The pressure to get faster, more efficient, and more productive is mounting. Healthcare reform, increasing regulations, and declining reimbursements are bearing down.
The pressure to get faster, more efficient, and more productive is mounting. Healthcare reform, increasing regulations, and declining reimbursements are bearing down. Radiology departments and imaging centers are looking for ways to improve performance. Dashboards are lighting the way.
Dashboards that analyze data going back months or years show where a facility or department has been. Executive dashboards show whether a facility is aligned with the goals of the facility. They are actually reports, prepared by hand, from data laboriously pulled from PACS or RIS and assembled in a spreadsheet. They provide a retrospective on the past quarter or year, laying the foundation for trending analyses, which can be used to predict the future path of an organization.
Tactical dashboards are different. They uncover problems that need to be fixed and the details necessary to put those fixes in place. They are the glitz of healthcare analytics, displaying information in real-time about key indicators such as scanner volumes or report turnaround times.
There aren’t many of these around. The most useful are custom-made at large academic radiology departments, such as the one at the University of Chicago. Commercially available tactical dashboards are typically tightly focused and, therefore, tell only part of the story.
“The big established companies that are already in the RIS/PACS business realize that collecting this information into sophisticated databases is going to become important,” said Dr. Raymond Geis, a neuroradiologist and advisor on imaging informatics for Advanced Medical Imaging Consultants, which interprets studies for 26 hospitals.
The need for such electronic dashboards is palpable. Efficiency has moved from a luxury to a necessity. Bottlenecks just happen, like traffic jams on a highway. Multiple factors, rather than a single, easily defined cause, are usually at fault.
Tactical dashboards hold the potential to make daily operations transparent, defining them in terms of metrics that quantify staff productivity, the utilization of scanners, and wait times. The complexity of these processes obscures them from view.
Questions about growing patient backlogs or increasing complaints about wait times may spur a facility to improve efficiency. Breaking workflow into its many parts reveals the process, like removing the cowling from a machine allows the gears underneath to be inspected. The one or several that are slipping become immediately obvious. Finding and correcting problems typically improves the quality of the service and, consequently, patient and referring physician satisfaction. It can also lead to increased revenue. Monitoring processes ensures that the fixes remain in place.
Finding the root problems may take some deep drilling. Techniques are being built into electronic dashboards to automatically track and process, then visually display certain key performance indicators. Depending on the sophistication of the dashboard, multiple sources of information may be tapped, including RIS, PACS, speech recognition, and order entry systems.
Paul G. Nagy, Ph.D., director of quality and informatics research and an associate professor of radiology at the University of Maryland, uses a home-grown web-based dashboard to automatically extract, process, and display performance indicators. The data are analyzed and presented twice monthly at departmental operational meetings.
“We want to stop driving with a rearview mirror,” Nagy said. “We want to look forward rather than backward.”
The University of Maryland experience typifies pioneering efforts to develop tactical dashboards. They are usually developed internally and under the direction of radiology chairs, he said.
“They wanted to measure things that they were getting called on the carpet about by the dean, or things that kept becoming fires,” Nagy said.
These efforts proved that specific metrics can be extracted from clinical information systems and evaluated to get a heads up on key operations. This led to more sophisticated systems, like that at the University of Maryland, where operational data are extracted and stored in a centralized database for analysis. Higher level analytics are performed on data collected over 24 months to reveal the root causes of problems in productivity, performance, and quality of service.
Key performance indicators in radiology are graphically displayed as four types of gauges. Together they provide a comprehensive picture of operations in the radiology department. Individually these metrics provide a look at specific potentially troublesome areas.
The percentage of outpatients seen in less than 15 minutes helps define performance in the “order and arrival” category. Image quality graded on a five-point scale is among the “acquisition” metrics. Percentage of reports not dictated within a month helps define “interpretation” performance. Reports unsigned after more than two weeks addresses “reporting.”
Nagy and colleagues can drill into any of these. Drilling for outpatient wait times in mammography, for example, reveals average outpatient waiting time as a function of the time of day.
