED
Many existing efforts to improve patient flow focus on emergency departments, which are experiencing record volumes as more and more physicians direct their patients to the ED rather than admitting them themselves. For this and other reasons, EDs are now responsible for about half of all hospital admissions and nearly all of the rise in overall admissions from 2000 to 2009.6 The increase in volume tends to increase wait times in the ED, which leads some sick or injured patients who grow tired of waiting to leave without being seen. It also encourages "boarding," the practice by which patients in need of admission wait in the emergency department for a bed to open up on an inpatient unit.
More than a decade ago, Peter Viccellio, M.D., an emergency medicine physician at Stony Brook University Hospital, in Long Island, N.Y., became concerned about the impact boarding had on patients. On a day when he had eight isolation patients in ED hallways waiting for beds, he queried colleagues in the New York State Health Department whether it was against the rules to move ED patients to the hallways of inpatient units even if there weren't beds ready for them. He found out there were no health or fire codes prohibiting it: hospital codes simply didn't address how to deal with circumstances in which the number of admitted patients exceeded the number of certified hospital beds. With approval from the Department of Health, his institution developed the full capacity protocol, which allows for redistribution of admitted patients to inpatient units during times of ED crowding, even if hospital beds aren't available. Stony Brook adopted the protocol in 2001. It has since spread to several but still a minority of hospitals across the country, even though research has found that it is safer to put patients in inpatient hallways than to board them in the ED, which is not designed to provide acute care over time to seriously ill patients.7
Implementing the full capacity protocol requires institutions to acknowledge that overcrowding is a hospital-wide problem that requires a hospital-wide response. Still, Viccellio describes the full capacity protocol as a stopgap measure—"moving a patient from one hallway to another hallway"—rather than a solution to the problem of overcrowding.
Another way hospitals attempt to relieve congestion is by discharging patients early in the day when appropriate. New York University Langone Medical Center, in Manhattan, had a long-standing policy of encouraging morning discharges, both to improve flow and enable more continuous care at home. Leaving the hospital early in the day gives "patients time to get home, get their prescriptions filled, and have a visiting nurse come in," says Brenda Ohta, Ph.D., senior director of care management. Still, the policy was rarely achieved in practice.
To change this, in 2012, hospital leaders established daily meetings to identify patients who might be ready to leave the next morning and sent email reminders to all members of the patients' care team assigning tasks—such as getting prescriptions ready or informing family members—to particular individuals. They also educated staff about the importance of early discharges and created an electronic dashboard to track daily progress on this goal. As a result, today more than 40 percent of the hospital's patients are discharged in the morning—compared with 15 percent or less before this initiative. Overall length of stay and readmission rates have also dropped, an indication that patients are not being inappropriately discharged and, perhaps, a sign that the focus on early discharge has created a more organized process overall.
Smoothing Patient Flow in Operating Suites
Hospitals also struggle with overcapacity because of the widespread practice of scheduling elective surgeries on Mondays and Tuesdays, rather than spreading them evenly throughout the week. Recognizing this, Monmouth Medical Center, a 527-bed academic medical center in Long Branch, N.J., began focusing on improving patient flow in its operating rooms as part of the Partnership for Patients initiative, which provides technical and financial support to hospitals working in hospital engagement networks (HENs) to reduce readmissions, infections, falls, and other adverse events.
Monmouth is one of 16 hospitals participating in a HEN led by the New Jersey Hospital Association. As part of this work, hospitals were able to apply to partner with the Newton, Mass.–based Institute for Healthcare Optimization (IHO), a nonprofit organization that consults with hospitals to separate elective and nonelective flows as well as inpatient and outpatient flows; smooth patient census for elective surgery; and estimate and "right-size" resource needs for each type of patient care. All of these steps aim to reduce what IHO terms "artificial variability" in patient census created not by urgent or emergent cases but by electively scheduled procedures and admissions (see Q&A).8
Monmouth staff reviewed five months of data, which showed the number of patients receiving elective surgery each day varied widely, leading to widely varying demand on staff and resources. The main culprit, it turned out, was the practice of scheduling a large number of orthopedic surgeries and three types of general surgery (vascular, ENT, and thoracic) early in the week—leading to overcrowding in the postoperative units. Like many surgeons, those operating at Monmouth preferred operating early in the week to avoid having their patients recover over the weekends when there are often fewer resources available to them.
