Using Artificial Intelligence to Enhance Patient Comprehension and Risk Perception of Pathology Reports: Findings From a Survey of Patients and Pathologists

Jordyn Franco, Brown University, Department of Family MedicineFollow
Osee Koulong Leussa, The University of Vermont, The Larner College of MedicineFollow
Amer K. Abu Alfa, Department of Pathology and Laboratory Medicine, University of Vermont Larner College of MedicineFollow
Mark K. Fung, Department of Pathology and Laboratory Medicine, University of Vermont Larner College of Medicine

ORCID

Osee Koulong Leussa: https://orcid.org/0009-0004-8822-7118

Abstract

Context: Electronic health records (EHRs) provide patients with direct access to their pathology reports, often before physician consultation. However, the complexity of medical terminology can lead to misinterpretation and anxiety. Artificial intelligence (AI), particularly large language models such as ChatGPT, may improve patient comprehension by simplifying medical language. Objective: This study evaluates ChatGPT's ability to enhance patient understanding and risk perception of pathology reports by generating patient-friendly explanations and assigning a numerical concern level. Pathologists assessed the accuracy of AI-generated summaries. Design: A publicly available colonoscopy pathology report was analyzed using ChatGPT, which produced a simplified summary and concern score. Two surveys were conducted: one among patients (n = 208) to assess comprehension and perceived usefulness and another among gastrointestinal pathologists (n = 33) to evaluate the accuracy and reliability of AI-generated concern levels. Results: Among patients, 83% found the AI-generated summary beneficial, with 58% rating it as "very helpful." However, 36% of pathologists reported omissions of key clinical details, and 26% identified misleading content. ChatGPT assigned a concern level of 6/10, whereas pathologists reported an average concern level of 2.75, suggesting an overestimation of risk. Conclusions: While AI tools like ChatGPT may improve patient comprehension of pathology reports, concerns regarding accuracy, reliability, and clinical applicability remain. Further research is required to evaluate the consistency of AI-generated interpretations and to establish standardized methodologies for integrating AI-assisted summaries into clinical workflows while ensuring alignment with expert assessment and medical guidelines.

Recommended Citation

Franco J, Leussa O, Alfa A, Fung M. Using Artificial Intelligence to Enhance Patient Comprehension and Risk Perception of Pathology Reports: Findings From a Survey of Patients and Pathologists. Patient Experience Journal. 2026; 13(1):69-78. doi: 10.35680/2372-0247.2103.