Department experiments with adding ‘like’ button into radiology reporting workflows

A department is experimenting with adding a “like” button into radiology reporting workflows in a bid to bring more peer recognition to the specialty. 

Diagnostic radiologists rarely receive routine feedback from referring clinicians on the quality and value of their work. Burnout, disengagement and poor job satisfaction are well-documented, with a zero-feedback workflow likely adding to this problem, experts write Wednesday in Academic Radiology. 

A Japanese hospital aimed to address this by adding a single-click “like” button into their live reporting workflow, similar to social media. The pilot has proven feasible, those involved note, with sustained voluntary adoption and high satisfaction among radiologists who chose to use it. 

“Fitting feedback into the workflow is itself a known barrier to keeping peer-feedback and peer-learning programs going in radiology, and the single-click design was meant to lower that barrier,” Rintaro Ito, MD, PhD, with the Nagoya University Graduate School of Medicine in Japan, and co-authors wrote Aug. 26. “Artificial intelligence, including large language models for report generation and interpretation, is increasingly used in the reporting workflow. A simple human peer-recognition layer complements these tools because it targets the motivation and social side of reporting rather than the interpretation itself.”

Nagoya University Hospital added the like button into its existing picture archiving and communication system-integrated reporting application last year. It appears as a small, nonintrusive interface element next to each finalized report. Any authenticated user viewing the report, whether a radiologist or a referrer, could press it with one click, and no written comment was required. The hospital server logged each like with its sender, recipient, report identifier and timestamp. 

Ito and colleagues piloted the like button in routine practice for about four months. They then assessed three factors—whether it was used based on server logs, what radiologists thought based on a survey of 25 docs, and if their reporting workloads changed. Out of two dozen eligible radiologists, 11 chose to use the system. 

Altogether, 684 likes were logged during the first four months of 2025 on a total of 658 radiology reports. Overall satisfaction was significantly higher among users than nonusers, the study found. Intradepartmental intent to keep using the program was strong. Motivation and perceived usefulness favored users but were not significant, after corrections, the study found. Researchers cautioned that like-button users and nonusers answered differently worded questionnaires, limiting the ability to compare the two groups. Also, group membership was self-selected, with radiologists predisposed to value peer recognition probably more likely to adopt the system. 

Despite these limitations, researchers believe their early findings warrant further investigation via a randomized or stepped-wedge trial that gives all participants the same questionnaire and links likes to objective endpoints. These could include report turnaround times, peer-review participation, discrepancy rates, or validated burnout-measuring instruments. 

“The data support two reasonable claims,” the authors advised. “First, the system is feasible, and the radiologists who used it generally expressed favorable attitudes toward it. The strongest signal was the intradepartmental intent to keep using the system, which was the only secondary item to pass correction for multiple testing…Second, the system did not add a measurable workload penalty in our small, paired sample, although the wide minimum detectable effect means small changes cannot be ruled out.”