A smart medication recommendation model for the electronic prescription

Area of Science:

• Health Informatics
• Medical Data Mining
• Clinical Decision Support

Background:

• Preventable medical errors and patient safety are critical concerns in healthcare.
• E-prescribing (electronic prescribing) was mandated by federal acts to improve healthcare quality and efficiency.
• Inappropriate prescribing contributes to a significant percentage of drug-related adverse events.

Purpose of the Study:

• To enhance e-prescribing system efficiency by reducing medication list length.
• To minimize the risk of inappropriate medication selection by physicians.
• To decrease the overall time physicians spend on prescribing.

Main Methods:

• Utilized 103.48 million prescriptions from Taiwan's national health insurance claims data.
• Computed Diagnosis-Medication associations from the large dataset.
• Developed a smart medication recommendation model using data mining association rules on 100,000 selected prescriptions.

Main Results:

• Introduced novel concepts: Mean Prescription Rank (MPR) and Coverage Rate (CR).
• Computed a proactive medication list (PML) based on MPR and CR.
• Significantly shortened medication drop-down menus in the e-prescribing system.

Conclusions:

• The developed model effectively reduces medication selection errors and prescription times.
• Physicians can still select appropriate medications, even with unintentional errors.
• The system enhances patient safety by improving the accuracy and efficiency of electronic prescribing.