Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Video

Updated: May 16, 2025

An Experimental Paradigm for the Prediction of Post-Operative Pain PPOP

An Experimental Paradigm for the Prediction of Post-Operative Pain PPOP

Published on: January 27, 2010

Multimodal machine learning for predicting perioperative safety indicators in spinal surgery.

Kyle Mani¹, Thomas Scharfenberger¹, Samuel N Goldman¹

¹Albert Einstein College of Medicine, Bronx, NY, USA.

The Spine Journal : Official Journal of the North American Spine Society

|March 31, 2025

Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Clinical Faceoff: Social Media for Orthopaedic Surgeons-Tool or Vice?

Clinical orthopaedics and related research·2026

Same author

Instrumentation failure after lumbar spondylectomy for spinal tumors: a systematic review and pooled individual patient analysis.

European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society·2026

Same author

Minimally invasive approaches to intramedullary spinal cord tumors: a systematic review of techniques and outcomes.

European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society·2026

Same author

Is preoperative body mass index associated with 30-day mortality after oncologic surgery for spinal metastases? An ACS-NSQIP analysis.

Journal of clinical neuroscience : official journal of the Neurosurgical Society of Australasia·2026

Same author

Vertebral metastatic disease: A paradigm shift.

Neuro-oncology advances·2026

Same author

Cluster analysis identifies high-risk phenotypic groups for revision following 1-3 level lumbar fusion in a disadvantaged inner-city population.

North American Spine Society journal·2026

Integrating free-text data with machine learning models significantly improves predictions for spine surgery patient safety indicators like length of stay and reoperation risk. This multimodal approach enhances accuracy over traditional methods using only structured electronic health records.

Area of Science:

Artificial Intelligence in Medicine
Spine Surgery Outcomes
Predictive Analytics in Healthcare

Background:

Machine learning (ML) algorithms leverage electronic health record (EHR) data for predicting perioperative safety.
Integrating unstructured free-text EHR data via natural language processing (NLP) can potentially enhance predictive accuracy.

Purpose of the Study:

To develop and validate a multimodal ML architecture combining structured EHR data and NLP-processed free-text inputs.
To improve prediction of perioperative safety indicators (extended length of stay, 90-day reoperation, ICU admission) compared to structured data-only models.

Main Methods:

Retrospective cohort study of 1,898 spine surgery patients (2018-2023).
Utilized NLP (quanteda package, bag-of-words) to process free-text EHR data and integrated it with structured data.

Keywords:

Electronic health record Explainable AI Machine learning Natural language processing Perioperative outcomes Spinal surgery

More Related Videos

Intraoperative Ultrasound in Spinal Surgery

Intraoperative Ultrasound in Spinal Surgery

Published on: August 17, 2022

A Spine Robotic-Assisted Navigation System for Pedicle Screw Placement

A Spine Robotic-Assisted Navigation System for Pedicle Screw Placement

Published on: May 11, 2020

Related Experiment Videos

Last Updated: May 16, 2025

An Experimental Paradigm for the Prediction of Post-Operative Pain PPOP

An Experimental Paradigm for the Prediction of Post-Operative Pain PPOP

Published on: January 27, 2010

Intraoperative Ultrasound in Spinal Surgery

Intraoperative Ultrasound in Spinal Surgery

Published on: August 17, 2022

A Spine Robotic-Assisted Navigation System for Pedicle Screw Placement

A Spine Robotic-Assisted Navigation System for Pedicle Screw Placement

Published on: May 11, 2020

Trained and validated two extreme gradient boosted (XGBoost) models: a baseline (structured data only) and a multimodal (structured + free-text) model using 10-fold cross-validation.

Main Results:

Multimodal models demonstrated superior performance across all outcome measures compared to the baseline tabular model.
Area Under the ROC Curve (AUC) for multimodal models ranged from 0.827 to 0.903, outperforming baseline models (AUC 0.770-0.779).
Key predictors included patient demographics, clinical covariates, and specific free-text findings like vertebral osteomyelitis and radiculopathy.

Conclusions:

The multimodal NLP model significantly outperformed the baseline model in predicting perioperative safety indicators.
Future work involves incorporating additional data dimensions (history of present illness, physical exam, imaging) and clinical implementation.