Three case studies on discharge innovation by Phyllis Maguire
Published in the July 2011 issue of Today's Hospitalist
Much has been made about the need to bolster transitions of care out of the hospital, and several big initiatives designed to improve discharge planning—including Project BOOST from the Society of Hospital Medicine and Project RED from the Boston University Medical Group—have gained national attention.
But at hospitals around the country, hospitalists are coming up with their own approaches to reinventing and improving discharges. The pressure on hospitalists has never been greater to free up beds as early as possible in the day—all while making sure patients who leave don't bounce back.
"There is no magic bullet. You need to try 10 different interventions to achieve the gain you're looking for."
–Victor Morris, MD Yale New Haven Hospital
hospitalists are finding that discharges involve a dizzying array of moving parts. "The analogy I like to use is an airplane trying to take off with all of its seats filled," says David Hoffmann, DO, who directs the hospitalist program at Chambersburg Hospital in Chambers-burg, Pa. "There are frequent breakdowns because there are so many different variables."
Here's a look at how three groups are trying to improve the discharge process, and the stumbling blocks that they're overcoming.
CASE 1: Chambersburg Hospital
At his hospital, Dr. Hoffmann wants to re-engineer discharges to not only reduce readmission rates, but to improve patient satisfaction. Many of the patient complaints he receives as the hospitalist medical director stem from fumbled discharges.
"In one out of every 30 discharges" Dr. Hoffmann explains, "things get missed. Patients were sent home who didn't know how to use their insulin or they got the wrong prescription."
When they dissected the discharge process, Dr. Hoffmann and his colleagues found that many problems could be traced back to two factors: missed communication among providers, and bungled coordination of all the variables in the discharge plan.
Communication gaffes would involve a specialist telling patients that they were ready to go home when hospitalists didn't think that was the case. Or a hospitalist would schedule a discharge when the patient's family was at work and couldn't pick him or her up.
Coordination problems could entail unanticipated insurance problems that cropped up the day of discharge, or patients not being educated about new medications or equipment. Those issues leave patients and family members waiting in their hospital room, stewing.
"The ball gets dropped because of missed opportunities or a bad pairing of expectations," Dr. Hoffmann notes. "The patient has one expectation, but the physician has another."
Moving to planned discharges
Dr. Hoffmann's group is trying several approaches to bridge those gaps. For several months, he and a few colleagues have been making half of the five or six discharges they each do every day a planned discharge.
"You have to pick patients who are fairly predictable," Dr. Hoffmann says. "You tell them one or two days before, 'I'm 85% certain that we'll discharge you on Wednesday at 11 a.m.' The nurses know that too, so everyone has a point to work from."
He's learned to space those planned discharges out over the day. "I'll put one on my schedule at 9, then one at 11 and maybe one at 1," Dr. Hoffmann explains. "You have to give yourself enough time in between to do the things that come up urgently."
He also does some paperwork for planned discharges a day or two ahead, and he makes sure to check in with specialists in advance.
"You have to tell them when you're planning to discharge a patient," says Dr. Hoffmann. For a new dialysis patient, for instance, "I'll call the nephrologist and say, 'I want to send this patient home on Wednesday after dialysis. Are you onboard with that?'"
He points out that he's done as many as four planned discharges in one day, but "the sweet spot is two or three. If you can effectively plan 50% of your discharges, you can improve patient satisfaction, but you definitely have to start out doing just one."
Electronic discharge checklist
Another innovation that Dr. Hoffmann's group is piloting takes even more direct aim at discharge communications.
He and his colleagues realized that in the hospital's electronic medical record system, doctors and nurses have to work to see each other's discharge-planning notes. Doctors must make seven or eight clicks of a mouse to get to the nurses' notes, while the nurses must quit their own module to see what doctors have entered in the chart.
That's why his group is about to try an electronic discharge checklist that will give both doctors and nurses a common area to list patients' discharge needs. The checklist will allow hospitalists to read, for instance, if nurses need them to write a nebulizer prescription, or that a particular patient's insurance won't pay for a wound vac, so a wet-to-dry dressing should be used.
Because the checklist will mean more documentation for hospitalists, the group plans to tie part of the physicians' quality bonus to how well they use the tool. And once the hospitalists and nurses work out any checklist kinks, Dr. Hoffmann plans to introduce it to other departments to get specialists using it as well.
