How to get from innovation to 'This is just what we do now'

By Jonah Comstock
Share

Sometimes innovation in hospitals means new technology, but sometimes it can just mean better workflows or ways of doing things. Or sometimes, for a new technology to really have value, doctors and nurses need to change their workflow around it. Creating and sustaining this behavior change can be the hardest part of an innovation.

At the Pop Health Forum in Boston this week, Dr. Leora Horwitz, director of the Center for Healthcare Innovation and Delivery Science at NYU Langone Medical Center, talked about bridging the gap between innovative idea and everyday practice.

“You have a great idea, you have an awesome intervention, you put it in place,” Horwitz said. “How do you make that stick? How do you make it ‘This is just what we do now’? How do you make something become routine in your organization?”

The first thing an organization needs to sustain an innovation is a champion, Horwitz said: Someone who will encourage hospital personnel to stick with a new way of doing things long enough for permanent behavior change to set in.

“A champion has to stick around for a while,” she said. “An intervention is like gluing a couple of pieces of wood together. Ultimately it’s going to be great, but until the glue sets you need to clamp it.”

Drawing on her experiences both at NYU and at Mount Sinai Hospital, Horwitz gave some examples of different kinds of interventions which require different approaches to sustain them.

The first category is “high value” interventions — not in the sense of value to the organization, but in the sense of immediate, obvious value to the people whose behavior is changing. Horwitz used the example of a new handoff software she created which reduced the number of notes users needed to write and rewrite by hand each day.

“All I did was say to the residents ‘Hey this new EHR has a sign off feature’ and adoption was like 100 percent instantly,” she said. “… High value interventions are the easiest because they will become sticky very fast and sustained over many years.”

But what if the intervention’s value is a benefit to a third party, and the people who have to change their behavior don’t see it? Horwitz listed two kinds of “low value” intervention. If a low-value intervention is fairly simple, the best thing to do is to automate it. Horwitz used the example of hospital doctors who weren’t putting the information in their discharge summaries that primary care doctors needed to see, like weight information for heart failure patients. Rather than trying to change the doctors’ behavior, she just made the system load that information in automatically.

But the hardest case of all is low-value interventions that are also complex. Here Horwitz talked about an ongoing effort at NYU to get patients discharged by noon in order to make more beds available for the ER and decrease ER wait time.

Her team was able to get doctors to change their behavior with a combination of performance standards, incentives, accountability, and mandates. But there is a big difference between the committed behavior of a high-value intervention and the merely compliant behavior in a low-value intervention.

A better solution, Horwitz said, is to turn low-value interventions into high-value interventions as much as possible, either by finding a way to create an intrinsic value for the people who are changing their behavior, or by helping them to see the results of it.

“Our whole dashboard is focused around how many patients we’re sending home by noon,” she said. “But I don’t really care how many patients are going home by noon, that’s not the point. The point is, can I make the beds available for my ED patients. So really what I want to show people is ‘What’s the ED wait time’?”

She suggested that creating a new dashboard that shows not just how many patients are discharged by noon but also emergency department wait times, doctors might start showing real commitment to the change.

“As time goes on you’ll get a sense of if [something] is a high value or a low value intervention,” Horwitz said, in summary. “If it’s a high value intervention, it will just happen. If it’s a low value intervention, either try to automate it or think about what kinds of standards, benchmarks, accountability, incentives, or mandates you need to set. Or try to convert it into a high-value intervention.”