Making analytics work for quality improvement? Harder than it looks
It's one thing to say you're going to embrace data analytics; it's another to do it. And it's something else entirely to actually turn the insights derived from clinical and business intelligence into better care and lower costs.
As Director of Government Programs at Providence Health & Services, the third largest non-profit health system in the U.S., Ray Manahan has his work cut out for him. His job is to lead a team that keeps track of ever-evolving government payment programs, researching their requirements and financial impacts and communicating them to organizational leadership, from the C-suite to quality chiefs and physicians.
Educating those disparate groups about the process changes and operational adjustments necessary to meet the moving targets established by the Centers for Medicare & Medicaid services would be challenging enough for a normal-sized system. The fact that Providence has 33 hospitals across five states, means ensuring compliance across the enterprise is a challenge, to say the least.
At the Healthcare IT News Big Data and Healthcare Analytics Forum in Boston next week, Manahan will offer some insights into the work he does to help prevent Providence from paying millions of dollars in penalties for not meeting CMS-required thresholds. It's not an easy task, and in some cases, the penalties are hard to avoid.
His presentation, Ready, Set, Go: Formulating Actionable Data to Drive Value, will explain the challenges he and his team have faced as the clinical and financial data across those 33 hospitals has proliferated and grown in complexity. He'll offer tips and best practices for data collection, analysis, visualization and, crucially, the keys to communicating the lessons learned from the data to the people responsible for turning insights into positive change.
[Learn more: Meet the speakers at the Big Data and Healthcare Analytics Forum.]
That, after all, is what "actionable data" is all about. Otherwise, it's just numbers and words and cool-looking dashboards.
The name of the game, Manahan tells Healthcare IT News, is "interpreting the data so our end-users understand what change needs to happen."
"When I say actionable data, it's simply taking these scores and putting them on a report card: Marking any given hospital within our 33-hospital system and rolling it up into a document that says you are red, green or yellow. And we need to draw attention to the appropriate stakeholders to make sure we're turning that green. Otherwise we're going to be hit with another penalty," he continues.
But of course that's easier said than done. Moving the needle on some of these CMS measures "truly does become a workflow issue for us: how our end-users are taking care of patients to provide a better patient experience, so this lands with our clinicians," says Manahan. "It's a lot of work to improve these scores."
Three specific measures Manahan will focus on in Boston: hospital acquired conditions, value-based purchasing and hospital readmissions.
For many of those, "we're in the reds and yellows, still," he says. "But we have some greens, and we need to home in on those."
That can be a challenge, especially because all of the measures have different timelines associated with them and, sometimes, the measurement thresholds seem unclear.
To take just one example, for hospital acquired conditions, "there are two measures in there, they call them Domain 1 and Domain 2," he says. "The threshold given by CMS is that we need to get below 5 for our score."
For one thing, the performance period for one Domain 1 (Patient Safety Indicators 90 or PSI 90, a weighted average of various expected safety measures) is from July 1, 2013 to July 30, 2015. The period for Domain 2, meanwhile (catheter-associated urinary tract, central-line associated bloodstream and surgical site infections or Cauti/Clabsi/SSI) runs from Jan. 1, 2014 to Dec 31, 2015.
"Why can't we align those performance periods if they're falling under a specific measure?" asks Manahan, not unreasonably.
Further muddying the waters, he says, is the fact the weight behind the performance scores for each domain is different – 35 percent for PSI 90, and 65 percent for CAUTI/CLABSI/SSI.
Performing analytics on hospital data for these very specific indicators is one thing. But turning that insight into concrete plans for performance improvement – and communicating them to hundreds or even thousands of clinicians – is quite another.
"There's all this math," says Manahan. "Now try explaining that to a doctor when their most important priority is to provide excellent patient care."
At Providence, by the way, there are about 4,500 of those.
"To make something actionable is very, very difficult here because of the size of our enterprise," he says.
That, as much as any complicated hoop-jumping required by Washington, is one of the biggest challenges faced by the sprawling Providence system.
So one strategy, says Manahan, has been to find areas where it's possible to make use of high peformers: "There are hospitals within our system that have great teams, primarily led by quality leaders – epidemiologists who understand these programs and can help us deliver the right message based on who the stakeholder is.
"It could be a nurse, could be a physician," he adds. "It could be someone from quality who works directly with the physicians. But I think finding the right peers to drive the message and finding some of these champions is one of these things we're accentuating more and more."
One strategy Providence Health & Services is embracing is to hold a summit this fall, convening those hospitals most often in the "green" to help offer insights, lessons learned and best practices to those still muddling through the reds and yellows.
These subject matter experts, these so-called "super users," are key to getting the message across to the clinical folks on the frontlines responsible for really making some of these quality improvement changes become a reality. Many of them, after all, are likely to be skeptical of a number-cruncher telling them how to do things.
They'll usually listen to their peers, however. "We're kind of the middle men," says Manahan. "They can have that kind of dialogue. Super users are important. If you have a physician champion, you're going to be in a good spot. They're the ones that can walk the floors with the other clinicians and really help understand the need for the change."
Ray Manahan presents on Thursday, Nov. 5, at the Healthcare IT News Big Data and Healthcare Analytics Forum. Register here.
See also:
Providence Health & Services upgrades EMR capabilities
Providence Health & Services expands data platform deployment