Prediction Models for Knee Osteoarthritis: Review of Current Models and Future Directions

Document Type : SYSTEMATIC REVIEW

Authors

1 1 Department of Health Sciences Research, Mayo Clinic, 200 First St SW Rochester, Rochester, Minnesota, USA 2 Department of Orthopedics, Mayo Clinic, 200 First St SW Rochester, Rochester, Minnesota, USA

2 Department of Health Sciences Research, Mayo Clinic, 200 First St SW Rochester, Rochester, Minnesota, USA

3 Department of Orthopedics, Mayo Clinic, 200 First St SW Rochester, Rochester, Minnesota, USA

Abstract

Background: Knee osteoarthritis (OA) is a prevalent joint disease. Clinical prediction models consider a wide range 
of risk factors for knee OA. This review aimed to evaluate published prediction models for knee OA and identify 
opportunities for future model development.
Methods: We searched Scopus, PubMed, and Google Scholar using the terms knee osteoarthritis, prediction model, 
deep learning, and machine learning. All the identified articles were reviewed by one of the researchers and we recorded 
information on methodological characteristics and findings. We only included articles that were published after 2000 
and reported a knee OA incidence or progression prediction model.
Results: We identified 26 models of which 16 employed traditional regression-based models and 10 machine learning 
(ML) models. Four traditional and five ML models relied on data from the Osteoarthritis Initiative. There was significant 
variation in the number and type of risk factors. The median sample size for traditional and ML models was 780 
and 295, respectively. The reported Area Under the Curve (AUC) ranged between 0.6 and 1.0. Regarding external 
validation, 6 of the 16 traditional models and only 1 of the 10 ML models validated their results in an external data set. 
Conclusion: Diverse use of knee OA risk factors, small, non-representative cohorts, and use of magnetic resonance 
imaging which is not a routine evaluation tool of knee OA in daily clinical practice are some of the main limitations of 
current knee OA prediction models. 
Level of evidence: III

Keywords

Main Subjects


Volume 11, Issue 1
January 2023
Pages 1-10
  • Receive Date: 13 September 2021
  • Revise Date: 17 February 2022
  • Accept Date: 23 February 2022
  • First Publish Date: 28 December 2022