To address this issue, Monmouth administrators and surgeons performed a "gap analysis" to see what additional support might be needed—such as physical therapy, case management, and radiology services—to provide the same level of postoperative care over weekends as is provided during weekdays. They then created a business plan to calculate how many additional surgeries per week would be needed to offset the costs of expanding staff over the weekend.
Working with the IHO team, they developed a new weekly scheduling protocol that sought to smooth their daily patient census over five days by spacing surgeries throughout the week. Asking surgeons to move their block time—a regularly scheduled period each week when mostly senior surgeons were given reserved time and resources—required "a real paradigm shift," according to Patricia A. McNamee, R.N., M.S., the hospital's assistant vice president of perioperative services. For Monmouth, like most hospitals, surgeries are a major source of revenue. And because nearly all of the surgeons operating at Monmouth are not employees—and thus have choices about where to bring their cases—hospital leaders didn't want to risk driving them away by denying them their preferred surgical slots.
To avoid this, schedulers proceeded by using carrots, rather than sticks, to encourage surgeons to change times, for example by offering those willing to move their block time from a Tuesday to Friday the opportunity to have support from a nurse surgical assistant and two operating rooms available to them so they could "flip" from one case to another and increase their productivity.
The new schedule has only been in place for two months, but on most weeks Monmouth has been hitting its target of smoothing patient census over five days. The increased revenue from being able to fit in more surgeries has more than offset the additional costs of the extra resources that were added to weekends. And better flow appears to be improving care, as measured by patients' ratings of nurse communication: in August and September, 87 percent of Monmouth patients surveyed said their nurses communicated well, compared with 83 percent in June and July.
"Under the old admission practices, with wild variation in daily census, nurses had trouble keeping up with the workload on busy days, nurses were floated in from different floors to help, and one nurse was pulled off of direct patient care to do all admissions," says McNamee. "Smoothing has led to less fragmented care: each patient is assigned a nurse who does an admission assessment and develops a care plan. This seems to have improved communication and continuity of care."
St. Louis Children's Hospital provides another example of how improving patient flow—in this case, during the admission process—can also improve continuity of care. In February 2010, the hospital established a call center to serve as a single point of entry for referring physicians seeking to admit a patient or confer with a specialist affiliated with Washington University School of Medicine. Pediatric nurses with critical care experience staff the center 24 hours a day, seven days a week, and can arrange an immediate consult with a specialist, schedule an appointment, or facilitate admission by assigning responsibility for the case to a hospital physician (who then takes part in a conference call with the nurse and referring physician). In the background, nursing staff begin the registration process and identify a bed for the patient.
Using the Children's Direct Transfer Center, the hospital has been able to admit some 4,000 patients a year directly to hospital units. Because the center's incoming calls are recorded, the hospital has been able to review cases to determine whether the routing was appropriate and handoffs were safely and efficiently handled. An oversight committee looks for cases in which patients were transferred to the PICU within 12 hours of admission as a measure of possible failure. "What we found is that our rate of transfer [to the PICU] from the emergency department, from doctors' offices, from our own clinics, or from an outside hospital is all about the same—it is just less than 1 percent," says Douglas Carlson, M.D., medical director of the transfer center.
Improving the admission process has also led to increased revenue from growing numbers of transfers and referrals from community physicians who appreciate the improved handoff process, says Julie Bruns, director of the call center. "We really try to ensure [physicians] only tell the story once," she says.
