JAMA Intern Med. 2023 Sep 5. doi: 10.1001/jamainternmed.2023.4192. Online ahead of print.
IMPORTANCE: An estimated 1.5% to nearly 5% of medications are dispensed after discontinuation in the electronic health record (EHR), with 34% meeting criteria for high risk of potential harm.
OBJECTIVE: To evaluate the association of the implementation of e-prescription cancellation messaging (CancelRx) with medication dispensing after discontinuation of e-prescriptions in the EHR.
DESIGN, SETTING, AND PARTICIPANTS: This case series with interrupted time series analysis included patients who had at least 1 medication e-prescribed in ambulatory care to a health system pharmacy and discontinued in the 2-year study period from 1 year prior to approximately 1 year after CancelRx implementation (January 15, 2018, to December 7, 2019). Prior to CancelRx implementation, changes to e-prescribed medications within the EHR were not electronically communicated to health system pharmacies, which used separate pharmacy management software. Statistical analysis was performed from November 2020 to June 2023 (primary analysis from March 2021 to May 2022).
EXPOSURE: Implementation of CancelRx.
MAIN OUTCOMES AND MEASURES: The primary outcome was the proportion of e-prescribed medications dispensed and sold to patients by pharmacies within 6 months after discontinuation in the EHR. A medication was defined as dispensed after discontinuation if the timestamp of dispensing was at least 1 minute and less than 6 months after the timestamp of discontinuation in the EHR. A secondary outcome was the proportion of discontinued medications that was reordered within 120 days.
RESULTS: A total of 53 298 qualifying e-prescriptions that were discontinued were identified for 17 451 unique patients (mean [SD] age, 50.6 [18.2] years; 9332 women [53.5%]). After CancelRx implementation, 22 443 (85.9%) of the 26 127 discontinued e-prescriptions resulted in a CancelRx transaction. In interrupted time series analysis, the proportion of prescriptions dispensed after discontinuation decreased from a baseline of 8.0% (2162 of 27 171) to 1.4% (369 of 26 127; P < .001), without a significant week-to-week trend (β = 0.000158; P = .37).
CONCLUSIONS AND RELEVANCE: In this case series with interrupted time series analysis, findings suggest that CancelRx implementation was associated with an immediate and persistent reduction in the proportion of e-prescriptions sold after discontinuation in the EHR. Widespread implementation of CancelRx may significantly improve medication safety through the reduction of medication dispensing after discontinuation by prescribers.