Optimizing methadone dose adjustment in patients with opioid use disorder

Area of Science:

• Pharmacology
• Data Science
• Clinical Medicine

Background:

• Opioid use disorder (OUD) presents a significant global health challenge.
• Methadone maintenance therapy is a cornerstone treatment for OUD.
• Optimizing methadone dosage is crucial for treatment efficacy and patient adherence.

Purpose of the Study:

• To develop and validate a model for optimizing methadone dose adjustments in patients with OUD.
• To integrate clinical data and patient factors for personalized methadone dosing.
• To leverage mixed modeling and machine learning for enhanced treatment management.

Main Methods:

• A retrospective study analyzed 40,530 dosing records and 1,508 urine tests from 96 OUD patients.
• Stage 1: Mixed modeling assessed factors influencing treatment discontinuation.
• Stage 2: Machine learning built a model for optimized methadone dosage.

Main Results:

• Reduced methadone doses were associated with a higher likelihood of treatment discontinuation.
• The developed model demonstrated a high correlation (0.995 ± 0.003) with physician-determined doses.
• The model's suggested doses were within an acceptable range compared to clinical practice.

Conclusions:

• A novel model for methadone dose adjustment in OUD patients was successfully developed.
• The model integrates urine opiate levels, adherence, and discontinuation likelihood for automatic dose suggestions.
• This approach is particularly valuable for remote patient management and when in-person consultations are limited.