Multicenter Validation of Natural Language Processing Algorithms for the Detection of Common Data Elements in Operative Notes for Total Hip Arthroplasty: Algorithm Development and Validation.

Peijin Han Sunyang Fu Julie Kolis Richard E Hughes Brian R Hallstrom Martha L Carvour Hilal Maradit-Kremers Sunghwan Sohn V G Vinod Vydiswaran

Published in: JMIR medical informatics (2022)

High performance across centers was achieved for the MedTagger-THA algorithms, demonstrating that they were sufficiently implementable, usable, and portable to different deployment sites. This study provided important lessons learned during the model deployment and validation processes, and it can serve as a reference for transferring rule-based electronic health record models.

Keyphrases

electronic health record
machine learning
total hip arthroplasty
deep learning
clinical decision support
adverse drug
big data
autism spectrum disorder
artificial intelligence
cross sectional
loop mediated isothermal amplification
double blind