Compliments to the surgeon: An artificial intelligence-based analysis of online patient reviews on cardiac surgical care in the United States.
5 affiliations
- 1Department of Cardiovascular Sciences, East Carolina University, Greenville, NC.
- 2Department of Psychology, East Carolina University, Greenville, NC.
- 3PEP Health Inc, Kansas City, Mo.
- 4ECU Health, Greenville, NC.
- 5Division of Cardiac Surgery, Department of Surgery, Johns Hopkins University, Baltimore, Md.
Abstract
Objective
Patients increasingly describe care in their own words on public review platforms, but this evidence has been too large and unstructured to use. We aimed to develop and validate a reproducible artificial intelligence (AI)-based method to make this source usable and apply it to cardiac surgery.Methods
Patient comments on online platforms (2020-2026) were collected. Cardiac surgery comments were identified by registered specialty or, where absent, by key word filtering and AI classification (Claude API [Anthropic], validated by physician, 95% sensitivity, and 85% specificity). Sentiment and praise and complaint themes were classified by the same approach and validated by a psychologist (91% accuracy). Cardiac surgery was compared against 4 noncardiac surgery groups, and 6 cardiac surgery subgroups were compared using Wilcoxon rank-sum tests for comment ratings and χ² tests for theme comparisons.Results
Of 28.8 million comments, 14,678 cardiac surgery comments were identified. Sentiment was positive in 82%. Cardiac surgery rated higher than every noncardiac comparator, including clinically comparable specialties (4.31 vs 4.13; P < .001). The most common complaints were interpersonal (unprofessionalism, coldness, dismissiveness, and poor communication), whereas surgical complications ranked sixth. Among subgroups, aortic surgery alone rated significantly higher; transplant patients showed uniquely lower praise across themes, whereas congenital surgery combined elevated gratitude with the only elevated complaint theme.Conclusions
Patients describe cardiac surgical care with overwhelming gratitude and rate it above every comparator group examined. Where they voice dissatisfaction, it centers on how patients are treated as people rather than on the surgical outcomes, identifying communication as the principal target for improvement.
Related
- Artificial Intelligence–Based Surgical Phase Analysis Enables Objective Assessment of Surgeon Skill in Robotic Distal Gastrectomy: A Multicenter Study
- Artificial intelligence for scoliosis surgical planning and postoperative prediction
- Increase in Surgical Case Mix Index Across a Multi-Hospital System with Structured Limited Surgeon Chart Review
- Theory of Artificial Intelligence Adaptation for Linguistic Concordance: Bridging Gaps in Patient-Centered Communication
- Wetter’s Digital Twin: A Patient-Specific Framework for Artificial Intelligence and Healthcare
BibTeX
@article{dogan2026compliments,
title = {{Compliments to the surgeon: An artificial intelligence-based analysis of online patient reviews on cardiac surgical care in the United States.}},
author = {M Melih Dogan and Samuel F Sears and Verena Sarrazin and Dina Dejanovic and Maeve M Sargeant and Julie Oehlert and Joseph G Taylor and Ahmet Kilic and Shahab A Akhter and Michael J Bates and Walter Douglas Boyd and Wiley L Nifong and Benjamin C Degner},
journal = {JTCVS open},
year = {2026},
month = aug,
doi = {10.1016/j.xjon.2026.102075},
url = {https://europepmc.org/article/MED/42948423},
note = {CC BY-NC-ND},
}
RIS
TY - JOUR
TI - Compliments to the surgeon: An artificial intelligence-based analysis of online patient reviews on cardiac surgical care in the United States.
AU - M Melih Dogan
AU - Samuel F Sears
AU - Verena Sarrazin
AU - Dina Dejanovic
AU - Maeve M Sargeant
AU - Julie Oehlert
AU - Joseph G Taylor
AU - Ahmet Kilic
AU - Shahab A Akhter
AU - Michael J Bates
AU - Walter Douglas Boyd
AU - Wiley L Nifong
AU - Benjamin C Degner
PY - 2026
DA - 2026/08/21
JO - JTCVS open
DO - 10.1016/j.xjon.2026.102075
UR - https://europepmc.org/article/MED/42948423
AB - Objective Patients increasingly describe care in their own words on public review platforms, but this evidence has been too large and unstructured to use. We aimed to develop and validate a reproducible artificial intelligence (AI)-based method to make this source usable and apply it to cardiac surgery.Methods Patient comments on online platforms (2020-2026) were collected. Cardiac surgery comments were identified by registered specialty or, where absent, by key word filtering and AI classification (Claude API [Anthropic], validated by physician, 95% sensitivity, and 85% specificity). Sentiment and praise and complaint themes were classified by the same approach and validated by a psychologist (91% accuracy). Cardiac surgery was compared against 4 noncardiac surgery groups, and 6 cardiac surgery subgroups were compared using Wilcoxon rank-sum tests for comment ratings and χ² tests for theme comparisons.Results Of 28.8 million comments, 14,678 cardiac surgery comments were identified. Sentiment was positive in 82%. Cardiac surgery rated higher than every noncardiac comparator, including clinically comparable specialties (4.31 vs 4.13; P < .001). The most common complaints were interpersonal (unprofessionalism, coldness, dismissiveness, and poor communication), whereas surgical complications ranked sixth. Among subgroups, aortic surgery alone rated significantly higher; transplant patients showed uniquely lower praise across themes, whereas congenital surgery combined elevated gratitude with the only elevated complaint theme.Conclusions Patients describe cardiac surgical care with overwhelming gratitude and rate it above every comparator group examined. Where they voice dissatisfaction, it centers on how patients are treated as people rather than on the surgical outcomes, identifying communication as the principal target for improvement.
SN - 2666-2736
ER -