Master of innovation: Roy Rosin on high-impact practices to accelerate change

When Health and Human Services Secretary Sylvia Burwell announced an aggressive timeline for the transition from volume- to value-based payment earlier this year, it was the equivalent of reading healthcare the riot act: The status quo in healthcare is now the enemy.

One way that leading-edge healthcare organizations are responding to the mandate for change is through the development of innovation centers where ideas to improve healthcare delivery can be rapidly designed, tested and implemented.

As chief innovation officer of the Penn Medicine Center for Health Care Innovation, Philadelphia, Roy Rosin, MBA, is three years into the initiative. The former innovation officer at the software company Intuit describes healthcare as “target-rich” and its practitioners as “passionate.”

In addition to addressing institutional targets, the center distributes innovation grants to applicants across the institution and then guides them through the innovation process with the objective of scaling those ideas that show promise for measurable impact in the areas of health outcomes, patient experience and new revenue streams.

“Healthcare is so interesting to me because there are all of these really passionate individuals on the front line, and they are lacking time and resources,” he observes. “I think a lot of times, their ideas just don’t go anywhere.”

Rosin acknowledges the barrier of conservatism in healthcare that may not exist in the world of software development, but he also points to a law of nature: inertia. Partner Voice talked with Rosin about how health care organizations can jumpstart their own version of change.

Change principles

Most of the principles of change that Rosin and colleagues are applying at Penn have to do with overcoming inertia: How do you get started? How do you actually get something done?

“The biggest change in innovation over the last 15 years or so has been the methods of how quickly you can test ideas,” he explains, “ways to generate some data and evidence, not at the level of proof of an idea, but just to get it started.”

Those methods boil down to an essential four principles:

  1. Quickly get some contextual, observational insight. “Go and walk in other people’s shoes and really see for yourselves the problem in the context in which the problem happens,” he urges. “That principle of having to see not what other people say is going on, but what actually is going on is critical.”
  2. Problem definition. “The problem is so often defined in a way that either limits innovation or, frankly, it makes you literally solve the wrong problem,” he explains.
  3. Divergence. Based on empirical work by innovation scientists, the first idea is rarely the one that succeeds, Rosin says. Innovation thrives with multiple iterations and hence the need for divergence and the use of concepts like analogy, where you imagine how different industries would handle the problem.
  4. Super rapid-cycle innovation. “If you don’t have a lot of time and resources and you want to see if there is anything valuable residing in these ideas, you need a way to get off the mark a lot faster, and there are techniques for doing it way, way faster,” Rosin says.

Solving the right problem

Properly defining the problem is a critical task, Rosin explains. In his recent talk at the Health Information Management Systems Society, he shared a story about being asked to solve a problem around time-of-day discharge that illustrates just how fraught with potential error this initial step can be.

Upon hearing the problem of discharge happening later and later in the day, most people would assume that the solution was moving the metric to an earlier time in the day.

Through contextual observation (or storytelling), he learned that there were 50 liquid oncology patients and 10 beds. The real need was to get some people out in order to intake others, and in Rosin’s mind, the problem changed from time-of-day discharge, to length of stay.

He pressed on, asking why a shorter length of stay was desired. As soon as he asked that, he realized that the goal was the time it took to get a bed. When he was told that patients needed a bed so that they could receive their chemotherapy, the problem definition changed again.

“So, it’s not time-of-day discharge, and it’s not length of stay, and it’s not time to bed, it’s actually time to treatment,” he says. The new definition of the problem changed the potential solutions dramatically, including outpatient clinics and other ways to change the actual treatment without adding beds—even without necessarily getting people out of their beds any faster.

Doing good problem definition ensures that you actually solve the right problem at the right level, and that the solution is not overly prescriptive, Rosin says.

Rapid-cycle techniques

Rosin described three techniques for initiating super-rapid innovation, essentially concepts for applying the scientific method to a hypothesis and generating a bit of evidence that demonstrates there might be something to that idea in your head.

Fake backend: This technique is a good choice when an innovator wants to test an operational innovation, to avoid building for a future that may never come. “Your fake backend might be that you have humans doing the work that technology would do, or you may have some kind of solution held together by chewing gum, scotch tape and paper clips,” he suggests.

