JMIR Publications

ABSTRACT

Background:

Medication errors are pervasive. One way to avert them is to limit the transcribing of prescription information. Electronic prescriptions (e-prescriptions) convey secure and computer-readable prescriptions from clinics to outpatient pharmacies for dispensing. After transmission, pharmacy staff perform a transcription task where they select the medications needed to fulfill e-prescriptions within their dispensing software and then verify that their medication selections are correct. Later, pharmacists manually double-check medications selected to fulfill e-prescriptions before dispensing. While pharmacist double-checks are mostly effective for catching medication selection mistakes, the cognitive process of doing medication selection is still prone to error due to heavy workload, inattention, and fatigue. Leveraging health information technology to improve medication selection accuracy during transcription in outpatient pharmacies supports the larger goal of making the United States health care system safer.

Objective:

The objective of this study is to determine the performance of an automated double-check that uses the RxNorm Application Programming Interface (API) and attempts to identify medication selection errors made in outpatient pharmacies.

Methods:

We conducted a retrospective analysis of 537,710 pairs of e-prescription and dispensing records from a mail-order pharmacy for the period 01/2017-10/2018. National drug codes (NDC) for each pair were submitted to the National Library of Medicine’s (NLM) RxNorm API and the API returned RxCUI semantic clinical drug/generic pack (SCD) identifiers associated with every NDC. The SCD identifiers returned for e-prescription were matched against the corresponding SCD identifiers from the pharmacy dispensing record. An error matrix was created based on hand-labeling mismatched SCD pairs. Performance metrics, including sensitivity, specificity, positive predictive value, false-positive rate, precision, and F1 score, were calculated for the e-prescription-to-dispensing record matching algorithm for both total pairs and unique pairs of NDCs in these data.

Results:

We analyzed 527,881 e-prescription and pharmacy dispensing record pairs. Four clinically significant cases of mismatched RxCUI identifiers were detected (i.e., three incorrect medication selections and one incorrect strength selection). Five-hundred forty six less significant cases of mismatched RxCUIs were found. Nearly all of the NDC pairs had matching RxCUIs (99.896% - 99.688%). The RxNorm API had a sensitivity of 1, a false-positive rate of 0.00104 to 0.00312, specificity of 0.99896 to 0.99688, precision of 0.00727 to 0.04255, and F1 score of 0.01444 to 0.08163. We found 872 pairs of records without an RxCUI.

Conclusions:

The NLM’s RxNorm API can perform an independent and automatic double-check of correct medication selection during e-prescription verification at outpatient pharmacies. RxNorm has near-comprehensive coverage of prescribed medications and can be a tool to prevent medication selection errors. In the future, tools like this may be able to perform automated verification of medication selection accurately enough to free pharmacists from having to perform manual double-checks of the medications selected within pharmacy dispensing software to fulfill e-prescriptions.