Machine learning models predicting risk of revision or secondary knee injury after anterior cruciate ligament reconstruction demonstrate variable discriminatory and accuracy performance: a systematic review
2025

Machine Learning Models for Predicting Knee Injury After ACL Surgery

Sample size: 125427 publication Evidence: moderate

Author Information

Author(s): Blackman Benjamin, Vivekanantha Prushoth, Mughal Rafay, Pareek Ayoosh, Bozzo Anthony, Samuelsson Kristian, de SA Darren

Hypothesis

Machine learning models would be superior in predicting outcomes compared to standard logistic regression models.

Conclusion

Machine learning models designed to predict the risk of revision or secondary knee injury demonstrate variable discriminatory performance.

Supporting Evidence

  • Nine studies comprising 125,427 patients were included in this review.
  • Machine learning models showed variable performance in predicting knee injury outcomes.
  • Some models achieved high accuracy, while others demonstrated significant miscalibration.

Takeaway

This study looked at how well computer programs can predict problems after knee surgery. Some programs did a good job, but others didn't work as well.

Methodology

A systematic review of nine studies using machine learning to predict outcomes after ACL reconstruction, analyzing data from three databases.

Potential Biases

Variability in study quality and adherence to reporting guidelines may introduce bias.

Limitations

The lack of external validation of existing models limits their generalizability.

Participant Demographics

The review included 125,427 patients with a mean age of 26.73 years, and 41.5% were female.

Digital Object Identifier (DOI)

10.1186/s12891-024-08228-w

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication