JMIR Publications

ABSTRACT

Background:

Pharmaco-epidemiology studies using electronic health records (EHR) data typically rely on medication prescriptions to determine which patients have received a medication. However, such data do not affirmatively indicate whether these prescriptions have been filled. External dispensing databases can bridge this information gap; however, there are few established methods for linking EHR data and pharmacy dispensing records.

Objective:

We described a process for linking EHR prescribing data with pharmacy dispensing records from SureScripts. As a use case, we considered the prescription and resulting fills for psychotropic among pediatric patients and assessed whether the use of pharmacy dispensing data influenced inference regarding association between prescription receipt and completion of a follow-up appointment with the prescribing provider.

Methods:

This retrospective study identified all new psychotropic prescriptions to patients under 18 at Duke University Health System in 2021. We linked dispensing to prescribing data using proximate dates and matching codes between RxCUIs (in the EHR) and national drug codes (in SureScripts). We described demographic, clinical, and service utilization characteristics to assess differences between patients who did versus did not fill prescriptions. LASSO regression was applied to evaluate the predictability of fills. Time-to-event models assessed associations between prescription filling and follow-up visits with the prescriber.

Results:

We identified 1,254 pediatric patients with a new psychotropic prescription. There were 976 patients (78%) who filled their psychotropic prescriptions within 30 days of their prescribing encounters. As such, we set 30 days as a cut-point for defining a valid prescription fill. Some of the greatest differences between those who did and did not fill their prescription were seen based on sex (standardized mean difference (SMD) = 0.115), race/ethnicity (SMD = 0.102), primary payer type (SMD = 0.305) and prescribing provider specialty (SMD = 0.384). The LASSO model achieved an AUROC of 0.816. Time to follow-up visit with the same provider was censored at 90 days after the initial encounter. Patients who filled their prescriptions showed higher levels of follow-up visits. The marginal hazard ratio (HR) of a follow-up visit with the same provider was 1.673 (95% CI: 1.463 - 1.913) for patients who filled their prescriptions. Using the LASSO model as a propensity-based weight, we calculated the weighted HR of a follow-up visit as 1.447 (95% CI: 1.257 - 1.665).

Conclusions:

Systematic differences existed between patients who did versus did not fill prescriptions. Prescription fulfillment was associated with increased likelihood of a follow-up visit with the prescribing provider. Incorporating external dispensing databases into EHR-based studies can inform medication receipt and associated health outcomes.