“You want to get details and try to drill through the data to get the actual story for individual failure modes,” he said.
Vendors lately have taken a swing at dashboards. Some provide a retrospective on a facility’s performance. Others provide real-time feedback on operations. A few fall somewhere in between, offering more or less of both.
Siemens offers its syngo Portal Executive. This dashboard tool works in concert with syngo software to keep track of referral patterns, analyzing, for example, whether patient flow from key referring physicians is continuing, increasing, or decreasing. Portal Executive can also be used to look for-and find clues to remedy-inefficiencies that are holding back exam volumes. This, however, is retrospective. Kurt Reiff, vice president of image knowledge and management for Siemens Healthcare, is driving the company toward a more ambitious future.
“What we really need is perspective,” Reiff said. “If my business develops in a [particular] direction, how do I have to invest my money? Do I buy a CT or an MR? Where is my best return on investment? I think this is where this is heading.”
In the meantime, the real-time feedback possible with some dashboards offers the opportunity to spot problems in their earliest stages. Needles, bars, or numbers moving into red, yellow, or green provide an instant read on what’s happening in the department or facility.
Carestream’s Digital Dashboard monitors server performance, user volumes, and storage utilization. The dashboard supports Carestream’s RIS/PACS, but also can verify that other vendors’ devices are connected and operating on the network. The data can be presented onscreen or through e-mails and text messages that alert recipients to problems such as a failure in network connectivity or storage issues.
Leslie M. Beidleman uses the Carestream dashboard to ride herd on PACS at about a half dozen hospitals and clinics in the Mercy Health Partners network in Toledo, OH. As regional PACS administrator, Beidleman uses the dashboard to keep tabs on the amount of space available on the facilities’ several servers, as well as to check processes such as archiving and study backup.
“It gives a good quick view of things,” she said.
Those data points can also be manually captured in a spreadsheet for later comparison, a trick that came in handy recently when Beidleman noticed that the number of studies needing backup was increasing. This led to finding the problem on the servers, which brought the number down very quickly.
At the June Society of Imaging Informatics in Medicine meeting in Minneapolis, GE provided a glimpse of its Centricity Dashboard, a work-in-progress scheduled to go commercial the end of this year. Metrics built into the dashboard are oriented toward productivity, reflecting patient throughput, turnaround time, and scanner usage. Color-coded displays are designed to tip off administrators to impending problems.
PACS, RIS, and other digital IT systems make dashboards possible, just as dashboards make radiology operations transparent. It’s a symbiotic relationship that bodes well for radiology. By spotting inefficiencies in workflow, dashboards may provide the means for finally achieving the potential of PACS, which entered radiology decades ago on the promise that replacing film and light boxes with digital workstations would accelerate productivity.
The key performance indicators on which individual dashboard gauges are based provide the immediate feedback to spot and fix problems when they occur. In the same way, detailed business analytics of cumulative data over months or years provide the basis for strategic decisions that may be critically important to the success of a radiology operation. Many indicators are chosen on the basis of what is important to individual facilities and departments.
A survey conducted eight years ago by the radiology department at Brigham and Women’s Hospital found that about 95% of academic radiology departments in the U.S. measure their performance. But there was little consistency from one to the next in the types of indicators or the frequency with which they were monitored.
Economic forces have intensified the pressure on departments to monitor operations. In response, administrators have tightened their focus on certain key metrics, according to Pablo Ros, who led the survey. Which of those indicators receives the most attention depends on the individual department and the people in it. Ros, now chair of the radiology department at Case Western Reserve University and University Hospitals of Cleveland, says preferences vary depending on the position: department chair or hospital director, individual radiologist or section chief.
“But there are certain things that are pretty basic,” he said.
When looking at the operation of a department, examination volume and full time equivalents (FTEs), are key indicators, as they relate to productivity and financial performance. Technical and professional relative value units (RVUs) are needed to correctly interpret the data.
“If you do 100,000 studies in a month, you might think that’s great. But it makes a difference if half are CTs and the other half are plain films,” Ros said. “The mantra today is to increase productivity and decrease the cost per unit.”
