Unsupervised machine-learning algorithms for the identification of clinical phenotypes in the osteoarthritis initiative database
Authors' conclusion
Two different clustering approaches yielded similar patient allocations primarily separating complex "comorbid" patients from healthier subjects, the latter divided in young/physically active vs elderly/physically inactive subjects.
Study context
- Population
- Progression and incident cohorts of the Osteoarthritis Initiative (OAI) dataset
- Outcome
- Identification of patient clusters (clinical phenotypes) using two machine learning algorithms
Methodological detailsunsupervised machine-learning cluster analysis of the Osteoarthritis Initiative database, deep embedded clustering (DEC), multiple factor analysis with clustering (MFAC), 157 input-variables at baseline
- Unsupervised machine-learning cluster analysis of the Osteoarthritis Initiative database
- Deep embedded clustering (DEC)
- Multiple factor analysis with clustering (MFAC)
- 157 input-variables at baseline
Main finding
As reported in the abstract
Patient phenotypes identified
DEC phenotypes
5
MFAC phenotypes
3
Observational data: associations, not effects.
Further findings
Pain and physical activity trajectories · 9 years
- Overall pain trajectories stable; only numeric rating scale (NRS) pain showed distinct increase; physical activity decreased in all clusters
Joint space change trajectories · 8 years
- Clusters showed different (though non-significant) trajectories
Original publication
Unsupervised machine-learning algorithms for the identification of clinical phenotypes in the osteoarthritis initiative database. Semin Arthritis Rheum. Published 19 November 2022. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →