There are three important things that healthcare chief information officers (CIOs) can learn from radiology data analytics -- and one important thing they can't, according to radiology informatics experts.
Analytics is at the heart of value-based care, with the potential to accurately measure the specialty's contribution to patient outcomes and provide artificial intelligence tools that could transform the practice of radiology in the future. But the approach still has limitations and needs to be viewed within a broader understanding of radiology workflow.
Here are three things healthcare CIOs can learn with access to radiology analytics.
1. Analytics can illustrate radiology's value to the healthcare enterprise.
The radiology department is a "well-oiled machine," according to Nadim Daher, principal healthcare analyst for market research firm Frost & Sullivan. That's largely because of the specialty's long-term embrace of technology tools such as PACS and RIS to improve productivity. Because it's such a data-driven specialty, there are real opportunities to redefine the value of radiology to the executive suite in large hospitals and integrated delivery networks, he believes.
To figure out the best metrics to track, Daher recommends that CIOs sit down with PACS administrators and radiology directors at their hospitals. Collaborating with these members of the radiology department will help determine the type and amount of data available, which Daher says is typically massive and generally untapped by the healthcare organization. Another goal for CIOs should be to figure out how to align the radiology department with the hospital's enterprise-level initiatives on the analytics front.
At a foundational level, radiology analytics can be used to measure the performance of radiologists, technologists, and the various modalities used within radiology departments, according to Katherine Andriole, PhD, director of imaging informatics at Brigham and Women's Hospital. With access to these tools, healthcare CIOs can determine metrics such as the number head and neck CT scans in a given month, for example. Another key measure tracked by radiology departments around the U.S. is report turnaround time, she said.
Provider organizations that want to reduce report turnaround times can take a hard look at the steps involved between booking a patient for an exam and the availability of the report for the referring physician. Radiology analytics can inform all of this, Andriole said.
2. Radiology has a meaningful contribution to make within the context of value-based care.
Payors and providers continue to collaborate on treatment pathways, which are evidence-based protocols for treating breast, colorectal, or lung cancer, for example. Frost & Sullivan's Daher believes that analytics can reveal the value of radiology to these pathways.
"People are no longer talking about imaging in a vacuum," he said. Rather, imaging is one of the many "weapons at hand" in the treatment of cancer. That's because in the world of value-based care, healthcare providers will need to show that the appropriate imaging procedure is being used and recommend the appropriate payment for that procedure -- and then tie that information to results, such as early detection of breast or lung cancer, he added.
For example, providers will be able to use radiology analytics to illustrate that the $1,000 advanced scan ordered by an oncologist is an important part of a cancer pathway, according to Daher. Still, the cost of drugs are really top of mind in healthcare today; next in line are imaging and labs, he said.
3. Artificial intelligence may transform the practice of radiology in the future.
IBM's artificial intelligence work with its Watson solution -- combined with the company's 2015 acquisition of Merge Healthcare -- could introduce the use of artificial intelligence to interpret imaging exams in the future. "If Watson could win at 'Jeopardy,' imagine what access to hundreds of radiology images could do to impact the interpretation of routine radiology exams," Daher said.
He speculates that using artificial intelligence for radiology interpretations is probably 10 years away. Still, Daher foresees a day when radiologists will focus primarily on interpreting the more complex exams that require a thorough understanding of the patient's history and an in-depth consultation with the rest of the patient's care team.
Describing the use of artificial intelligence as "burgeoning," he anticipates seeing increased use of the technology over the next five to seven years for interpreting more routine exams, such as x-rays. Any changes to interpreting behavior would need to be approved by the U.S. Food and Drug Administration (FDA), since all interpretations must be signed off by a radiologist, he said.
Further, with its Watson solution, IBM is just one of a number of companies undertaking this type of work. There are also start-up firms that are getting into this space, Daher said.
Despite the potential, there's one thing radiology analytics can't do for CIOs: Radiology analytics does not tell you everything you need to know.
When it comes to measuring the productivity of radiologists, CIOs should know that some images take longer to interpret than others. What that means on a practical level is that CIOs shouldn't criticize one of their neuroradiologists because he or she is not reading as many exams as a colleague who is interpreting chest x-rays in the intensive care unit, Andriole said.
CIOs also need to understand the typical workflow of radiologists, she said. Physicians who work in the clinic environment and orthopedic surgeons, for example, review a patient's chart, look at images and labs, and then make a diagnosis -- and that can include disease management or surgery or rehabilitation. Her point being that these physicians have a patient-centric workflow.
Meanwhile, radiologists have a very exam-centric workflow. While other physicians look at their day in terms of the number of patients they'll see or the number of surgeries they'll perform, radiologists view their days by the number of exams they need to interpret.
She also pointed out that radiologists are most efficient when they can work in a quiet environment, which is something else CIOs might not understand.
"The clinic environment is very hectic with a lot of interruptions, and radiologists will be less efficient in these environments," Andriole said.
