Adopting Text Mining on Rehabilitation Therapy Repositioning for Stroke

Area of Science:

• Neurorehabilitation
• Medical Informatics
• Biomedical Engineering

Background:

• Stroke significantly impacts daily life, necessitating effective rehabilitation strategies.
• Current rehabilitation guidelines lack specificity, and innovation in techniques is slow.
• Identifying novel or repositioned therapies for stroke is crucial for improving patient outcomes.

Purpose of the Study:

• To identify a candidate rehabilitation therapy for stroke repositioning using a text mining approach.
• To evaluate the clinical effectiveness of the identified therapy for upper limb dysfunction in stroke patients.

Main Methods:

• Utilized the ABC model (Stroke, Assessment Scales, Rehabilitation Therapies) with PubMed data.
• Compiled datasets of stroke-related scales and therapies.
• Crawled co-occurring therapies and filtered for potential repositioning candidates.
• Conducted clinical validation of the top-ranked therapy.

Main Results:

• Identified Hand-arm bimanual intensive training (HABIT) as the top repositioning candidate.
• HABIT demonstrated significant improvements in Fugl-Meyer Assessment (FMA) and Action Research Arm Test (ARAT) scores.
• The ABC model proved effective for rehabilitation therapy repositioning.

Conclusions:

• Hand-arm bimanual intensive training (HABIT) is a promising repositioned therapy for stroke patients with upper limb dysfunction.
• The ABC text mining model is an effective tool for discovering and validating rehabilitation therapies.
• Findings support clinical knowledge discovery and improved stroke rehabilitation strategies.