Prognosticators often look to Australia for an idea of what kind of flu season we can expect in the United States. Australia just finished its flu season, which started early and was unusually severe. Whether that accurately portends a similarly severe season in the United States remains to be seen.
Still, hospitals and EDs should be prepared to manage the kind of capacity problems that result when hospitals accustomed to running at or near full capacity are slammed with a steady flow of influenza patients, many of whom require hospitalization.
In fact, there are several tactics that can help hospitals and EDs manage the kind of capacity challenges that occur during flu seasons and other periods of peak volume. However, such interventions require some leadership and planning to carry them out successfully. Many of these tactics were highlighted in a presentation conducted by the Institute for Healthcare Optimization (IHO), a nonprofit focused on applying science and operations practice to healthcare delivery, on Oct. 23, 2019. Aptly titled “Optimizing Patient Flow in Preparation for the Flu Season,” presenters broke down specifically where and why the most common types of bottlenecks occur, and how hospitals and EDs can select and implement the best solutions for their needs.
Consider Artificial Peaks
Eugene Litvak, PhD, president and CEO of IHO, noted that a common misconception about the variability that occurs regarding hospital patient volume is that most of it is due to admissions from the ED. However, he stressed that is not the case.
Instead, Litvak noted that in many hospitals, investigators find that it is elective or scheduled admissions that are at the root of most variability in a hospital’s census, a phenomenon Litvak calls artificial variability.
“It is not random or nonpredictable. It is manmade,” he said. These elective admissions do not just involve patients who are scheduled for surgery. They can involve telemetry beds or medical admissions, too. Still, these elective, scheduled admissions tend to create what Litvak called artificial peaks.
Further, he noted that when this type of variability is not well-managed, and then a flu epidemic breaks out, there can be many consequences, such as overcrowding, medical errors, infections, and even a higher risk of mortality.
What else causes bottlenecks? Sandeep Green Vaswani, MBA, senior vice president at IHO, noted that the criteria hospitals use for determining when to move patients from one care level to another often are fuzzy, making it difficult for clinicians to make objective decisions.
“There is just too much room for error around the specific criteria, and that is something that needs to be addressed,” he said. “[Also,] the criteria often tend to take into account legacy processes or specific physician preferences or agreements that may have been in place for a variety of reasons.”
Instead, patient-centric, clinical, and technical criteria should be used to determine how patients should be moved through a hospital system, Vaswani suggested. For instance, he observed that hospitals often maintain long lists of criteria for admission to different units and various levels of care. However, many of these criteria are so outdated that clinical staff largely disregard them.
“Nobody really uses them to methodically determine where a patient should be admitted and when they should move through different levels of care,” Vaswani explained. This adds another layer of variability to patient throughput.
Even after hospitals and EDs have updated their criteria and taken other steps to address bottlenecks, it is not uncommon for clinicians to still run into obstacles when trying to put patients in a particular unit they deem best for the patient’s needs.
“There may still be some misplacement even after you have streamlined your admissions, discharge, and transfer criteria,” Vaswani observed. “If that is the case ... figure out exactly how many beds are needed for each level of care for each type of service.”
To make such determinations, Vaswani recommended hospitals conduct queuing analyses and simulation techniques to determine what changes are needed to meet patient needs.
“What we find time and again is that once variability is removed, these tools — queuing and simulation — can help hospitals determine the right-sizing of these patient units,” he said. “What this kind of analysis allows you to do is be much more methodical about managing ED boarding rather than being surprised every day when you see there are not enough beds and there are too many patients waiting in the ED.”
However, Vaswani stressed this is not the kind of work that one can borrow from another hospital. “You need to figure out exactly what your problems are and where your bottlenecks are. Then, accordingly, determine what the appropriate intervention would be,” he said. “For instance, you could find that you have more admissions when certain physicians are in the ED or rounding on your unit. That may be an indication there is some artificial or manmade variability or patterns at your hospital on certain units.”
