Self-Service Research Tool Empowers Clinicians in the ER and Beyond
By Matt Phillion
It can often be a struggle for healthcare researchers to find the data they need to deliver new insights, in part because access to that data can be limited by important patient-privacy regulations and technology challenges. UChicago Medicine, in the search for a path to “data democratization,” has adopted a self-service data solution that allows clinicians, researchers, staff members, and administrators to engage with their own data to ask and answer their own questions as they seek to improve quality of care.
The initiative helped the hospital:
- Empower emergency department clinicians and all UCM stakeholders to ask questions, build cohorts, and validate hypotheses, in real time
- Accelerate improvement cycles by removing data access delays that previously slowed progress in emergency care and beyond
- Identify workflow bottlenecks and protocol variations using real-world and synthetic data
- Foster a systemwide culture of inquiry and transparency
- Enhance patient safety by driving rapid-cycle learning and implementing changes informed by actual patient patterns
“Clinicians want to do good- they want to find insights, improve care, and do more to help their patients,” says Tom Spiegel, MD, chief quality officer with UChicago Medicine. “There are so many bottlenecks out there.”
One is time. Every query is unique and customized—it takes time to build those queries and reports, creating those bottlenecks.
“I think the overwhelming sentiment is we could be doing more, but the data upon which this all works is a specialized field and has to be handled correctly or you may venture off in the wrong direction,” Spiegel says. “Wanting to do more is what keeps many clinicians up at night. There’s always a quest for more data than what data teams can deliver in a timely manner.”
This process really does start with the ability to sit down at a terminal and request what you need specifically, when you need it, he explains.
“A few weeks ago, I was working a shift in the ER on a Friday night, and the doctors were talking about reviewing the use of a certain anti-nausea medication,” says Spiegel. “The situations we use this medication for have been growing, but this medication costs 100 times more than the seven other alternatives. Now, the seven alternatives cost 44 cents and this drug costs $44, but multiply that three or four times a day over a year and you’re at $50,000. If there’s not a compelling argument to use that specific costlier medication, why do it?”
And so, as part of this conversation, the physicians said: Let’s look at the data. They opened up their self-service tool, MDClone, and looked at how many patients had received this medication over the past calendar year.
“It was thousands,” says Spiegel. “This was institution-wide, with the ER the second-largest user.”
Using the tool, they were able to drill down into which departments, which clinicians had been using the drug, and more.
“In just minutes we were able to turn around and answer this question: we had the ability to do that at 10:30 on a Friday night, answer the question, do further analysis, and export that data to do more with it,” says Spiegel.
A tool researchers are looking for
Spiegel says the self-service tool can be an incredible recruiting tool for clinician talent.
“I’ve asked MD/PhDs we’re trying to recruit: ‘How long did it take to get your data requests fulfilled when you were getting your PhD?’” Spiegel says. “And they’ll answer, ‘Six weeks to three months.’ Most institutions have long delays. So, then I’d show them the tool and how you’re able to ask a fairly simple [or complex] question but get the answers right then and there. If you are working on research on a Saturday night, you can sit down and get your PhD work done. It’s such a stark contrast to the world where most academicians live.”
Data requests are also iterative. That weeks- or months-long request may then result in additional queries, whittling down or expanding the question asked to get at what you want.
The self-service system does have its limits, of course.
“You get what you ask for,” says Spiegel. “If you ask the wrong question or structure the query in a way that is nuanced, you’re going to have different results.”
To help alleviate this, the team has set up office hours to help researchers make the best use of the tool.
“We’ve instituted a process where a team of us go through and verify the queries: are they getting the information they need?” says Spiegel. “There are ways to support individual users.”
Ensuring data privacy
Healthcare starts from a strong respect for patient data privacy, but there are also additional steps that help ensure the data received in these queries maintains the expected level of privacy and security.
“There are two parts. There’s the real data that has all of the safeguards that patients would expect, and then there’s synthetic data that is also safeguarded but since it’s not tied to individual patients, it’s not as strict in terms of privacy concerns,” says Spiegel. “If you asked hospital data administrators what the number one topic they’ve spent time on in the past year would be, it’s privacy.”
Beyond the standard safeguards, there are protections on how the data can be shared so it can’t end up on a personal computer or other device.
The ER’s unique view
While the research tool can benefit anyone in the organization, the ER has a unique opportunity to use it, Spiegel says.
“In the ER, overnight we’ll have lulls between patients, when we can brainstorm, share ideas, and have phenomenal discussions,” he says. “We enjoy collegial moments and at those times, we can really refine some of those thoughts and theories.”
Those discussions can often be about operational costs, which has the added benefit of not only helping patients, but justifying the cost of the tool itself. Once the tool pays for itself, it’ll be there for other queries as well.
“A lot of the projects we’re looking at now have to do with patient safety and quality,” says Spiegel. “Early projects focused on cost justification so we can get to the point where we’re improving care. Some examples of projects we’re pursuing now include looking at preoperative care for patients, optimizing their status pre-surgery so they can recover faster. Our Quality Chief in cardiology is looking at how to better measure adherence to evidence-based medicine and follow AHA guidelines. It enables us to look at areas [where] we’re doing well and to support it, as well as see where we’re not doing as well and get better.”
There are also conversations in radiology about the number of CT scans a patient receives in their care journey to identify risks and, if necessary, find safer options.
The former UChicago Department of Medicine Quality Chief is currently retired but spending time on MDClone getting answers to questions he always wanted to know during his career, says Spiegel.
“He was a PCP and he’s been going through data on patients with high blood pressure, looking at what medications patients are on and looking for ways to optimize their care,” says Spiegel. “He’s worked with teams to build protocols for 568 patients to improve their adherence. That’s the kind of thing he’s able to do now.”
At the moment, the tool has had limited deployment, enabling champions and early adopters to work with it as additional security enhancements and other improvements are implemented.
“Once we do that, we’ll have a campus-wide launch,” says Spiegel.
They are also running a parallel project, a self-service quality improvement toolkit that walks clinicians through the steps to change a care pathway or order set based on their research: which committees they need to speak with, what forms they need to fill out, and more.
“I’m on the frontlines tonight and tomorrow and I know the frustrations clinicians have,” says Spiegel. “I think empowering people to assess their ideas and directly make changes will be amazingly impactful.”
Matt Phillion is a freelance writer covering healthcare, cybersecurity, and more. He can be reached at matthew.phillion@gmail.com.