Patient-friendly radiology reports generated by AI significantly improve how families understand pediatric imaging results, with 92.6% of surveyed families reporting better comprehension when complex medical terminology is converted into simple, interactive explanations.
- Stanford and Duke researchers found that 92.6% of families said patient-friendly reports improved their understanding of radiology results
- 91.8% of respondents were more likely to choose the hospital in the future after accessing simplified radiology reports
- A natural language classification algorithm automatically converts standard medical reports into patient-centered interactive web formats
- Advanced imaging modalities like MRI (19.7%) and CT scans (18.9%) had higher patient-friendly report access rates than radiographs (5.4%)
- The Cures Act Final Rule requires real-time patient access to radiology reports, making simplified versions increasingly important for informed decision-making
Patient-friendly interactive radiology reports generated by a natural language classification tool can enhance family comprehension of pediatric imaging results, according to a recent study
Researchers from Stanford University in Stanford, CA, and Duke University in Durham, NC, assessed family use and experience with patient-friendly radiology reports at their institution’s children's hospitals, with findings suggesting the approach is a viable tool in the pediatric setting, noted lead author Andrew Johnston, MD, of Stanford, and colleagues.
“Terminology in radiology reports can be difficult for patients and families to comprehend, which may cause anxiety and impact the ability of patients and families to make informed decisions about the care they receive,” the group wrote. The study was published August 3 in Pediatric Radiology.
The Cures Act Final Rule implemented in 2020 requires that patients have real-time access to their radiology reports, which use terminology designed for communication among healthcare providers, the authors wrote. Efforts that have employed AI to improve the readability of radiology reports have improved patient comprehension in adult populations, yet little investigation has been done in the pediatric population, they noted.
To bridge the gap, the researchers made patient-friendly interactive reports available through their institution’s patient portal for all pediatric imaging exams performed between January 2024 through November 2025. They used a commercially available natural language classification algorithm (Scanslated Inc, Tampa, FL), which automatically converts standard radiology reports into patient-centered interactive web page formats. For the study, the group tracked portal access rates and view times and surveyed families on experience and comprehension.
Across both institutions, 391,713 imaging exams were performed during the study period. Standard reports were viewed by 50.5% of families, while patient-friendly reports were viewed by 9%, representing 17.7% of those who had already viewed the standard report.
Families who opened patient-friendly reports spent an average of 3.29 minutes viewing each one. Radiograph reports were accessed less than advanced imaging modalities, with patient-friendly access rates of 5.4% for radiographs versus 19.7% for MRI, 18.9% for CT, and 14.2% for ultrasound, the group reported.
In secondary results, a survey found that 91.8% of respondents said access to the patient-friendly report made them more likely to choose that hospital in the future; 91% felt more prepared for follow-up visits, and 92.6% said the explanations and diagrams improved their understanding.
“While initial utilization of the tool was low, providing patient-friendly interactive radiology reports is feasible for pediatric patients and may improve subjective comprehension of radiology reports,” the group wrote.
Future research efforts should focus on increasing access, assessing the impact on patient portal engagement, and exploring supplementary technologies such as full report summarization, the group concluded.
Read the full study here.