Improving Continuity in Primary Care
Scheduling is one key to smoothing flow in hospitals, but it is also crucial in primary care clinics, which experience high demand from patients with highly variable needs. One of the clinics associated with the Cambridge, Mass.–based Cambridge Health Alliance, Malden Family Medicine Center, serves 11,000 patients with 40 different primary care physicians, who are a mix of faculty and residents associated with Tufts University School of Medicine. Trying to schedule appointments that ensure continuity with so many different doctors is a challenge, according to Greg Sawin, M.D., a family practice physician at Malden. "It involves layering multiple calendars," he says, "and after a few layers down the human brain begins to falter."
Malden received help from the Healthcare Systems Engineering Extension Center, led by James Benneyan, Ph.D., an industrial engineer at Northeastern University who received a Health Care Innovation Award grant from the Center for Medicare and Medicaid Innovation to bring technical expertise in systems engineering and operations management to health care. Benneyan hopes "to demonstrate how engineers can meaningfully partner with health systems by applying systems improvement methods used in other complex industries in ways that will impact the triple aim of better health, better care, and lower costs."
Under the extension center model, a team of senior engineers, graduate students, and undergraduate students partners with providers on a particular project, sometimes "embedding" themselves in hospitals or clinics to learn about the real-world issues affecting their operations. Northeastern engineers worked with Malden to design a scheduling model that maximizes care continuity by ensuring that as often as possible patients are seen by a member of their designated care team—a process that requires taking into account residents' rotations, faculty members' teaching schedules, vacation and sick time, and many other complicating factors.9 If a patient's primary care physician is not available, he or she is seen by the next-most "optimal" provider on their care team (a medical assistant, nurse practitioner, or another appropriate provider). The theory, consistent with the clinic's efforts to provide a patient-centered medical home, is that having patients seen by the same group of providers over time will improve their compliance with recommended care and provide the foundation needed to deliver higher-quality care.
"It's all about continuity," says Sawin. "If you have to see complex patients during a 20-minute visit, and they have a problem list as long as your arm, it helps for providers to have increased familiarity." It also engages patients if they see the same front-desk staff, nurses, and medical assistants at each visit, he says. "If Charlene is the medical assistant who rooms my complex patient over and over again, then when she calls her to schedule a mammogram [the patient] is more likely to listen. It builds trust, which is crucial to engaging patients."
Spreading Proven Models
Efforts to make more efficient use of existing resources by smoothing patient census and improving admission, handoffs, and discharge processes can have significant benefits not only in terms of improved safety and more continuous care, but also in terms of increased efficiency and revenue.10 This makes it hard to understand why proven approaches to improving flow haven't been more widely adopted.
One reason may be that health care is often regarded as an exceptional industry—fundamentally different than complex systems such as airlines or hotels that have benefited from operations management approaches—and therefore subject to different rules. There is undoubtedly some truth to this, but there is also much about a complex system such as a hospital that can be understood and even predicted, experts say. For example, modeling hospital admissions—even emergency admissions, which are not as unpredictable as one might think—can be done with an 80 percent to 85 percent confidence level, according to Kirk Jensen, M.D., chief medical officer of Best Practices, Inc., and an IHI faculty member. "We know who is coming tomorrow, we know when they are coming, and we know what they will need," says Jensen. "We just don't know their names."
With the federal government backing efforts to spread operations management in health care through grants to the Institute for Healthcare Optimization and the Health Care Systems Engineering Extension Center, these initiatives may be gaining traction. Providers may also be encouraged to improve patient flow by the awareness that, with the insurance expansions under health reform, they may experience even greater demand.11
In the end, improving patient flow in hospitals and other health care settings may mean doing more to accommodate patients' schedules (rather than physicians') and rethinking how and when care should be delivered. In primary care, this is already happening with the growing number of retail clinics and those embedded in schools or other convenient locations. Hospitals, too, are slowly changing their way of doing business to accommodate "after hours" care.
