AI-Based Automation for Medication Reconciliation: Scoping Review

Area of Science:

• Health Informatics
• Artificial Intelligence in Healthcare
• Patient Safety

Background:

• Medication reconciliation (MedRec) is crucial for patient safety, ensuring accurate medication information across care transitions.
• MedRec involves creating medication histories, identifying discrepancies, and resolving them.
• Artificial intelligence (AI) offers potential to improve MedRec efficiency and accuracy.

Purpose of the Study:

• To systematically review and map the application of AI in MedRec tasks and subtasks.
• To assess the level of automation achieved by AI in MedRec processes.
• To identify research gaps and future directions for AI in MedRec.

Main Methods:

• A comprehensive scoping review was conducted, searching major databases (MEDLINE, Embase, Web of Science, IEEE Xplore, Compendex).
• Studies were screened for AI application in MedRec tasks, excluding rule-based systems.
• A 4-stage human information processing model guided the assessment of automation levels.

Main Results:

• 94 studies met inclusion criteria, all addressing medication history creation.
• Only 2.1% of studies addressed discrepancy identification, with limited automation beyond information acquisition.
• Most studies utilized electronic health record text data and machine learning models, with a focus on model development using public datasets.

Conclusions:

• Current AI applications in MedRec are preliminary, primarily focused on information extraction.
• Significant gaps exist in automating discrepancy identification and resolution.
• Future efforts should address data challenges, implement AI models in practice, and evaluate real-world usability.