Data Analytics: Are They Powering Your Decision-making?

Medical imaging is in a big battle with big data. There’s too much data in too many locations, and most often they are not well managed. Data are clearly imaging’s most abundant yet most underutilized strategic asset. Radiology practices and imaging departments are deploying interactive visual analytics and dashboards in their quest to leverage existing data for heretofore hidden insights. Topping the list of benefits are enhancing care and efficiencies, improving clinical and operational processes, and saving money.  

It comes as no surprise that analyzing data is the No. 3 concern of radiology practice leaders, according to a recent poll by Radiology Business Journal and The leaders who contributed include radiologists, executives, administrators, IT managers and directors. This article is the third of three to dive into the top three pain points radiology practices face based on the survey—and to offer effective solutions.

Data analytics are the key to increasing understanding of performance to achieve the best outcomes at the lowest cost. Analytics offer readily available, actionable information that allows leadership to manage operations, performance and outcomes. To get a good result, practices need to define their metrics carefully and consolidate access to key clinical, operational and financial data. To peer inside data effectively, healthcare organizations need to leverage technology, services and staffing investments. That’s according to Francis Magann, the senior manager of business intelligence & data analytics for Change Healthcare, who has helped implement data-driven analytics solutions in health systems across the globe. “Using data well improves availability, quality, and reliability,” he says.

Magann compares data to oil. It’s one thing to know a rich resource is in the ground but quite another to find it, remove it, refine it and make it available at a consistent level of purity to make it useful for all circumstances. “Healthcare needs to turn data into information and insight, and thus make it usable,” he says. “For imaging groups with multiple facilities or health systems with multiple hospitals, it is essential to get a true enterprise picture of what’s going on to make better business decisions. Many people say, ‘Oh, analytics, we’re good. We’re getting great data out of our modalities and EMR system.’ But they aren’t. Most of them spend several days a month mining data and pulling together reports because everything is in separate silos. It’s a nightmare, and all the while imaging leadership is not getting the information they need to run their departments well.” By the time they manage to pull all the streams of information together from different sources, the data are already out of date. Many organizations are making clinical, operational and financial decisions using stale information. 

Magann has helped many imaging practices and health systems achieve a better look at the big picture by deploying analytics via a partnership with TIBCO. Change Healthcare Explorer Analytics uses TIBCO Spotfire visual analytics. Here’s how it works: Brush-linked interactive visual analytics enables users to mashup data from multiple sources, monitor key performance indicators, and spot trends and outliers—with simple drill-down to root cause. The solution includes an inbuilt AI engine that automatically identifies relationships in the data, suggests insights, and recommends visualizations directly on the data at hand. The solution’s data mashup and wrangling provides data lineage from source to analysis, and is easy to edit and apply across evolving data updates. It enables the “why,” not just the “what,” whereby analysts and consumers can quickly get to insights in the data. This is crucial to address the varying business needs across the radiology data landscape. 

Data analytics projects focus on getting the right granularity of data to the right people at the right time and in the right way so that information is relevant, current and easy to understand. The right people include C-level leaders, directors, business managers and data scientists. Each role has a different focus on the value of data and what it needs to look like to be valuable. And everyone needs to view the data through the right lens.

When data are viewed like a pyramid, the key strategic drivers of the organization sit at the apex. The C-suite may view key performance indicators (KPIs) through a high-level balance scorecard or dashboard with traffic lights. Directors and business managers get a bit more detailed and tactical to understand the day-to-day whys, whats and whens of the group or department. “As you move down the pyramid, you find business analysts and scientists who need to mine the data in greater detail and look at it in different ways,” Magann offers. “The pyramid gives you a structure to organize your data as part of a business intelligence or analytics strategy. But, more curiously, it actually works in the reverse from the base back up to the apex. That gives you your governance.”

Getting started

Breaking off the rearview mirror is one of the first lessons of data analytics. Don’t recreate reports you have always done. Instead, think: “How can I drive improvement in my organization?”

To be successful, the business intelligence strategy needs to line up with the strategic drivers of the organization, Magann urges. “Imaging groups and departments need to develop into learning organizations that are constantly improving,” he says. “That's why it's important to structure your information and information strategies appropriately to align with those strategic drivers that are pertinent to the organization at that particular point in time because business strategy must constantly refine and evolve keeping pace with the forces at play in your market.” 

