JUST BECAUSE YOU CAN doesn’t necessarily mean you should. That’s the object lesson that Michael Maniaci, MD, enterprise physician lead for Mayo Clinic’s hospital at home program, takes away from the less-than-stellar implementation of an electronic alert that he and his team devised.
Their goal with that alert was to try to engage frontline clinicians in identifying good home hospital candidates. But both the tool and its rollout landed with a thud, as Dr. Maniaci and his colleagues describe in a study published in January by the Journal of Hospital Medicine.
The best-practice advisory (BPA) alert was introduced in two Mayo sites in March 2021 and then studied for nine months. While close to 9,000 of the alerts were fired over that time on more than 2,800 patients who were potential candidates, providers referred only 324 to the hospital at home team.
“You really have to link your vision to strategy and execution.”
~ Michael Maniaci, MD
Mayo Clinic
Moreover, the team decided that only 31 of the referred patients met clinical and social criteria, a success rate of less than 10%. All the while, the hospital at home program in those two sites was admitting between 80 and 100 patients every month—but those patients weren’t being identified via the BPA alert.
Here’s what Dr. Maniaci says he learned: It’s very easy to launch an alert that relies on technology and data that are readily available through an EHR. But what allows such a tool to succeed are all the things that Dr. Maniaci and his team didn’t put together at the time, but have since spent time crafting. Those include home hospital protocols, training, communication processes, scripting, and feedback chains for when something goes wrong or clinicians have questions.
“You really have to link your vision,” he says, years after the first version of the alert proved spectacularly ineffective, “to strategy and execution.”
What went wrong
As Dr. Maniaci explains, hospital at home is fundamentally different from every other care setting.
“In every other patient situation, it’s a ‘push,’ not a ‘pull,’ ” he says. “Patients are pushed out of the ICU or out of the ED to the floor. On any other service, I never have to ask to get a new patient.”
But hospital at home is the exact opposite. Dr. Maniaci and his team have to “pull” appropriate patients from those who are already in the hospital or waiting to be admitted. That screening process takes time and labor.
The BPA project tried to speed up that process by giving clinicians in the ED or on the floors a heads-up about a patient’s potential eligibility for hospital at home. Hospital-based teams, the thinking went, would then engage in considering patients’ clinical eligibility for the program and refer them.
But that’s not what happened, and the JHM write-up is great reading for what can go wrong with research projects. One big problem: The alert—which, in its first iteration, was fired based on only a few patient data points including zip code, payer, diagnosis and some hospital at home exclusion criteria—”was so very basic to where almost anybody would qualify for hospital at home,” Dr. Maniaci says.
When receiving that alert, a provider could either refer the patient to the home hospital team, dismiss the notification or defer the decision.
“Of the alerts we sent during the study period, people just cancelled them because they thought they were annoying,” he says. “Alert fatigue is real, and this alert just made that worse.”
The need for tailored training
Another problem: The alert was fired before clinicians even had a chance to examine or get to know the patient.
“In the ED or on the floors, this alert was the first thing that came up when providers opened the chart,” says Dr. Maniaci. “Before you know one thing about the patient, you’re being told, ‘This patient qualifies for home hospital, do you want to refer?’ The vast majority of those who received it were like, ‘I don’t want to deal with this now.’ ”
Another thing his team initially got wrong: “You need to tailor your rollout and training to who’s actually doing the referring on the ground in that particular hospital,” he points out. Before the initial rollout, he and his team met once with hospitalists, and they also mentioned the alert a few times in department newsletters and meetings. Those very brief introductions proved to not be enough.
Study results were pooled from both the Mayo campus in Jacksonville, Fla., where Dr. Maniaci is based, and one in Wisconsin.
The Jacksonville facility is a teaching hospital and in the JHM writeup, almost 42% of providers making referrals to the home hospital program in that site were trainees. (An additional 36% were attendings, while 23% were NPs, PAs or RNs.) In Wisconsin, by contrast, 84% of the home hospital referrals in the study came from attendings, while trainees made only 6%. NPs, PAs and RNs made the remaining 10%.
“Training in Wisconsin has to focus on providers—and providers are more stubborn than anyone else,” says Dr. Maniaci. “We’re very possessive of our patients.”
With attendings and with APPs who have years of experience, he adds, “you have to tell them why hospital at home is good for their patients.” But when targeting APPs with less experience and medical learners, “you really have to stress what the criteria are for including patients in home hospital and who’s the right patient.”
Fortunately, he adds, home hospital is now getting widespread enough around the country with enough literature that the benefits of the care model are more well-known. Too, capacity constraints are still such a problem that doctors and nurses are themselves interested in freeing up beds when they can.
A better way
So what’s changed since the BPA alert first proved to be a dud? The alert is still being used, although its inclusion criteria have been updated. That means the alert is firing more consistently on patients who are more appropriate for hospital at home.
More importantly, the alert has been linked to an algorithm that weighs 50 different patient criteria and produces a “fit” score for that patient for hospital at home, one that produces a stoplight icon with the alert that’s either green, yellow or red.
“The fit score at least gives providers a little more information to decide whether or not to refer the patient,” Dr. Maniaci says.
And in addition to beefing up and targeting provider training, “we really paid attention to the communication plan,” he adds. Dr. Maniaci and his team worked with patient experience advocates and staff leadership to let them know the “why” behind home hospital. They also came up with scripts that providers use when approaching potential patients and worked with nurse managers, nurse administrators and individual specialty departments.
They also figured out a way to get feedback from and give help to clinicians if they weren’t sure how the program worked or if they had questions.
“All these things had to be built and rolled out, and that took us years,” Dr. Maniaci says. “Now, we have all these communication trees and protocols, but we’d done only 1% of that when we first rolled this alert out. At the time, we just said, ‘Here’s a tool, and here’s how to use it.’ That’s all we did.”
All high-acuity care
So has the home hospital team started firing the BPA alerts at a different time in the care process? They haven’t yet, Dr. Maniaci replies, but they are studying when the right time to receive the alert might be for different patients and providers.
“We’re trying to decide if there should be a delay—and if so, what that delay should be,” he says. “Does the patient first need a scope or a radiological procedure or a consult before home hospital is considered? That will ultimately need to be what we move to.”
As for screening patients for hospital at home, Mayo now maintains a dedicated APP admitter. But other team members who need to be engaged with home hospital patients—pharmacy, case management, social work—are the same staff treating and placing hospitalized patients.
“We’re trying to get away from splitting them into two silos: hospital at home and brick-and-mortar care,” says Dr. Maniaci. “Instead, we want to look at both types of patients as those who need high-acuity care.”
He also points out that frontline clinicians in his health system haven’t yet achieved peak “push” rate. But clinician engagement in home hospital is certainly much further along than when the first BPA alert was fired.
The same way that doctors assess patients in the hospital to see if they still need a certain level of monitoring or a Foley catheter or IV antibiotics, “part of that clinician checklist now is, ‘Does this patient need this hospital building?’ That’s now just another question they ask themselves, and it’s part of everyone’s thought process and standard of care.”
Getting there took him and his team three years. “We went the easy way up front, but ‘easy’ isn’t always the best for providers or patients,” Dr. Maniaci acknowledges. Putting the time and logistics in, “you certainly end up with a better payoff.”
Phyllis Maguire is Executive Editor of Today’s Hospitalist.
For more on Mayo Clinic’s hospital at home program see How to get hospitals at home up to scale.
