Predicting Anterior Cruciate Ligament Reconstruction Outcome - Machine Learning Analysis of National Knee Ligament Registries

Aug 2, 2026

The dissertation from the Norwegian School of Sport Sciences examines how machine learning can predict outcomes after anterior cruciate ligament (ACL) reconstruction using national knee ligament registry data. It develops supervised prediction models across multiple studies, addressing patient selection, data preparation, missing data handling, model performance evaluation, and external validation using Danish and international datasets. A sixth study extends the approach with combined unsupervised learning to derive additional outcome-related patterns and evaluates model outputs to support clinically relevant risk prediction.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages196

Reading Time8 h 14 min

AccessFree to download

Summary

Introduction

Background

The Challenge with Anterior Cruciate Ligament Reconstruction Outcome Prediction

Machine Learning for Outcome Prediction

Machine Learning and National Registries to Predict ACL Reconstruction Outcome

Specific Aims of the Dissertation

Materials and Methods

Ethics

General Comments

Prediction Model Development (Papers I-III)

External Validation (Papers IV and V)

Unsupervised Learning (Paper VI)

Results

Paper I – Norwegian Revision Risk Prediction

Paper II – Norwegian Inferior Patient Reported Outcome Risk Prediction

Paper III – Combined Norwegian and Danish Revision Risk Prediction

Paper IV – External Validation Using the Danish Registry

Paper V – External Validation Using the STABILITY I Patients

Paper VI – Combined Norwegian and Danish Unsupervised Machine Learning Analysis

Discussion

Prediction Model Development (Papers I-III)

Main Findings

R . Ky le Martin // Pred icting Anterior Cruciate L igament Reconstruction Outcome DISSERTATION FROM THE NORWEGIAN SCHOOL OF SPORT SCIENCES 2025 R. Kyle Martin Predicting Anterior Cruciate Ligament Reconstruction Outcome Machine Learning Analysis of National Knee Ligament Registries R. Kyle Martin Predicting Anterior Cruciate Ligament Reconstruction Outcome Machine Learning Analysis of National Knee Ligament Registries DISSERTATION FROM THE NORWEGIAN SCHOOL OF SPORT SCIENCES • 2025 ISBN 978-82-502-0634-2 For my Nan, who instilled in me the importance of empathy and a good bedside manner. Table of Contents Acknowledgements ............................................................................................................. I List of Papers .................................................................................................................... III Abbreviations ..................................................................................................................... IV List of Figures..................................................................................................................... V List of Tables ................................................................................................................... VII Summary ......................................................................................................................... VIII Introduction ........................................................................................................................ 11 Background ............................................................................................................................................. 11 The Challenge with Anterior Cruciate Ligament Reconstruction Outcome Prediction ............ 12 Machine Learning for Outcome Prediction....................................................................................... 15 Machine Learning and National Registries to Predict ACL Reconstruction Outcome .............. 18 Specific Aims of the Dissertation .......................................................................................21 Materials and Methods ...................................................................................................... 22 Ethics ....................................................................................................................................................... 22 General Comments................................................................................................................................ 22 Prediction Model Development (Papers I-III) .................................................................................23 Patient Population ............................................................................................................................. 23 Data Preparation ................................................................................................................................ 24 Machine Learning Analysis .............................................................................................................. 26 Missing Data ....................................................................................................................................... 28 Model Performance Evaluation ...................................................................................................... 29 External Validation (Papers IV and V)............................................................................................... 30 Patient Population ............................................................................................................................. 30 Data Preparation ................................................................................................................................ 31 Missing Data ................................

FAQs

What problem does this dissertation address?

It focuses on predicting outcomes after anterior cruciate ligament (ACL) reconstruction and the challenge of accurately forecasting revision risk and inferior patient-reported outcomes.

How is machine learning used in the study?

Machine learning is applied to develop supervised prediction models using registry data, followed by performance evaluation and external validation on separate cohorts.

What additional approach is included beyond supervised prediction?

The dissertation includes an unsupervised learning study that combines Norwegian and Danish data to uncover outcome-related structure in the registry information.

Aug 2, 2026

Upload by Evangeline