Machine Learning Models in the Prediction of Adverse Outcomes in Peripheral Arterial Disease: Meta-Analysis.
Ann Vasc SurgMay 8, 2026 (epub)
Bharadhwaj Ravindhran, Laila Ubhi, Shahani Nazir et al.
✦ AI-curated · Sources linked
In 30 seconds
This meta-analysis evaluated the effectiveness of machine learning (ML) models versus traditional statistical methods in predicting adverse outcomes in patients with peripheral arterial disease (PAD). Thirteen studies were included, revealing that gradient-boosted models had the highest predictive performance with a diagnostic odds ratio (DOR) of 36.593. The study highlights the advantages of ML techniques in this context.
Key findings
- Gradient-boosted models achieved a DOR of 36.593 for predicting major adverse limb events.
- The least performing ML models still showed moderate predictive ability with a DOR of 8.616.
- Traditional regression models had a lowest DOR of 3.326, indicating inferior performance compared to ML approaches.
- 54% of studies included were rated as low risk of bias.
Why it matters
The findings suggest that machine learning models can significantly enhance the prediction of adverse outcomes in PAD patients, potentially leading to better risk stratification and management strategies in clinical practice.