For cases in which hospital leaders suspect there may be unnecessary or inappropriate admissions, that is another chance for studying variability closer, Vaswani noted. For example, one may find there are long delays in the ED for admission to certain units while there are no delays for admission to other, similar units.
Alternatively, one could find discharge or transfer delays that are due to nonclinical reasons. There could be process delays or holdups in receiving appropriate orders, Vaswani shared. “If you find you are transferring patients too often from one unit to another to accommodate new admissions, that is another indication that you may not be right-sized, and there may be artificial reasons affecting patient flow,” Vaswani added. “Similarly, length of stay [LOS], depending on the team or provider of record, is an indication of artificial variability.”
Another red flag is when there is a different LOS for patients who are similar. “That may be an indication that something is going on,” Vaswani said. “For instance, access to services over the weekend may not be adequate, which may prolong LOS.”
When patients are treated in the right place at the right time, there often is an ability to decrease the number of beds required while also accelerating throughput, Vaswani said. “That also then helps to smooth out the workload for nursing, which obviously has a significant staff satisfaction aspect to it,” he observed. “Patients are certainly happier to be treated in the right place, and you see satisfaction scores go up as a result of improved flow.”
Look Beyond ED
When the waiting room in the ED is packed, and there are long waits for care, people tend to become frustrated, including administrators and staff.
“Everybody is sort of at each other’s throats about where the problem is,” observed Peter Viccellio, MD, FACEP, clinical director of emergency medicine at Stony Brook University. “There has been a longstanding culture in the hospital industry that since this problem sits in the ED, the ED should do something about it.” Perhaps there are too many unnecessary visits. Or, if there is a boarding problem, then the ED is admitting too many people unnecessarily. Further, if it takes too long to secure a CT scan, then the solution is to not order so many of them, Viccellio noted.
“It is true that if we closed our doors, we probably wouldn’t have this [capacity] problem,” he observed.
Viccellio explained that EDs tend to focus on things they can control. For instance, leaders might place physicians or other providers in triage so that blood work and other tests can begin earlier.
In fact, that was a step Viccellio’s own department took, which reduced LOS by an average of 16 minutes. Still, that only represented about 1.1% of the ED’s problem in terms of boarding and crowding.
“There is no right way to do the wrong thing,” Viccellio said. “If you want to fix the problem that you have, you have to attack that problem, not another one.” Further, the consequences of failing to fix a capacity problem in the ED puts patients at risk, increases LOS, hikes malpractice claims, and drives up nursing turnover, Viccellio added.
How can hospitals effectively address ED overcrowding or boarding? Viccellio said there are several known solutions, but it is important to pick the right solution and to implement that solution to a sufficient degree. For instance, he noted most hospitals are familiar with the concept of facilitating earlier discharges so that patients awaiting admission in the ED can move upstairs faster. However, if hospitals are increasing the number of their early discharges from 5% to 6%, the intervention probably has not been implemented to a sufficient degree to solve the problem.
Hospitals have to commit to a process in which stakeholders identify obstacles that must be addressed, and then move forward. It sounds simple enough, but some may be so caught up in the process they lose sight of the desired results, Viccellio explained. “They’re still looking at data, and still doing little projects, but have not really moved the numbers,” he said.
To prevent the process itself from becoming the goal, leaders must establish desired results at the outset, Viccellio explained. In this results-oriented approach, process leaders can say they want discharges by noon to increase by 30% in four months. This is in contrast to trying a solution, and then reporting back in a year.
“This does require leadership in order for [changes] to happen,” Viccellio stressed. “Once the change happens, though, then it is safer for the patient, it is easier for the physician and for the staff, and the financial benefits to the hospital can be huge.”
Viccellio noted research suggests that boarding in the ED is associated with one extra day in the hospital as an inpatient. A concerted focus on early discharges can enable one to affect LOS positively while also reducing boarding hours dramatically.