Business intelligence and data analytics is a foundational capability for driving a continuous improvement culture. Establishing a business intelligence strategy is step one. Step two is creating a baseline of current clinical performance. And step three is getting your team on board. “About 70 percent of a successful data business change is the people, 20 percent is the process and 10 percent is the technology,” he says. “People are the key to developing continuous improvement cultures in our organizations. They need to move forward with new ways and not revert to inefficient ways they measured in the past.”

But who should promote that change? The chief medical officer, chief executive officer, chief finance officer, director of radiology or imaging, shift leaders and supervisors. Since change starts at the top but happens from the bottom, organizations need to engage each layer to be effective. 

Magann recommends working closely with subject matter experts to identify which data to measure and where to find them in the existing IT systems. If the organization doesn’t have the system integration skillset to pull the right data from the right systems, lean on a vendor that does. In Change Healthcare's case, they approach it like a historian going to the primary source of information because it’s the most accurate. They tap the EMR for patient information, RIS for the order, PACS for the images and dictation system for the report. And in many instances, multiple PACS must be linked even though they are configured differently. All of the data need to be standardized. 

Next they bring the data together in a single data warehouse or a more virtual and federated data option. 

“Before we go live in any BI system,” Magann notes, “we run reports on the source systems and the equivalent report on the new BI system. We compare them to validate the data, which is critical for quality because users must trust the information. We train people using their own data because it’s more meaningful to them. We switch on the feeds for a little while, typically about a month, and we find there are a lot of ah-ha moments.”

He cites the example of a heat map his team created with one imaging department. The x-axis showed the time of day over a 24-hour period while the y-axis showed the volume of studies. They selected data on two technologists over a week. Immediately the supervisor, pointing to color-coded data, said, “Hold on a second. Those guys swapped a shift without telling me.” 

And that’s where the visual storytelling of the true operational picture comes in. “Not large tables of information but digestible and actionable insight, for each role in the chain,” Magann says. “It's difficult to spot patterns when you're looking at raw data tables. With color-coding, patterns emerge.”

Bringing positive change

Next, organizations need to assess the changes. Did they turn the dial in a positive direction? The answer is almost always yes for a collection of gains in operations, finance, quality, governance and compliance. Day to day, that means increasing patient capacity and staff productivity, reducing patient wait times, improving revenue reporting, compliance and accreditation. 

Case in point is a hospital that used data analytics to reduce wait times for outpatient exams. When the team looked at exams by time of day, they noticed wait times increased first thing in the morning and trailed off mid to late morning for lunch and again after lunch. They also saw a second peak after the close of the work day. They surmised that outpatients preferred appointments in the morning and toward the end of the day after work hours. To level out the demand, they started imaging inpatients during downtimes to fill the gaps and best utilize resources. The team also adjusted their shift rotations to have more technologists starting later in the day and continuing later into the evening to cater to that secondary peak.

The imaging department also noticed that typically exams did not occupy the entire time scheduled. “So in principal, they could offer an extra few slots at the end of the day for more patients,” Magann says. “But this facility chose to go in the opposite direction and move away from an appointment-based system to a first-come, first-served basis." 

Thus, by imaging inpatients during down times, staffing appropriately for high-demand timeslots, and making more effective use of their rooms, modalities and people, the department doubled the amount of outpatient exams they did over a three-month period. Outpatient waiting times dropped from six weeks to six days. “This facility was in a major metropolitan area where there are many other imaging options,” he notes. “Clearly, they did more imaging and retained patients who didn’t need to go to competing systems down the street.”

Analytics also allow departments to dive into reimbursements and cross-reference work done versus reimbursements paid. “That’s why data validation from the source systems back into our system is so important,” Magann says. “We often find organizations are leaving money on the table. For one, that added up to close to half a million dollars annually in work that was done in the department that wasn't being reimbursed. They obviously made some changes right away.”

Taking a closer look

With experience and insight, imaging groups and departments are winning the battle with big data. By visualizing data better and gaining clinical, operational and financial insight, they are changing the way they do business. Efficiency, productivity, and overall care are improving as well. Facilities investing in data analytics prove to be more competitive and better organized. 

Data analytics is “more than bringing data scientists and skills to build data warehouses and lakes,” Magann says. “It’s understanding the imaging data, the mechanics of handling it along with the people, the process and the technology. It's really about bringing that balanced perspective to any customer engagement in a constant, systematic way.”

As he sees it: “Anyone who isn’t using data analytics should take a much closer look. It will absolutely change their business for the better."


Read about the other two concerns and solutions from the Radiology Business Practice Leaders Survey:

Orchestrating Radiology Workflow: Measuring, Managing and Load Balancing

Radiology Practice: The Answer to Managing IT Complexity