A New Clustering Method for Knee Movement Impairments using Partitioning Around Medoids Model

Area of Science:

• Orthopedics
• Rehabilitation Science
• Biomedical Engineering

Background:

• The movement system impairment (MSI) model aids in classifying, diagnosing, and treating knee impairments.
• Partitioning around medoids (PAM) clustering can group patients into homogeneous clusters using discriminative variables.
• Simplifying the MSI model is crucial for efficient clinical application.

Purpose of the Study:

• To reduce the number of clinical examination variables for knee impairments.
• To identify the most important variables for classifying knee pain patients.
• To simplify the MSI model using PAM clustering.

Main Methods:

• A cross-sectional study involving 209 patients with knee pain in Shiraz, Iran.
• Assessment of knee, femoral, and tibial movement impairments and pain levels during various functional tasks.
• Application of PAM clustering analysis using pain patterns and impairment types.

Main Results:

• PAM clustering categorized patients into valgus and non-valgus groups.
• Further sub-clustering identified distinct patient subgroups based on hypomobility and other characteristics.
• Only 23 out of 41 variables were essential for sub-clustering, significantly reducing the data needed.

Conclusions:

• A simplified direct knee examination method was developed.
• The method organizes key discriminative tests, requiring fewer clinical variables.
• This approach enhances the efficiency of the MSI model for knee impairments.