CASE 2: Yale New Haven Hospital
Yale New Haven Hospital in New Haven, Conn., began scrutinizing discharges in 2008 because it was always operating at maximum capacity. "We're 966 beds, and we're always full," says Victor Morris, MD, medical director of the hospitalist program.
The hospital started by bringing in consultants to shadow everyone in the hospital for three weeks. The first big lesson learned, says Dr. Morris, was that "there is no magic bullet. You need to try 10 different interventions to achieve the gain you're looking for."
Dr. Morris says the hospitalists quickly realized that the work surrounding discharges needed to be moved up a day or two. To keep everyone, including nurses and care coordinators, abreast of where each patient is in terms of being discharged, the hospital uses a red/yellow/green color code system.
Green indicates that patients are likely to be discharged the next day, while yellow indicates that patients may be clinically ready for discharge, pending factors like lab results. Red means that patients aren't yet close to leaving.
Patients' electronic charts display their color, and "there's a big, flat-screen monitor sitting in the center of the nursing station in each unit" with a color next to every patient's name, says Dr. Morris. The color codes also appear on each hospitalist's daily list of assigned patients.
"I print out my list of patients in the morning," Dr. Morris says, "and I can see which are red, green or yellow." Nurses and care coordinators use the system to make sure all patient education is finished, transportation is in place and follow-up appointments have been scheduled the day before.
Discharge huddles, 11 a.m. discharge
But deciding when patients can be deemed yellow or green requires a great deal of planning. Every morning during sign-out, for example, physicians and nurses discuss where patients are in terms of being discharged.
And to ensure that imminent discharges are on track, each floor holds three discharge huddles a day. The first of these 10-minute meetings happens early in the morning, while the second is held in the late morning and the last happens at 3 or 4 p.m.
In these huddles, the charge nurse, the care coordinator, an intern or resident, and a hospitalist or hospitalist PA run through the list of patients. The goal is to keep each other up to date on whether an expected discharge is proceeding as planned or has hit a snag.
And when doctors see that a discharge is expected the next day, they know to complete the discharge summary the night before. The hospitalist group now tracks how many discharge summaries physicians complete the day before discharge—and have made that number part of the formula used for hospitalists' annual performance incentive plan.
The hospital has created another incentive for the hospitalists: increasing the number of patients discharged by 11 a.m. When the hospital first started looking at ways to improve discharges, a paltry 6% of patients were being discharged by that time.
Today, 21% of patients are discharged by 11 a.m., which the hospital wants to push to 24% over the next several years. Dr. Morris notes that one factor that's allowed the hospitalists to meet that 11 a.m. target is the group's substantial number of PAs and APRNs.
"When we get our patient list in the morning," Dr. Morris explains, "the doctors go see the sick patients, and we send the midlevels to see the green patients."
CASE 3: John Muir Medical Center
The hospitalists with John Muir Medical Group, who provide care in two hospitals in San Francisco's East Bay, use a different approach to designate which patients they expect to discharge soon. They rely on the notation "48," as in 48 hours to probable discharge.
At John Muir Medical Center in Walnut Creek, "We just click a button on our computer and that puts the number next to the patient's name," explains Deborah Arce, MD, the medical director of John Muir's hospitalist program. At the other hospital campus in Concord, Calif., the process is done on paper. "We send the case management office our census list with '48' written next to the patient's name," Dr. Arce says. "Once case management at either campus see the '48,' they know to mobilize."
According to Viviane Alfandary, MD, a hospitalist with the group, the hospitalists for many years have also been calling the patient's family the day before discharge to identify any problems, and they've reviewed medications with patients the day before discharge. In addition, the hospital has funded an extra RN case manager on weekends, Dr. Alfandary says, "so weekend discharges won't lag because of having only a skeleton crew."
Another strategy is having hospitalists complete their discharge orders before noon on the day of discharge. Patients may be discharged later in the day, Dr. Arce explains, but having orders completed by noon allows the hospital to know how many nurses to staff for the evening and night shifts.
"We started by setting the target for discharges by noon at 50%, and each year we raise the bar," she says. "The hospitalists now receive a tiered bonus if they hit 60%, 70% or 75%."
Funding pilot projects
According to Dr. Arce, the hospitalists are ramping up several other discharge innovations to make their jobs more manageable. Those innovations are the result of a strong collaboration between the John Muir hospitalists and the physicians with a competing hospitalist group, East Bay Hospitalists, which also serves both hospital campuses.
