ChatGPT vs. surgeons on pancreatic cancer queries: accuracy & empathy evaluated by patients and experts
Open accessAbstract
BACKGROUND: Artificial intelligence (AI) offers potential support in patient-clinician interactions, but its impact on such communication remains unexplored. METHODS: In this study, ChatGPT was compared with two pancreatic surgeons in responding to ten pancreatic cancer surgery-related questions, co-designed with the Patient Advisory Board of the Surgical Society's Study Center. A blind evaluation of these responses, considering content congruency and clarity for non-specialists, was conducted by patients and surgeons. RESULTS: From June 23 to July 21, 2023, 24 patients and 25 surgeons participated, of which eleven patients and ten surgeons completed the survey in full. Utilizing a quantitative scale from 1 (strong-disagreement) to 5 (full-agreement), consensus was observed among patients and specialists concerning the content delivered by ChatGPT. The metrics for comprehensibility to a non-specialist audience consistently showed positive reception. In the evaluation of empathetic resonance, ChatGPT's responses mirrored those of the surgeons in the patient's view. A significant proportion ranked Surgeon 1's contributions foremost, followed closely by ChatGPT. DISCUSSION: This study demonstrates that surgeons and ChatGPT answer common queries from patients regarding pancreatic cancer surgery comparable regarding reliability, lay comprehension and empathy as evaluated by patients and surgical experts. These findings highlight the potential of AI in enhancing patient-provider interactions.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
AI personalized conversations and treatment recommendations help patients make informed decisions and improve treatment adherence.
“This helps patients make informed decisions and improves treatment adherence [57].”
From Empathic AI for Patient-Centered Cancer Care: A Scoping Review of Patient Navigation, Support, and Clinical Practice (2026) · cites this paper as [57]
The Discussion reports that ChatGPT’s answers to general pancreatic-cancer surgery questions were understandable and showed promise for patient education. It does not establish that personalized conversations or treatment recommendations helped patients make informed decisions, and it did not assess treatment adherence. (LLM, conf=0.80)
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
Survey of Hallucination in Natural Language Generation
Ji, Z., Lee, N., Frieske, R., Yu, T., Su, D., Xu, Y., Ishii, E., Bang, Y. J., Madotto, A., & Fung, P. (2023) · ACM Computing Surveys · doi:10.1145/3571730 · Cited by 5,025 · see citing works
O'Sullivan, S., Nevejans, N., Allen, C., Blyth, A., Leonard, S., Pagallo, U., Holzinger, K., Holzinger, A., Sajid, M. I., & Ashrafian, H. (2019) · The International Journal of Medical Robotics and Computer Assisted Surgery · doi:10.1002/rcs.1968 · Cited by 437 · see citing works
Hopkins, A. M., Logan, J. M., Kichenadasse, G., & Sorich, M. J. (2023) · JNCI Cancer Spectrum · doi:10.1093/jncics/pkad010 · Cited by 259 · see citing works
Flagged first, then by citations
Research works printed but not linked: 2
- Antel R, Abbasgholizadeh-Rahimi S, Guadagno E, Harley JM, Poenaru D. (2022) The use of artificial intelligence and virtual reality in doctor-patient risk communication: a scoping review. Patient Educ Counsel 105:3038-3050.
- Flesch R. (1948) A new readability yardstick. J Appl Psychol 32: 221-233. https://doi.org/10.1037/h0057532. · DOI · Paper page
Cited by
Sorted by citations
White, B. M., Aitken, G. L., Zink, J. A., Kheirkhah Rahimabad, P., Kumsa, F. A., Hashtarkhani, S., Rashid, R., Kheirinejad, S., Girdwood, T., Brett, C. L., et al. (2026) · JMIR Cancer · doi:10.2196/82336 · Cited by 2 · see citing works
“Study characteristics are summarized in Table S1 in Multimedia Appendix 1 [ 14-57 14-57 ].”
- Authors
- Klotz, R., Pausch, T. M., Kaiser, J., Joos, M. C., Hecktor, R., Ahmed, A., Dörr-Harim, C., Mehrabi, A., Loos, M., Roth, S., Michalski, C. W., & Kahlert, C.
- Affiliations
- University Hospital Heidelberg (DE)
- National Center for Tumor Diseases (DE)
- Corresponding author
- Kahlert, Christoph
- ORCID
- Venue
- HPB
- Publisher
- Elsevier BV
- Published
- 2025-03
- Volume
- 27
- Issue
- 3
- Pages
- 311-317
- ISSN
- 1365-182X (ISSN-L)
- OpenAlex
- W4404976008
- Cited by
- 18
- References
- 3
- Citations by year
- 2026: 9 · 2025: 8 · 2024: 1
- FWCI
- 1.68
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- Locations
- HPB (publishedVersion, cc-by)
- PubMed (publishedVersion)
- Topics
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Health Sciences › Health Professions › General Health Professions › Patient-Provider Communication in Healthcare
- Health Sciences › Health Professions › General Health Professions › Health Literacy and Information Accessibility
- Keywords
- ChatGPT, artificial intelligence, pancreatic cancer surgery, surgeon-patient communication, patient-provider interaction, empathy, patient understanding, patient advisory board
- MeSH terms
- Generative Artificial Intelligence, Adult, Aged, Artificial Intelligence, Attitude of Health Personnel, Communication, Empathy, Female, Humans, Male, Middle Aged, Pancreatic Neoplasms, Physician-Patient Relations, Surveys and Questionnaires, Comprehension, Social Media, Surgeons
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
