Auto-filling ultrasound software significantly decreases errors in radiology reports

Radiology report quality was improved with a software solution that auto-fills ultrasound measurements and helped radiologists focus better on their imaging interpretations, according to research published Oct. 7 in Current Problems in Diagnostic Radiology.  

Reliable transmission of ultrasound measurements into radiology reports can be tainted with error due to measurements being hand-written on worksheets, dictated from worksheets or errors with images themselves on the report, wrote lead author Meir H. Scheinfeld, MD, PHD, a radiologist at Montefiore Medical Center in the Bronx, New York, and colleagues.  

To decrease valuable physician time spent checking and editing error-filled reports, the researchers implemented a software that auto-fills ultrasound measurements throughout their institution. 

“Auto-filling of ultrasound measurements streamlines ultrasound reporting and allows the radiologist to focus on the imaging findings rather than the required documentation of numerical data,” Scheinfeld et al. wrote. 

Scheinfeld and colleagues created and mapped a total of 128 fields for 39 exam templates. After implementation of the software, ultrasound technologists were trained to use institutional measurement fields to perform studies (a requirement for measurement transmission).  

The researchers found that technologist utilization was 86 to 96 percent and radiologist utilization was 92 to 93 percent, according to the researchers. Additionally, radiology resident utilization was 95 to 96 percent.  

Overall, the team estimated the new software solution would save more than $800,00 over the next five years, and that accuracy was a top priority throughout the entire software implementation and training process.  

“Training the technologists and radiologists to be vigilant for errors was an important component of the implementation,” the researchers wrote. “Technical failure and human error are inevitable, and these would need to be identified and corrected as they would occur. Therefore, despite the system working in the vast majority of cases, measurements would still need to be reviewed.”