Drug grouping learning for improving evidence-based treatment recommendations.

Area of Science:

• Pharmacology
• Clinical Practice Guidelines
• Data Science

Background:

• Clinical practice guidelines (CPGs) translate scientific evidence into clinical recommendations.
• Formulating clinical questions, especially for pharmacological interventions, is a critical first step in CPG development.
• The variability in evidence quantity and quality for some interventions impacts treatment recommendations.

Purpose of the Study:

• To present a novel method for formulating clinical questions involving pharmacological interventions.
• To group drugs based on shared characteristics, considering both patient/clinician desired outcomes and pharmacological features.
• To demonstrate the feasibility of this approach in the context of Attention Deficit Hyperactivity Disorder (ADHD).

Main Methods:

• Developed a new method to learn personalized distances among drugs, incorporating user preferences.
• Designed an ensemble clustering method to identify optimal drug groupings for specific clinical queries.
• Applied and validated the approach using data related to ADHD treatment.

Main Results:

• The proposed method successfully groups drugs based on desired outcomes and pharmacological properties.
• Personalized drug distances and ensemble clustering effectively identified suitable drug groupings.
• The approach demonstrated feasibility in the context of ADHD treatment.

Conclusions:

• The presented method offers a robust approach to formulating clinical questions for pharmacological interventions.
• Personalized drug grouping can enhance the development of evidence-based clinical practice guidelines.
• This methodology has the potential to improve treatment recommendations by considering patient and clinician preferences.