How BayCare Health System excels in raising data literacy

Martha Heller: What does data literacy mean to BayCare Health System?

William Walders:It means that every team member operates at the top of their license. When the environmental services team who cleans our operating rooms has the data to flip an OR quickly to get a new patient in, they work more efficiently. A third of our staff are nurses. If we understand the volume of patients in the hospital and the level of care they need, and can predict future staffing needs, we provide better care for less cost.

And when scheduling patient appointments, if we know historically what percentage of them are no-shows, we’ll improve the number of patients seen by our clinical staff. So if we can see the data behind low appointment times, we can create incentive programs to book those slow times. With data literacy, our team has the capacity to ask the right questions and use data to improve our operations. 

What data is most important for you right now?

We care a lot about access to care. On average, most health systems take six weeks from when a patient needs to see a doctor to when they have an appointment. We’re using data to reduce that wait time.

Reducing healthcare costs is another key part of our mission, so healthcare economic data is very important. Then there’s clinical data: we’re investing in genomics to get ahead of cancer, which includes precision medicine and the ability to know whether a patient carries a certain gene. When we tell patients to eat healthier and exercise more to avoid heart failure, they don’t always follow that advice. But using data to advise patients to come in for imaging every year instead of every five is actionable. So, there’s a lot of data that’s important to us. For me, it depends on the day.

What are some real examples of how data is driving positive outcomes?

Our access to financial data is much faster than in the past, so we can pivot daily on financial decisions. We’ve improved patients’ ability to schedule appointments, which has led to a 30% productivity gain. The number-one revenue generator in health systems is OR time, and we have 400 ORs. We can use analytics to show the turnaround times for specific doctors. We know, for example, that Dr. Smith, who is coming in at noon, likes the bed with the head near the door, which is not our standard practice. If we don’t get that right, and the head is at the other end, she walks in, throws her hands in the air, and says, “I’ll be back in 20 minutes,” which costs us around $10K. The ability to leverage data to understand and plan for those behaviors is extremely important.

How did you improve the organization’s data literacy?

Once we set up a data architecture that provides data liquidity, where data can go everywhere, we had to teach people how to use it. We provided basic dashboards like length of stay and patient utilization, and suggested they look at them every day.

That’s the first part of data literacy: here’s the data we have. Is it helpful to you? Right away, you get 60% of people nodding their heads. The 10% who are PhD researchers want the keys to the kingdom, and we’ll make that work for them, but our first step was to get our staff to work with the data we have.

Once we established basic data literacy, we taught people some advanced data practices and trained them on our tools. My data team no longer spends 90% of their day doing pivot tables. Our workforce is doing that themselves. Staff also shares data and builds new data sets for each other. Now that they’re thirsty for it, we’re all using data more to solve problems. Data value has become a crowdsourced entity.

What’s the next step for data literacy?

We have unique and disparate lines of business: 16 hospitals, and hundreds of medical practices, surgery centers, and retail settings. When I pull a lever on one side of the system, what does it mean to the other side? That’s where we currently are from a literacy perspective, showing people a multivariate data capability, how they can work with each other, and what data literacy means to the overall system. If you’re a patient, you can go to 12 of our services, and we do a good job of ushering you through disease management. But what are we doing to optimize your experience and make sure you’re being billed in one place? What are we doing to eliminate explanation of benefits friction? How do I know what capacity looks like as I’m sending people out the door to referrals for imaging and other services? That’s the level of literacy we’re moving toward, where people think outside of their day job, and start asking more questions.

We’ve also set the expectation that if you’re in a leadership role, you need capabilities that are beyond your functional job description, including program management, financial acumen, and data literacy. That’s what I hire for at a leadership level: can you do something with data? Data literacy is becoming a condition of employment.

What is your advice to other CIOs setting up a data capability?

When people start doing smart things with data, you need to create reporting standards and version control. Let’s say I pulled data a month ago and dropped it on my laptop. By the time I’m ready to report on that data, it’s no longer accurate. We can’t let accuracy drop as more people have access to data. What’s the source of truth? Where did you get the data from? How timely is it? Good data programs need mechanisms for standards and version control.

To get a data program off the ground, you also need to show clear first order wins. For example, with the right data, we’re saving OR time and thousands of dollars. You have to build the business case, in proforma, as to why the program is going to benefit the organization, and you have to tell that story in milestones. Don’t just go in and ask for a big lump of money. You need to work in stages and then report back on the positive outcomes. And don’t get stuck on the tooling. Just build a prototype in an agile way and start delivering. From that point, your data capabilities will grow.

The other watchout is analysis paralysis. The fact is, your organization has been using data all day long for years. Start by increasing the value of the data it’s been using. You don’t have to make everything perfect before you begin.

Finally, make sure you have strong data governance. Our data governance is as good as I’ve ever seen. We know where every resource is, and we understand the next challenge, how to reprioritize as necessary, and get the resources we need. Governance is essential, as much as architecture and training the workforce. We’re by no means perfect, but we’re 80% there.