"If you look back 30 or 40 years when the average length of stay was 10 days and average acuity level was very moderate, it made sense to run the place 9 to 5 Monday through Friday with a skeleton crew on evenings and weekends," says Stony Brook's Viccellio. "But today 60 to 80 percent of patients are admitted as unscheduled emergencies over seven days of the week—and yet we're still trying to fit that problem into a five-day-a-week solution."
Some have suggested that improving flow in hospitals may in fact drive up health care costs. For example, improving flowing in the OR may open the door for hospitals to deliver more elective surgeries, some of which may be medically unnecessary. But advocates say that capacity is not the only constraint on unnecessary surgeries—and that the benefits of improved flow outweigh such concerns.
"There are a number of drivers that mitigate this concern—innate professionalism, peer review, the threat of malpractice, development of professional networks, and the further influence of data on hospital and physician practices," says Jensen. What's more, new payment models (such as shared risk arrangements in ACOs) should encourage providers to ensure patients receive care in the most appropriate settings and take steps to hold down admissions and ED use.
"Improving flow has the capacity to improve access for those patients who don't have it," he says. "It can improve safety and reliability—leading to fewer unnecessary process or procedural steps, better professional caregiver-to-patient ratios, and fewer delays."
1 "Optimizing Patient Flow: Moving Patients Smoothly Through Acute Care Settings," IHI Innovation Series white paper (Cambridge: Institute for Healthcare Improvement, 2003).
2 A. T. Pedroja, "The Tipping Point: The Relationship Between Volume and Patient Harm," American Journal of Medical Quality, Sept.–Oct. 2008 23(5):336–41.
3 Peaks in patient numbers lead to increased numbers of patients being assigned to nurses. M. D. McHugh and C. Ma, "Hospital Nursing and 30-Day Readmissions Among Medicare Patients with Heart Failure, Acute Myocardial Infarction, and Pneumonia," Medical Care, Jan. 2013 51(1):52–59.
4 D. R. Baker, P. J. Pronovost, L. L. Morlock et al., "Patient Flow Variability and Unplanned Readmissions to an Intensive Care Unit," Critical Care Medicine, Nov. 2009 37(11):2882–87.
5 L. Aiken, S. P. Clarke, D. M. Sloane et al., "Hospital Nurse Staffing and Patient Mortality, Nurse Burnout, and Job Dissatisfaction," Journal of the American Medical Association, Oct. 23/30, 2002 288(16):1987–93; A. J. Singer, H. C. Thode, Jr., P. Viccellio et al., "The Association Between Length of Emergency Department Boarding and Mortality," Academic Emergency Medicine, Dec. 2011 18(12):1324–29.
6 K. Gonzalez Morganti, S. Bauhoff, J. C. Blanchard et al., The Evolving Role of Emergency Departments in the United States, RAND Research Report, 2013.
7 A. W. Viccellio, C. Santora, and A. J. Singer, "The Association Between Transfer of Emergency Department Boarders to Inpatient Hallways and Mortality: A Four-Year Experience," Annals of Emergency Medicine, Oct. 2009 54(4):487–91; and Singer, Thode, Viccellio et al., "The Association Between Length of Emergency Department Boarding and Mortality," 2011.
8 Other hospitals working with the IHO in the New Jersey group have reduced inappropriate use of telemetry beds, improved transfer and discharge processes, and reduced length of stay and ED boarding.
9 Other efforts undertaken by the Healthcare Systems Engineering Institute focus on patient safety, appointment access, shared savings models, home health, ED observation units, readmissions, and other issues.
10 S. Allen, "No Waiting," Boston Globe, August 30, 2009. E. Litvak and H. V. Fineberg, "Smoothing the Way to High Quality, Safety, and Economy," New England Journal of Medicine, Oct. 24, 2013 369:1581–83.
11 E. Litvak and M. Bisognano, "More Patients, Less Payment: Increasing Hospital Efficiency in the Aftermath of Health Reform," Health Affairs, Jan. 2011 30(1):76–80.