Another potential intervention involves taking steps to smooth elective admissions across all the days of the week rather than frontloading these admissions early in the week. For example, perhaps boarding problems at one facility are occurring most commonly on Monday and Tuesday, the same days when elective admissions were most prominent.
“At this same institution, with 600 beds and an average of 30 boarders, had these [elective admissions] been smoothed across five days of the week, there would be no boarding. Instead of needing 600 beds, it would need just 570 beds,” he explained. “If someone wanted to make the leap and [smooth elective admissions across] seven days of the week to address the [boarding] problem, the 600-bed hospital would only need 500 beds. That is assuming no change in LOS.”
Considering most hospitals must run at or near full capacity from a financial standpoint, it is important to put a plan in place for when the hospital is full, but patients are still coming to the ED, Viccellio noted. This is certainly the reality for many hospitals during periods of peak flu volume.
In these cases, Viccellio suggested rather than boarding patients in the ED while they await an inpatient bed, place them on the appropriate inpatient floors. In each case, there may be no bed available, but the right expertise is there to care for the patient, even if that patient is placed in a hallway temporarily. “If you believe in safety, you have to compare the patient’s experience [boarding in the ED] vs. if they are upstairs in a hallway,” Viccellio offered.
He added there should be a curtain around the patient, a call bell the patient can use to signal the nurse’s attention, and a bathroom the patient can use, but the patient winds up where he or she needs to be. “We found that about 25% of [such patients] get into a room right away, another 25% get into a room within an hour, and about 50% wait eight to 10 hours before they get into a room,” Viccellio reports.
Rather than boarding dozens of patients in the ED, one is asking other units to take over the care of one to three patients to redistribute the load. “If you are adding a couple of patients per unit, you are not having to add staff ... but the staff have to understand why they are [taking added patients] on,” Viccellio observed. “They are not doing it to make the emergency nurse’s job easier. They are doing it to make the patient’s care better.”
To implement this type of policy in his own hospital, administrators held meetings about the plan adjacent to the ED so that staff from upper floors could see what patients experience when they are boarded in the ED. It was clear that one emergency nurse cannot optimally tend to 10 or 12 patients requiring inpatient hospitalization. “This group of nurses at my institution was actually able to frame things as what is best from the patient’s point of view,” Viccellio reported.
If there are patients ready for discharge on upper floors, consider moving them into hallways so newly admitted patients can move into a room, Viccellio suggested. “If you do that, it may also help to facilitate early discharge,” he offered.
Investigators have found that this full capacity protocol is safer for patients, producing a lower mortality rate and a shorter LOS; the patients prefer this protocol, too, Viccellio noted. “We surveyed patients who were in the hallways in the ED vs. in the hallways upstairs. Close to 90% preferred being upstairs,” he said.
Again, hospitals and EDs must focus on solutions that work rather than pet projects or Lean events that may just nibble around the edges, leaving EDs with the same problem they are trying to solve. “The front end is not the problem. The throughput on the back end is. This should draw your attention to what your fixes are,” Viccellio stressed.
While hospitals and EDs grapple with their capacity problems, Viccellio recommended administrators consider every queue that exists. “You have to get patients out as quickly as you can, so what are they waiting for?” he asked. “Are they waiting for physical therapy, an echocardiogram, or a CT scan? Do all of these services go home at [3 p.m.] so that those patients have to wait until the next morning?”
Answering these questions should help hospitals pinpoint throughput problems. Then, administrators can find ways to redistribute staff so that services ordered today are fulfilled the same day. “This may help with throughput short term in a dramatic way,” Viccellio predicted.
Although gathering good data is important to any change process, Viccellio cautioned that continuous requests for new data should never be an excuse for inaction. “You can keep asking for data to enforce inaction for months to years,” he said. “These things have to be done in a framework of clear leadership and a time frame. You need your crucial data, but it is possible to keep asking for more and more data just to put off action.”