Such a solution might work for the length of the experiment and used to test in one unit or one practice for a few hours or a few days, just to see what happens.

Rosin tells the story of the orthopedic practice manager who felt strongly about same-day scheduling. Everyone loved the idea, but they could never make it happen. “He became the fake backend call center, he published same day appointments in a billboard on the web site and gave a fake call center number,” Rosin says. “It was his cell phone number.”

Vapor test. Vapor refers to something that does not exist, great for what Rosin calls contextual demand testing. Surveys and focus groups can be used, but there is a huge gap between what people say and what they do.

“Many times, when you get an out-of-stock message on the Web, that thing is not actually out of stock, it never existed,” he explains. “In health care, you may see things like, ‘We are not accepting new patients at this time,’ so there’s some soft landing there.”

Fake front-end. A fake front-end employs tactics like paper protocoling in the context of redesigning workflow, either quickly on paper or using some other similar medium.

Rosin describes work being done by a care unit that was seeing children present with sickle cell anemia. “They felt that part of the population had a low chance of the probability of a serious bacterial infection, but there was a debate about whether those people really could be sent home because, historically, they were all admitted,” Rosin explains.

The solution was to have the physicians decide whether they thought patients could go home or not, but they admitted 100% of the patients so that they could compare the decision to what actually happened.

With time-to-change a priority in healthcare, Rosin sees rapid-cycle experiments as a pipeline to rapid innovation. He cites the orthopedic practice manager who became his own call center by putting a new idea into motion and seeing what happened. “Again, the evidence isn’t like randomized controlled trials, gold standard absolute proof evidence,” he says. “It’s the fact that something here might work. There is some reason to invest.”

Leadership, metrics and technology

There’s a huge role for leadership to play in innovation, Rosin says.  Leaders must create some capacity for innovation. “For people to reimagine their world, they need time to do so,” he says. “Finding that time, setting clear goals and targets is up to leadership, both in terms of what matters to the system, or beyond the system, in the world of public policy.”

In turn, the emergence of rapid-cycle innovation has changed the role of leadership. Instead of thinking they can get to the next solution by making a few big bets, leaders must make many more smaller bets and guide the teams to make sure there are clear hypotheses and that the experiments are set up to generate data more quickly.

Likewise, metrics have changed. Initially, efficacy metrics are used to measure whether, on a small scale, you are moving in the right direction, whether you have moved the needle. For instance, did you lower the readmission rate?

“It is very hard to tell which ideas will make it and which won’t,” Rosin says. “Most innovation fails from scaling prematurely: You try to go big before getting it right, when it’s all about getting it right.”

Technology plays a number of roles at the Penn innovation center. Big data has enabled teams at Penn to change the problem definition for treatment of sepsis from getting the drugs to patients within 60 minutes to moving detection back a few days. Smart pill bottles are enabling the health system to better monitor medication compliance and open the door to building social support around the patient.

“While I don’t equate technology with innovation, technology is required to scale innovation,” Rosin says.

A multitude of targets

Pretty much anywhere you look in healthcare, you will find a target for innovation, Rosin says. His group focuses on healthcare delivery and, by definition, behavior-change work, such as medication adherence, factors surrounding readmissions and ways to detect developing issues between normally episodic care to enable earlier intervention.

“We try to find the issues that drive the most cost and the most misery,” he says, listing hypertension, smoking, colorectal cancer screening rates and avoiding unnecessary care.

Making sure incidental findings on radiology reports don’t fall through the cracks is a target being pursued by a team of Penn radiologists through an innovation grant from the center. “They created a really interesting tracking system and database where they were able to actually notate the request to follow-up and get in sync with the referring physician for whom the follow-up was appropriate,” he says.

The system tracks whether the follow-up appointment was scheduled and happens, and raises a red flag if necessary. “Sometimes these things developed into problems for the patients, and sometimes it created a liability issue for the system,” he notes.

With the advent of accountable care, Rosin is enthusiastic about the prospect of payment incentives aligning with high value care. “It’s really a fun time to be in healthcare,” he says. “Innovation now means that you…get to do the right thing. It is hard to put your finger on where to play because the [opportunities are] all over the map.”