461 The Write Medicine: Comparing AI and Human Clinical Documentation in Orthopaedics
British Journal of Surgery · 10.07.2026 · 113 (Supplement_6)
DOI: 10.1093/bjs/znag063.598 ↗Kurz gefasst
Fachartikel in der Zeitschrift British Journal of Surgery, online erschienen am 10.07.2026. Verfasst von J Potter und 3 weiteren Autor:innen.
Abstract (Originalsprache)
Aim
Quality clinical documentation is essential in orthopaedic practice, but adherence to British Orthopaedic Association standards has declined amid increasing caseloads. AI medical scribes like Heidi AI use large language models to generate real-time structured documentation, potentially reducing administrative burden. However, concerns about AI accuracy persist. This study compares Heidi AI with conventional dictation-transcription for efficiency, accuracy, and readability in orthopaedics.
Method
Twenty-eight orthopaedic consultations were transcribed using Heidi AI and conventional human dictation-transcription by one consultant surgeon and one medical secretary. Heidi AI transcription was instantaneous; human transcription time included G2 software initialisation, dictation, and transcription. Accuracy was evaluated using a 10-category error classification framework applied by a GMC-registered orthopaedic doctor. Readability was assessed using SMOG Index, the Readability Index, Flesch-Kincaid Grade Level, and Flesch Reading Ease.
Results
Heidi demonstrated significantly fewer errors (5 vs 18; 0.14% vs 0.39%, p=0.0234) and instantaneous processing versus 6 minutes per consultation. G2 produced more readable text across all metrics (p<0.05). Heidi showed greater error consistency. Critically, one G2 document contained an error altering patient care plans; no such errors occurred with Heidi. Heidi AI eliminated approximately 6 minutes of administrative work per consultation, totalling nearly 3 hours saved across 28 patients.
Conclusions
Heidi AI demonstrated superior accuracy and efficiency compared to conventional transcription, eliminating critical errors that could alter patient care. Although conventional transcription achieved better readability, this is outweighed by Heidi AI's enhanced accuracy, particularly given mandatory physician review. Heidi AI represents a viable solution addressing documentation burdens while maintaining superior transcription fidelity.
Zitieren
APA
Potter, J., Miller, A., Azeez, Z., & Noureldin, S. (2026). 461 The Write Medicine: Comparing AI and Human Clinical Documentation in Orthopaedics. British Journal of Surgery, 113(Supplement_6). https://doi.org/10.1093/bjs/znag063.598
Vancouver
Potter J, Miller A, Azeez Z, Noureldin S. 461 The Write Medicine: Comparing AI and Human Clinical Documentation in Orthopaedics. British Journal of Surgery. 2026;113(Supplement_6). doi:10.1093/bjs/znag063.598
BibTeX
@article{Potter2026,
author = {Potter, J and Miller, A and Azeez, Z and Noureldin, S},
title = {461 The Write Medicine: Comparing AI and Human Clinical Documentation in Orthopaedics},
journal = {British Journal of Surgery},
year = {2026},
volume = {113},
number = {Supplement_6},
doi = {10.1093/bjs/znag063.598}
}
EUVIMED