Key Points
Question
Was implementation of e-prescription cancellation messaging (CancelRx) associated with a decrease in medication dispensing after discontinuation of e-prescriptions in the electronic health record?
Findings
This case series with interrupted time series analysis found that, after CancelRx implementation, the proportion of prescriptions dispensed after discontinuation decreased to 1.4% from a baseline of 8.0%, without a significant week-to-week trend.
Meaning
This study suggests that CancelRx implementation was associated with an immediate and persistent reduction in the proportion of e-prescriptions sold after discontinuation in the electronic health record.
Abstract
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.
This case series with interrupted time series analysis evaluates the association of the implementation of e-prescription cancellation messaging (CancelRx) with the discontinuation of medication dispensing after e-prescription in the electronic health record.
Introduction
Preventable harm due to medication errors is estimated to affect 1 in 30 people, with severe or life-threatening harm in 25% of cases.1 An estimated 1.5% to nearly 5% of medications are dispensed after discontinuation in the electronic health record (EHR),2,3 with 34% meeting criteria for high risk of potential harm.2 Dispensing of discontinued medications leads to patient harm through the unintended continuation of medications4,5 or the inadequate management of disease with incorrect medication dosing.5
Dispensing of discontinued medications may result from a lack of communication of discontinuation from prescribers to pharmacies. In 2006, the National Council for Prescription Drug Programs (NCPDP) added CancelRx to the SCRIPT standard for e-prescribing.6 When implemented, CancelRx enables EHRs to send electronic cancellation messages for e-prescriptions through the health information network to a pharmacy, in a process similar to e-prescribing. However, adoption of CancelRx has been slower than adoption of e-prescribing, and this functionality remains underused.6
In previous studies within academic health systems, more than 90% of CancelRx transactions resulted in a successful cancellation.7,8 Implementation of CancelRx was associated with reduced discrepancies between health system and pharmacy medication lists,8 but the association of CancelRx with medication dispensing has only been measured in a pilot study.9 The objective of this study was to evaluate the association of CancelRx implementation with medication dispensing after discontinuation of e-prescriptions in the EHR.
Methods
Intervention and Setting
On January 15, 2019, CancelRx was implemented for prescriptions sent from ambulatory practices to pharmacies associated with an urban academic medical center. The health system ambulatory practices used a single EHR (Epic; Epic Systems Corp), while the 11 health system pharmacies used a separate pharmacy management software (EnterpriseRx; McKesson). As a result, changes to e-prescribed medications within the EHR were not electronically communicated to the health system pharmacies prior to CancelRx implementation. Prior to widespread implementation of CancelRx, a proactive risk assessment and pilot implementation were conducted,7 which informed implementation decisions. These results and CancelRx configuration options were then presented to the multidisciplinary ambulatory design committee, followed by the joint design committee, which oversees both ambulatory and inpatient design committees. When CancelRx was implemented, the EHR was configured to send a CancelRx message when an e-prescription was discontinued unless the reason for discontinuation specified that the prescription was being renewed or was a duplicate. At the time of this analysis, we did not suppress CancelRx messages based on order class or medication expiration status. Based on the prior estimates of the large volume of inbasket notifications, most of which would not be actionable, these were suppressed during implementation.7 The warning to contact the pharmacy was always displayed if an external pharmacy did not accept CancelRx messages; however, all prescriptions in this analysis were e-prescribed to internal health system pharmacies that accepted CancelRx messages. This project was reviewed and approved by the Johns Hopkins School of Medicine institutional review board with a waiver of consent granted as it was not feasible to obtain consent from this number of participants, it was a minimal risk study with use of data from clinical care, and used a limited data set. This study followed the reporting guideline for case series.
Study Design
We conducted an interrupted time series analysis of the association of CancelRx with medication dispensing after e-prescription discontinuation in the EHR. In addition, we conducted pre-post analyses to examine the association of CancelRx implementation with medication reorders after discontinuation and variation in dispensing after discontinuation by pharmacy and medication pharmaceutical class.
Study Population and Data Sources
Patients were included if they had at least 1 medication e-prescribed in ambulatory care to a health system pharmacy and discontinued during the 2-year study period from 1 year prior to approximately 1 year after CancelRx implementation (January 15, 2018, to December 7, 2019). We excluded e-prescriptions for over-the-counter medications and those with topical, otic, and ophthalmic administration.
