MétaCan
Menu
Back to cohort

Evaluation of ChatGPT-4o as a Patient Information Tool for Common Orthopaedic Surgeries: Accuracy, Completeness, and Clinical Utility

2025· article· en· W4417524748 on OpenAlexaff
Levi M. Travis, Soumil Prasad, Selina Deiparine, William A. Marmor, Michael G. Rizzo

Bibliographic record

VenueJAAOS Global Research and Reviews · 2025
Typearticle
Languageen
FieldMedicine
TopicArtificial Intelligence in Healthcare and Education
Canadian institutionsObject Research Systems (Canada)
Fundersnot available
KeywordsInformation systemOrthopedic surgeryMEDLINEPatient care

Abstract

fetched live from OpenAlex

INTRODUCTION: Artificial intelligence chatbots, such as ChatGPT-4o ("omni"), a large language model developed by OpenAI that integrates text, image, and audio processing with web connectivity, have gained traction as potential patient education tools in orthopaedic surgery. This study aimed to evaluate the accuracy, completeness, and clinical utility of ChatGPT-4o's responses to common patient questions about six widely performed orthopaedic procedures. METHODS: We assessed ChatGPT-4o's responses to five standardized patient-oriented queries for total knee arthroplasty, total hip arthroplasty, anterior cruciate ligament reconstruction, rotator cuff repair, anterior cervical diskectomy and fusion, and carpal tunnel release. Responses were generated using ChatGPT-4o's web-enabled version in January 2025. Two resident orthopaedic surgeons independently rated each response for accuracy, completeness, layperson clarity, misleading content, and conciseness using a structured binary rubric. The validated DISCERN instrument (16 items, max score 80) was adapted for quantitative assessment of information quality. Interrater reliability was assessed with Cohen kappa. RESULTS: Overall, ChatGPT-4o generated accurate and structured responses, free of overt errors. The average DISCERN score across procedures was 43.5, classifying the information as fair. The highest average DISCERN score was for anterior cervical diskectomy and fusion (mean 45.8 ± 10.1), whereas the lowest was for rotator cuff repair (mean 41.6 ± 5.9). Factual accuracy was high (>90%), but 36% of responses contained some misleading or incomplete information. Responses explaining treatment alternatives were the most accurate and complete, whereas those outlining surgical risks performed worst. Interrater agreement was good (Cohen kappa = 0.64). DISCUSSION: ChatGPT-4o provided generally accurate, clear, and empathetic explanations of common orthopaedic surgeries, offering a promising adjunct to conventional patient education. However, key limitations particularly regarding alternative treatments, nuanced risks, and lack of tailored advice limit its stand-alone use in clinical practice. Careful oversight and clinician vetting remain essential. CONCLUSIONS: ChatGPT-4o can supplement orthopaedic patient education by offering accessible, engaging content. However, notablenotable gaps in detail and occasional misleading information necessitate careful review and contextual explanation by orthopaedic surgeons.

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)0.040
metaresearch head score (Gemma)0.162
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: Observational
GenreCandidate signal: Empirical · Consensus signal: Empirical
Teacher disagreement score0.040
Threshold uncertainty score0.211

Distilled classifier scores by category (both heads)

CategoryCodexGemma
Metaresearch0.0400.162
Meta-epidemiology (narrow)0.0010.000
Meta-epidemiology (broad)0.0010.001
Bibliometrics0.0040.002
Science and technology studies0.0010.001
Scholarly communication0.0010.001
Open science0.0010.004
Research integrity0.0010.001
Insufficient payload (model declined to judge)0.0020.001

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 head0.528
GPT teacher head0.621
Teacher spread0.093 · 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

Citations0
Published2025
Admission routes1
Has abstractyes

Explore more

Same venueJAAOS Global Research and ReviewsSame topicArtificial Intelligence in Healthcare and EducationFrench-language works237,207