Currently submitted to: JMIR AI
Date Submitted: Aug 20, 2026
Open Peer Review Period: Aug 28, 2026 - Oct 23, 2026
(closed for review but you can still tweet)
NOTE: This is an unreviewed Preprint
Warning: This is a unreviewed preprint (What is a preprint?). Readers are warned that the document has not been peer-reviewed by expert/patient reviewers or an academic editor, may contain misleading claims, and is likely to undergo changes before final publication, if accepted, or may have been rejected/withdrawn (a note "no longer under consideration" will appear above).
Peer review me: Readers with interest and expertise are encouraged to sign up as peer-reviewer, if the paper is within an open peer-review period (in this case, a "Peer Review Me" button to sign up as reviewer is displayed above). All preprints currently open for review are listed here. Outside of the formal open peer-review period we encourage you to tweet about the preprint.
Citation: Please cite this preprint only for review purposes or for grant applications and CVs (if you are the author).
Final version: If our system detects a final peer-reviewed "version of record" (VoR) published in any journal, a link to that VoR will appear below. Readers are then encourage to cite the VoR instead of this preprint.
Settings: If you are the author, you can login and change the preprint display settings, but the preprint URL/DOI is supposed to be stable and citable, so it should not be removed once posted.
Submit: To post your own preprint, simply submit to any JMIR journal, and choose the appropriate settings to expose your submitted version as preprint.
Warning: This is an author submission that is not peer-reviewed or edited. Preprints - unless they show as "accepted" - should not be relied on to guide clinical practice or health-related behavior and should not be reported in news media as established information.
Automating the Comprehensive Complication Index Using Large Language Models: A Methodological Validation Study
ABSTRACT
Background:
The Comprehensive Complication Index (CCI) is a validated measure of postoperative morbidity that quantifies total postoperative complication burden. It remains underutilised because manual Clavien-Dindo Classification (CDC) grading and score aggregation is tedious and susceptible to subjectivity and inconsistency. Although large language models (LLMs) can classify postoperative complications from free-text clinical documentation, their ability to compute the more complex CCI remains unclear.
Objective:
To evaluate the feasibility, accuracy and reproducibility of contemporary LLMs for automated CCI derivation from real-world surgical discharge summaries.
Methods:
Six commercially available LLMs were assessed using a tiered validation framework. After conceptual testing, ChatGPT-5.2 was selected for the main analysis as the model with the highest accuracy in complication extraction and CCI computation. 40 synthetic discharge summaries were analysed under naïve and structured prompting conditions. The prompted pipeline was then applied to 200 retrospectively anonymised real-world surgical discharge summaries. Outputs were compared with human reference values using Bland-Altman analysis, intraclass correlation coefficients, and kappa statistics.
Results:
Compared with naïve analysis, structured prompting increased agreement in LLM-derived vs human reference CCI scores of synthetic cases, reducing mean bias and narrowing the limits of agreement (0.36; LoA -8.96 to 8.24 vs 0.13; LoA -0.90 to 0.65, respectively). Prompted analysis of real-world discharge summaries equally showed a high level of agreement, and limits of agreement remained narrow (mean bias 0.18; LoA -4.71 to 5.07). Exact CCI agreement occurred in 193/200 cases (97%), and highest CDC grade agreement was near-perfect (κw = 0.99).
Conclusions:
Structured prompting results in accurate and reproducible CCI derivation from surgical discharge summaries. LLM-based CCI computation is therefore a promising tool for scalable, automated postoperative morbidity assessment. Multicenter validation is warranted before wider implementation.
Citation
Request queued. Please wait while the file is being generated. It may take some time.
Copyright
© The authors. All rights reserved. This is a privileged document currently under peer-review/community review (or an accepted/rejected manuscript). Authors have provided JMIR Publications with an exclusive license to publish this preprint on it's website for review and ahead-of-print citation purposes only. While the final peer-reviewed paper may be licensed under a cc-by license on publication, at this stage authors and publisher expressively prohibit redistribution of this draft paper other than for review purposes.