A trained pharmacy data analyst (T.W.) exported EHR data on discontinued medications that had been e-prescribed to a health system pharmacy and matched these to dispensing data from the pharmacy management software. e-Prescriptions were matched on the e-prescription message order identification number, order date, and medication name, which were present in both data sets. Additional data elements from the EHR included patient self-reported race and sex, prescriber, discontinuing user, reason for discontinuation, and discontinuation date. Additional data elements from the pharmacy management software included the dispensing pharmacy, the date and time when the medication was last sold or, if never sold, the last date and time when the prescription was in process to be filled. Pharmacy data included dispensing through July 15, 2021, to allow 6 months of follow-up for all discontinued e-prescriptions.
We excluded reordered prescriptions, which were identified by a discontinuation reason indicating that the prescription was reordered (which is automatically added if the reorder function is used) or by meeting all of the following criteria: a duplicate prescription was placed within the hour before or after the date or time of discontinuation to the same pharmacy and the discontinuation reason was not a dose or form adjustment.
Outcomes
Our primary outcome was the proportion of e-prescribed medications dispensed and sold to patients (subsequently referred to as dispensed) by a health system pharmacy within 6 months after discontinuation in the EHR. We defined a medication 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 as a balancing measure.
Statistical Analysis
Statistical analysis was performed from November 2020 to June 2013 (primary analysis from March 2021 to May 2022). We categorized medications based on the pharmaceutical class from First Databank associated with the medication in the EHR and the dispensing pharmacy. Characteristics of patient, encounter, pharmacy, and therapeutic class were examined using frequencies and proportions for categorical variables and median values and IQRs for continuous variables. Variables were compared prior to and after CancelRx implementation using χ2 tests and mean values with 95% CIs.
For each calendar week, the proportion of discontinued e-prescriptions that was filled and sold within the 6 months after discontinuation in the EHR was tabulated. Calendar weeks with partial data (weeks containing January 15, 2018, and January 15, 2020) and the week of implementation (week containing January 15, 2019) were excluded. After the original study design, changes in CancelRx implementation occurred in week 98 and temporarily limited the number of CancelRx transactions sent. As a result, we excluded e-prescriptions discontinued during weeks 98 to 102, providing 51 weeks of data prior to CancelRx implementation and 46 weeks of data after implementation. Weekly data over the 97 weeks of the study period were analyzed using interrupted time series modeling.
In the interrupted time series analysis, we specified a priori that the primary outcome of CancelRx would be an immediate reduction in dispensing after discontinuation (ie, a change in level without lag). However, we first visualized the data with a scatterplot and nonparametric data analysis using locally weighted scatterplot smoothing to describe the trajectory of our outcome over time to confirm this assumption. We assessed the correlation between time points using visual and statistical testing of residual values to identify autocorrelation and nonstationarity. We assessed for autocorrelation up to lag 51 using the Durbin-Watson statistic, which indicated higher-order autocorrelation. As a result, we used stepwise autoregression to select the order of the autoregression error model. Autoregressive parameters were sequentially removed until the remaining autoregressive parameters had statistically significant t tests, retaining a single parameter (autoregressive parameter for lag 37). We tested for nonstationarity using the Dikey-Fuller test for unit root (P = .001, rejecting the null hypothesis that the mean is not stationary). All analyses were conducted using SAS, version 9.4 (SAS Institute Inc). All P values were from 2-sided tests, and results were deemed statistically significant at P < .05.
Results
We identified 53 298 qualifying e-prescriptions among 17 451 patients (mean [SD] age, 50.6 [18.2] years; 9332 women [53.5%]) written by 3986 unique prescribers (Table). The median time from e-prescribing to discontinuation was 55 days (IQR, 17-141 days). After CancelRx implementation, 22 443 (85.9%) of the 26 127 discontinued e-prescriptions resulted in a CancelRx transaction. Among discontinued e-prescriptions in which a CancelRx was not sent, 1634 (6.3%) were not expected to generate a CancelRx because they were discontinued as a duplicate prescription, leaving 2050 (7.8%) for which a reason was not identified.
Table. Characteristics of Discontinued e-Prescriptions, Associated Patient Characteristics, and Dispensing Outcomes Before and After CancelRx Implementation.
| Characteristic | Before CancelRx (n = 27 171 [51.0%])a | After CancelRx (n = 26 127 [49.0%])b | Total (N = 53 298 [100%])c |
|---|---|---|---|
| Unique prescribers (discontinue order), No. | 2893 | 2987 | 3986 |
| Time between original prescription and discontinuation, median (IQR), d | 56 (19-140) | 54 (16-145) | 55 (17-141) |
| