MétaCan
Menu
Back to cohort
Record W4408972937 · doi:10.5435/jaaos-d-24-01072

The Challenges of Using ChatGPT for Clinical Decision Support in Orthopaedic Surgery: A Pilot Study

2025· article· en· W4408972937 on OpenAlexaff
Michael A. McNamara, Brandon G. Hill, Peter L. Schilling

Bibliographic record

VenueJournal of the American Academy of Orthopaedic Surgeons · 2025
Typearticle
Languageen
FieldMedicine
TopicArtificial Intelligence in Healthcare and Education
Canadian institutionsObject Research Systems (Canada)
Fundersnot available
KeywordsMedicineMedical physicsGeneral surgery

Abstract

fetched live from OpenAlex

BACKGROUND: Artificial intelligence (AI) technologies have recently exploded in both accessibility and applicability, including in health care. Although studies have demonstrated its ability to adequately answer simple patient issues or multiple-choice questions, its capacity for deeper complex decision making within health care is relatively untested. In this study, we aimed to delve into AI's ability to integrate multiple clinical data sources and produce a reasonable assessment and plan, specifically in the setting of an orthopaedic surgery consultant. METHODS: Ten common fractures seen by orthopaedic surgeons in the emergency department were chosen. Consult notes from patients sustaining each of these fractures, seen at a level 1 academic trauma center between 2022 and 2023, were stripped of patient data. The history, physical examination, and imaging interpretations were then given to ChatGPT4 in raw and semistructured formats. The AI was asked to determine an assessment and plan as if it were an orthopaedic surgeon. The generated plans were then compared with the actual clinical course of the patient, as determined by our multispecialty trauma conference. RESULTS: When given both raw and semistructured formats of clinical data, ChatGPT4 determined safe and reasonable plans that included the final clinical outcome of the patient scenario. Evaluating large language models is an ongoing field of research without an established quantitative rubric; therefore, our conclusions rely on subjective comparison. CONCLUSION: When given history, physical examination, and imaging interpretations, ChatGPT is able to synthesize complex clinical data into a reasonable and most importantly safe assessment and plan for common fractures seen by orthopaedic surgeons. Evaluating large language models is an ongoing challenge; however, using actual clinical courses as a "benchmark" for comparison presents a possible avenue for further research.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

How this classification was reachedexpand

Full frame machine prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.

metaresearch head score (Codex)0.042
metaresearch head score (Gemma)0.155
Version: metacan-v3-hybrid-931329e0061cValidation status: machine_predicted_unvalidated
Candidate categoriesnone
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Observational · Consensus signal: none
GenreCandidate signal: Empirical · Consensus signal: Empirical
Teacher disagreement score0.042
Threshold uncertainty score0.223

Distilled classifier scores by category (both heads)

CategoryCodexGemma
Metaresearch0.0420.155
Meta-epidemiology (narrow)0.0010.001
Meta-epidemiology (broad)0.0010.001
Bibliometrics0.0010.001
Science and technology studies0.0010.002
Scholarly communication0.0030.004
Open science0.0030.004
Research integrity0.0020.003
Insufficient payload (model declined to judge)0.0050.001

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.266
GPT teacher head0.495
Teacher spread0.229 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it

Classification

machine, unvalidated

Machine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.

The models applied no category: nothing in the taxonomy fit this work.
Study designObservational
Domainnot available
GenreEmpirical

How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".

Quick stats

Citations5
Published2025
Admission routes1
Has abstractyes

Explore more

Same venueJournal of the American Academy of Orthopaedic SurgeonsSame topicArtificial Intelligence in Healthcare and EducationFrench-language works237,207