Evolving Best Practices to Help with Faster Radiology Report Turnaround

GE Healthcare

In critical care, you often hear the phrase “time is brain.” Multiple studies have attempted to quantify this. One investigation into the challenges of detecting intracranial hemorrhage[1] found that for every minute of delay, the patient accelerates through 3.1 weeks of neurologic aging. If that doesn’t underscore the need for speed in radiological reporting, what will?

The demand for this reporting continues to grow. Yet it’s clear that as radiology evolves to meet the demand, one constant remains: better patient care depends on fast turnaround on critical patient reads, especially in emergencies.

Speed isn’t everything—you need a healthy balance between timely turnaround and accuracy in providing quality patient care. And several factors can influence clinician efficiency and turnaround variability. But with all things being equal, let’s explore how radiologists can best improve report turnaround times (RTAT) even despite increasing demand for imaging.

Both proven and emerging solutions to this challenge tend to ladder up to the same three themes: prioritization, communication, integration. The following best practices may not be new to your organization but can be better supported with new advances in technology.

Some of the following ideas are taken from our recent panel webinar, Reimagining Radiology Workflow[1] .

Smarter Prioritization

Not all exams are of equal importance, and many require specialist interpretation. It’s vital to understand what exams should be given to which radiologists and when⁠—and radiologists themselves need to know which is most important to read next.

This can be challenging, especially with the large volume of imaging requests at peak times and in organizations without an effective way to prioritize them.

Many radiology workflows have two levels of prioritization: STAT and routine. Of course, this means everything comes through as STAT, and when everything is urgent, nothing is urgent. According to research presented at RSNA 2017, the overuse of STAT designation is now a key contributor to slowing down radiology workflows.

In efforts to separate STAT from truly critical, thus reducing RTAT for critical exams, radiologists often start by reorganizing the worklist.

Traditionally, this meant simply including more levels of priority. But the most forward-thinking radiology departments are beginning to leverage systems that use rules-based logic to review and prioritize all images, then quickly direct the right studies to the right radiologists.

These organizations understand that, for peak productivity, it makes sense to assign cases⁠—whether a normal head CT or the post-ops plans⁠—to the person most efficient at them, or to the most appropriate subspecialist based on availability. A recent study found that changing the reporting system from modality-based to subspecialized reduced the RTAT at major hospitals from 288 to 245 minutes[2].

Incorporating artificial intelligence (AI) allows radiologists to triage* their worklist more effectively, and ideally, simply presents the specific radiologist their next exam upon completing their last report.

Streamlined Communication

Technology can be leveraged for greater efficiency at many more stages along the radiology workflow chain beyond the read. Multidisciplinary communication is a big part of it.

In one study of an emergency department’s CT TAT delays, a three-level strategy of weekly meetings, daily huddles and secure messaging across disciplines was the major contributor to success in streamlining communication. By pushing through interventions and resolving potential delays, this simple strategy led to a 30% reduction [PR(H2] in reports that radiologists struggled to reach ED physicians, and the perceived average wait time to reach an ED physician decreased from 1-10 to 1-5 minutes[3].

Secure messaging was a tactic that came up in a recent panel discussion with Dr. Michael Lev[3] , Director of Emergency Radiology and Emergency Neuroradiology at Massachusetts General Hospital, whose care team has been focusing their efforts on auto protocoling.

In their landmark study, Quantifying the Impact of Noninterpretive Tasks on Radiology Report Turnaround Times[4], Dr. Lev and colleagues discovered that that for every time a radiologist is interrupted by a phone call, it adds about 4.25 minutes to their TAT. Many of these interruptions were just texts asking for a protocol or calls from the clinician asking after the report. The irony is that without the interruption, the report would be on its way much faster.

AI-based systems help save the time radiologists need to improve this human aspect of the care process⁠—the more efficient collaboration that needs to happen to make better decisions and provide high-quality patient care. By streamlining the worklist, AI naturally streamlines communication along the workflow chain. It has a ripple effect.

And this is important, because increasingly, true integration means more than integrating AI into parts of the workflow chain.

True Integration

Improving TAT is easier said than done in the wake of ever-increasing caseloads⁠—along with the continuous growth of image density and complexity, the threat of potential radiologist shortages[5] and the resulting pressure on radiologists. Delay reporting undermines their very role in improving quality of patient care. As the American College of Radiology’s guidelines suggest[6], radiologists are able to provide quality care only if their reports are available in a timely fashion.

Over the past two decades, new technology has driven dramatic improvements in turnaround times for radiology reports, but as these challenges persist, it’s clear that healthcare AI still holds untapped potential.

This means two things: we’ll continue to adopt more AI tools to cover the spectrum of use cases in healthcare, and these tools will become harder to manage disparately. That’s why technology enablers alone aren’t enough for truly efficient radiology workflow.

Today, “true” efficiency isn’t integrating AI for one use case or another but weaving it seamlessly throughout the radiology workflow. For some organizations, this might mean reinventing the workflow altogether.

Learn more

We discussed this in more depth in a panel discussion about why so many organizations are turning to intelligent workflow redesign to achieve clinical excellence. We heard from Dr. Lev and other imaging leaders who are evolving their radiology workflows to better meet the needs of their patients and providers, and what they’ve learned about AI’s potential today and in the future.

Watch the panelists’ discussion [4] and let their innovative thinking guide your own organization’s workflow optimization strategies.

References

  1. Saver J L. Time is brain⁠—quantified. Journal of the American Heart Association. 8 Dec. 2005. www.ahajournals.org/doi/full/10.1161/01.str.0000196957.55928.ab. Accessed 10 October 2021.
  2. Zabel A et. al. Subspecialized radiological reporting reduces radiology report turnaround time. Insights Imaging. 30 Oct. 2020. pubmed.ncbi.nlm.nih.gov/33123830/. Accessed 10 October 2021.
  3. Perotte R et. al. Improving emergency department flow: reducing turnaround time for emergent CT scans. AMIA Annual Symposium. 5 Dec. 2018. www.ncbi.nlm.nih.gov/pmc/articles/PMC6371246/. Accessed 10 October 2021.
  4. Glover M et. al. Quantifying the impact of noninterpretive tasks on radiology report turn-around times. Journal of the American College of Radiology. Nov. 2017. www.sciencedirect.com/science/article/abs/pii/S1546144017308943. Accessed 10 October 2021.
  5. Stempniak M. Physician shortage in radiology, other specialties could surpass 35,000 by 2034, but AI also a factor. Radiology Business. 16 Jun. 2021. www.radiologybusiness.com/topics/artificial-intelligence/physician-shortages-radiology-aamc-artificial-intelligence. Accessed 10 October 2021.
  6. Towbin A J et. al. Practice policy and quality initiatives: decreasing variability in turnaround time for radiographic studies from the emergency department. RadioGraphics. 2 Mar. 2013. ​​pubs.rsna.org/doi/full/10.1148/rg.332125738. Accessed 10 October 2021.