Despite optimism toward AI implementation in radiology, how AI affects radiologist burnout in the real world remains a "black box," and the impact is yet to be determined, researchers noted in an article published on January 16 in European Radiology.
Projections include streamlining repetitive tasks, improving workflow, and reducing radiologist burnout, but when it comes to reducing burnout alone, those promises aren't bearing out just yet, according to Jay Parikh, MD, of the University of Texas MD Anderson Cancer Center in Houston, and Frank Lexa, MD, from the Radiology Leadership Institute of the American College of Radiology.
"Our analysis found inconclusive, limited data that AI reduces radiologist burnout," they stated. "The balance of data does not support that AI improves the drivers of burnout."
The authors identified the following drivers of radiologist burnout:
- Workload and job demands
- Control and flexibility
- Work-life integration
- Social support and community at work
- Organizational culture and values
- Efficiency and resources
They also identified nearly 20 key factors contributing to burnout, including high productivity expectations, compensation linked to volumes, and uneven distribution of caseload, unsupported vacation time, poor leadership behavior, and others.
For their study, Parikh and Lexa reviewed publications of AI affecting the established drivers of radiologist burnout.
First focusing on China, where there has been widespread AI adoption, the authors highlighted a national cross-sectional study of 6,726 radiologists -- 3,017 in the AI group -- that found that the weighted prevalence of burnout was significantly higher in the AI group compared with the benign AI group (40.9% vs. 38.6%; p < 0.001).
Next, in a survey of 675 members of the European Society of Radiology (ESR), AI remained significantly associated with an increased workload (p < 0.001, with an odds ratio of 10.64), according to findings.
The key findings included the following:
- Joint exposure to AI use alongside either high workload or low AI acceptance was associated with an additional risk of burnout.
- Duration of AI use amongst radiologists was significantly negatively correlated with burnout.
- Increased workload attributed to AI is at least partially from elevated processing and interpretation times.
Furthermore, Parikh and Lexa raised awareness of an issue that's been percolating.
"Stand-alone image interpretation by AI has been proposed," they noted. "In this circumstance, the radiologists will need to feel comfortable relinquishing complete autonomy to the AI platform."
The ESR survey found that 41% of responders foresaw the radiologist taking the legal responsibility of an adverse AI systems outcome, while 41% of responders favored shared responsibility by the radiologist and the manufacturer, they added.
"Some radiologists, when they disagree with the AI, now incorporate a statement in their reports as to why they are countering the AI's false positives," the team wrote. "This practice can slow down the workflow of the radiologist and reduce productivity while creating resentment and burnout from added low-value activities."
Parikh and Lexa also offered advice for leaders and radiologists. Leaders must choose systems carefully with a radiology-centric vision. Radiologists need to provide feedback to leaders who have the courage to jettison systems and/or replace programs that aren’t adding value.
Ultimately, "more primary investigations are needed to help unveil the 'black box' and help radiologists understand the true impact of this evolving technology on their overall already compromised wellness," the authors concluded.
Read the complete article here.
