ChatGPT-generated informed consent forms in interventional radiology: A randomized controlled evaluation of comprehension, attitudinal responses, and readability
European Journal of Radiology, cilt.195, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 195
- Basım Tarihi: 2026
- Doi Numarası: 10.1016/j.ejrad.2025.112624
- Dergi Adı: European Journal of Radiology
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, CINAHL, EMBASE, MEDLINE, Academic Search Ultimate (EBSCO)
- Anahtar Kelimeler: ChatGPT, Large language model, Interventional radiology, Informed consent, Comprehension, Readability, Attitudinal response
- Sağlık Bilimleri Üniversitesi Adresli: Hayır
Özet
Objectives To evaluate whether ChatGPT-generated informed consent forms improve comprehension, attitudinal responses, and readability compared with standard informed consent forms in two interventional radiology procedures. Materials and methods This prospective, single blind, randomized controlled study was conducted at a tertiary-care interventional radiology department between March and August 2025. A total of 122 participants with no prior exposure to the procedures were randomly assigned (1:1) to receive either a web-based Chat Generative Pre-trained Transformer (GPT)-4o-generated or a standard informed consent form for transarterial radioembolization (TARE) or percutaneous hydatid cyst treatment (PHCT). All participants completed a ten-item comprehension test and a four-item attitudinal survey. Readability was analyzed using multiple indices (Flesch–Kincaid, Gunning Fog, Coleman–Liau, SMOG, Automated Readability Index, and Ateşman) and the Patient Education Materials Assessment Tool (PEMAT). Results Participants who read ChatGPT-generated consent forms achieved higher comprehension scores than those who read standard forms (82.9 % vs. 77.3 %, p = 0.04). The benefit was consistent across procedures and education levels, with the greatest relative improvement among participants with lower educational attainment (p = 0.04). Readability analysis showed that ChatGPT-generated forms had shorter sentences and lower linguistic complexity (Flesch 39.2; Ateşman 53.9) than standard forms. Test–retest results indicated better knowledge retention for ChatGPT-generated forms, particularly in the TARE subgroup. No significant attitudinal differences were observed between groups. Conclusion ChatGPT-generated informed consent forms demonstrated superior comprehension and equivalent attitudinal responses compared with standard documents, supporting their potential as communication tools to enhance clarity and engagement in interventional radiology.