Decision Support for Managing Common Musculoskeletal Pain Disorders: Development of a Case-Based Reasoning...

Area of Science:

• Artificial Intelligence in Healthcare
• Musculoskeletal Pain Management
• Physiotherapy Decision Support

Background:

• Musculoskeletal pain disorders often lack effective, evidence-based treatments due to patient heterogeneity.
• Existing treatment guidelines have limited applicability in clinical practice for diverse patient populations.
• Artificial intelligence, specifically case-based reasoning (CBR), offers a potential solution by leveraging past patient experiences.

Purpose of the Study:

• To develop a CBR-based decision support system, SupportPrim PT, for physiotherapy care in primary care settings.
• To demonstrate the system's capability in identifying similar patients with musculoskeletal pain disorders.
• To enhance personalized treatment approaches for musculoskeletal conditions.

Main Methods:

• Collected data from Norwegian primary care physiotherapy patients to build a case base for SupportPrim PT.
• Employed the local-global principle in CBR, utilizing weighted prognostic attributes for global similarity and clinically important differences for local similarity.
• Assessed system performance by comparing similarity scores against the Örebro Musculoskeletal Pain Screening Questionnaire (ÖMSPQ) and the Musculoskeletal Health Questionnaire (MSK-HQ).

Main Results:

• The SupportPrim PT system was developed with 29 weighted attributes and local similarities, using an initial case base of 105 patients.
• The system identified similar patients with mean absolute differences of 9.3 points on the ÖMSPQ and 5.6 points on the MSK-HQ.
• A larger case base (N=486) resulted in higher mean similarity scores, indicating slightly improved patient similarity matching.

Conclusions:

• Successfully developed SupportPrim PT, a CBR system for managing musculoskeletal pain in primary care physiotherapy.
• The system effectively identified similar patients based on established screening tools and outcome measures.
• SupportPrim PT shows promise for improving decision-making and potentially personalizing care for patients with musculoskeletal pain.