A year ago, Dr. Arce explains, both she and the medical director of East Bay Hospitalists, Peter Rowe, MD, decided to approach the hospitals' administration together. They came away with a far-reaching commitment on the part of the hospital system to fund an innovative 90-day rapid cycle improvement project. That funding, Dr. Arce says, has allowed both groups to collaborate on "pilot projects aimed at enhancing efficiency and hospitalist satisfaction, as well as improving discharges."
As part of that process, for instance, both hospitalist groups are now piloting the use of RN clinical care coordinators. These coordinators focus on communicating with families and the care team, providing patient education around discharge planning, scheduling follow-up appointments and calling patients after discharge.
A lower census
And two pilot projects experimented with ways to lower hospitalists' daily census. To reduce census, Dr. Arce's group embedded a hospitalist in the ED to do all daytime admissions, while Dr. Rowe's group assigned an extra hospitalist to the day shift.
The idea behind the lower census, Dr. Arce points out, was to give hospitalists the time to visit patients a second time every day after their rounds are over. "During the revisit, our hospitalists were encouraged to provide patient education, prepare discharges for the following day and communicate with the care team," she says.
How low did the census need to go to make those revisits possible? "If we cut physician census to 14 or fewer every day, we ask them to go back and revisit patients," says Dr. Arce. She adds that while the average hospitalist census for both groups until now has been 15 or 16, that figure can grow to 19 or higher in the winter.
During the six-week pilot project, participating hospitalists made over 200 revisits. "They also decreased the ED length of stay by 0.47 hours, reduced inpatient length of stay by four-tenths of a day and decreased our readmission rate by 2.5%," Dr. Arce says.
Those findings convinced administrators that having hospitalists make second visits every day is a key to better discharges. As a result, the hospital system has agreed to fund an additional hospitalist rounder for each campus— and for each of the two hospitalist groups.
Phyllis Maguire is Executive Editor of Today's Hospitalist.
What doctors communicate around discharge
WHILE THE DISCHARGE PROCESS is being scrutinized for its impact on throughput and readmissions, one group of researchers is using another yardstick to measure patient discharges: medical professionalism.
Jennifer Bracey, MD, a hospitalist with Johns Hopkins Bayview Medical Center in Baltimore, was the lead author of an abstract presented at this year's Society of Hospital Medicine meeting that looked at hospitalists' communication behaviors around discharge.
When Dr. Bracey and her colleagues asked nearly 30 hospitalists to report on what they communicate at discharge to patients and other providers, many admitted that they didn't routinely complete several important components. For example, 55.6% never or rarely used teach-back techniques when giving patients discharge instructions, while 41% never or rarely called a patient's primary care physician at discharge.
Most (93%) never or rarely called a patient after discharge. And only 52% usually or always spoke with a nurse about the patient's postdischarge care plan.
Surveyed hospitalists did do some things right. For instance, 89% usually or always talked to patients about discharge medications. And 74% of them usually or always completed discharge summaries within 24 hours of the patient being discharged.
According to Dr. Bracey, the hospitalists knew that each of the communication techniques was important, but they simply didn't have enough time to pursue them all.
The surveyed hospitalists did, however, feel that completing a discharge summary in a timely manner was one high-yield communication, as was making sure that patients had a follow-up appointment. (Fifty-six percent of the hospitalists usually or always personally ensured that patients had a scheduled follow-up appointment.)
Dr. Bracey and her group are considering what to do with the information they gathered. If most hospitalists can't find time to contact primary care physicians at discharge and patients after discharge, perhaps hospitalist schedules need to include some hours at the end of the week for such calls.
And Dr. Bracey and her colleagues are starting pilot projects to explore ways to more actively involve nurses in discharge planning. One project is changing the times that doctors hold multidisciplinary rounds.
"Many nurses are administering patients' medications at 10 a.m., which is when we do our rounding with case man- agers," Dr. Bracey says. "That makes it harder for nurses to participate."
Another project involves giving patients a piece of paper in their admission packet for them to jot down concerns about their own discharge, such as not being able to afford medications or find rides to follow-up appointments. Patients will be encouraged to share those with their physicians or nurses.
"That's another way," Dr. Bracey says, "to get both patients and nurses more involved in discharge care."