Linking Electronic Health Record Prescribing Data and Pharmacy Dispensing Records to Identify Patient-Level Factors...

Area of Science:

• Pharmacoepidemiology
• Health Informatics
• Pediatric Health

Background:

• Electronic health record (EHR) data often track prescriptions but not fills, limiting pharmacoepidemiology.
• External dispensing databases can bridge this gap, but linking methods are scarce.
• Accurate medication receipt data is crucial for understanding treatment adherence and outcomes.

Purpose of the Study:

• To describe a novel process for linking EHR prescribing data with pharmacy dispensing records.
• To evaluate the impact of dispensing information on identifying patients receiving psychotropic medications.
• To assess the association between prescription fills and subsequent pediatric health behaviors.

Main Methods:

• Retrospective linkage of new psychotropic prescriptions in pediatric patients (Duke University Health System, 2021) with Surescripts dispensing data.
• Utilized RxNorm and National Drug Codes for matching, with a 30-day window for fill definition.
• Employed LASSO regression for fill predictability and time-to-event models for health behavior associations.

Main Results:

• Of 1254 pediatric patients, 77.8% filled psychotropic prescriptions within 30 days.
• Patients who did not fill prescriptions exhibited distinct characteristics (e.g., higher BMI, neighborhood disadvantage, public insurance).
• Prescription fills were associated with increased follow-up provider visits (weighted hazard ratio: 1.447).

Conclusions:

• Significant demographic and clinical differences exist between patients who fill and do not fill prescriptions.
• Integrating external dispensing data into EHR-based studies provides a more accurate picture of medication receipt.
• This enhanced data linkage is vital for robust pharmacoepidemiological research and understanding treatment impact.