Productivity, finances, and quality are woven tightly together. If workflow improves, financials go up, as does the quality of services. Reduce the number of retakes by improving scan quality and productivity improves, as does the bottom line. Patient and referring physician satisfaction rise. Demand may rise. Workflow becomes more challenging. Pressure mounts for greater efficiency, which is achieved by optimizing scan volume and minimizing patient wait times.
Demand may vary across several sites of an enterprise, as referring physicians send more or fewer patients later or earlier in the day to one place or another. When managing that enterprise, scheduling changes smooth the flow. Proactive monitoring keeps services on track while trending analyses allow strategic planning to meet future needs through asset utilization.
Such introspection works well when the goal is to improve individual performance. But a myopic view of the world can lead to missed opportunities. With McKesson’s benchmark collaborative, radiology departments can see how they stack up against others that are similar in size, makeup, and procedures. Reports can be customized to include certain specific metrics that have meaning to just your radiology department. This, according to McKesson executives, produces business information about inefficiencies, which can be turned into models that illustrate how those inefficiencies can be fixed. These lead to strategies about how to use staff and equipment more efficiently and effectively.
“It is a gold mine of information about their internal operational performance that they simply did not have before,” said George Kovacs, director of product marketing in McKesson’s Medical Imaging Group.
McKesson’s report allows the user to drill down starting at the most general comparison, a piece of a bar or pie chart, and see what’s underneath. Users can follow evidence that leads to the root cause of an operational inefficiency and then see how different actions, such as those taken by others in the collaborative, may resolve the problem.
Such comparisons provide the context for making strategic decisions. But even such complex and contextual analyses are only that, analyses. The ultimate expression of the dashboard is as a means to solve the problems. The automation that goes into summing up operational data needs to be extended to automatically fix problems rather than simply alert administrators to their presence.
“We want to identify the problem and instantiate the appropriate business logic to fix it,” said Dr. Paul Chang, vice chair of radiology informatics and pathology informatics at the University of Chicago.
For example, rather than programming an electronic dashboard to flash red when patient wait times go beyond a certain point, software might automatically reroute patients among scanners to spread the load.
Business intelligence can play a major role in safety as well, for example, in radiation monitoring. Rather than documenting the number of patients overexposed to radiation and waiting for human intervention to change the protocols, Chang advocates building software that recognizes and then stops the overexposure before it happens.
Chang, the founder of Stentor, a PACS company acquired by Philips in 2005, today runs a business intelligence analytics group at the University of Chicago focused on identifying and building upon the business intelligence efforts developed in nonmedical industries, such as banking and manufacturing, where software finds and automatically fixes problems, then tells human operators about them.
“We are a good 10 years behind any other industry when it comes to business intelligence and analytics,” Chang said. “There is nothing fancy about this.”
Dashboards are not where we need to be technologically, Chang said, but they are a good first step. Problems must be understood to be solved. Much can be learned about finding and solving them by looking outside medicine.
“Everyone thinks that our problems are very different from those of other industries, but they really aren’t,” Chang said. “Other industries have to be safe; they have to be efficient. They have had more experience and applied more resources to look at their problems.”
As efforts to include medical imaging in the electronic medical record proceed, the importance of considering the global nature of healthcare IT must be considered. A radiology dashboard may tap several sources of data, but to improve the overall efficiency of caring for patients, data from other IT systems, such as ones associated with pathology, pharmacy, patient scheduling, and computerized physician order entry must be added. Extending the dashboard to include these data and evolving the technology to become a business intelligence solution will make it useful to widely divergent groups.
Today’s radiology dashboards are informational at best. The widespread adoption of RIS and PACS provides operational data to be tapped. Ros believes there is tremendous opportunity for vendors to create dashboards that deliver comprehensive operational analyses. But Geis thinks it may be awhile before this opportunity is fully realized.
“It’s a couple years off before there are truly viable products and their use can be widespread,” he said.
When they do arrive, dashboards will provide the means to several ends: information that can be used by radiology departments and imaging facilities to increase productivity, financial performance, and quality.