# Patient Perceptions of Artificial Intelligence-Generated Kidney Transplant Information: Comparing ChatGPT With the National Kidney Foundation

**Authors:** Hwarang Stephen Han, Jihye Lee

PMC · DOI: 10.1016/j.xkme.2026.101247 · Kidney Medicine · 2026-01-08

## TL;DR

Patients with chronic kidney disease preferred AI-generated information about kidney transplants over traditional sources, suggesting AI could enhance patient education when used with professional guidance.

## Contribution

This study is the first to compare patient perceptions of AI-generated transplant information with that from a trusted health organization.

## Key findings

- Participants preferred ChatGPT responses over National Kidney Foundation responses in 81.3% of comparisons.
- ChatGPT was rated higher in information quality, empathy, and learning outcomes.
- Findings suggest AI can present transplant information in patient-friendly ways.

## Abstract

Generative artificial intelligence (AI) may help patients better understand the complexities of kidney transplantation. However, little is known about how individuals with chronic kidney disease (CKD) perceive AI-generated health information. This study assessed patient perceptions of AI-generated responses to common kidney transplant queries compared to those from a trusted health resource.

A cross-sectional online survey.

A total of 216 adults with CKD, including kidney transplant recipients, residing in the United States participated in the study.

Participants compared kidney transplant-related query responses generated by ChatGPT (GPT-4o), a widely used generative AI tool, with those provided by the National Kidney Foundation (NKF).

Participant perceptions across several domains: overall preference, perceived information quality, empathy, and learning outcomes.

Participants reviewed paired responses from both ChatGPT and NKF, presented without source attribution. Results were analyzed using mixed-effect models.

Participants preferred ChatGPT-generated responses over NKF’s in 81.3% of comparisons (P < 0.001). ChatGPT responses were rated significantly higher than NKF’s in terms of information quality, empathy, and perceived learning outcomes (all P < 0.001).

The web-based survey may not fully represent the diverse populations served by transplant centers. Limited prompts were used, which may not capture the full range of transplant scenarios. We were also unable to determine which specific features influenced participant preferences.

Generative AI platforms like ChatGPT may present information in ways that resonate with patients, potentially enhancing their education and engagement. However, as these tools are still in the early stages of integration into everyday life, their use should be guided by careful human oversight.

This study examined how adults with chronic kidney disease, including transplant recipients, engage with generative artificial intelligence (AI) to learn about kidney transplantation. Participants reviewed answers to common transplant queries from both ChatGPT, a popular AI tool, and the National Kidney Foundation, a trusted educational source, without knowing which source provided each answer. Participants preferred AI-generated responses in 81.3% of cases and rated ChatGPT higher than National Kidney Foundation in information quality, empathy, and perceived learning. These findings suggest that generative AI has the potential to present information in ways that resonate with patients. Although promising, AI tools should still be used under the guidance of healthcare professionals, as they remain in the early stages of integration into everyday life.

## Linked entities

- **Diseases:** chronic kidney disease (MONDO:0005300)

## Full-text entities

- **Diseases:** AI (MESH:C538142), CKD (MESH:D051436), Kidney Disease (MESH:D007674), hallucinations (MESH:D006212)
- **Species:** Homo sapiens (human, species) [taxon 9606]

## Full text

_Full body text omitted from this summary view._ Fetch the complete paper as Markdown: https://tomesphere.com/paper/PMC12936935/full.md

## Figures

1 figure with captions in the complete paper: https://tomesphere.com/paper/PMC12936935/full.md

## References

43 references — full list in the complete paper: https://tomesphere.com/paper/PMC12936935/full.md

---
Source: https://tomesphere.com/paper/PMC12936935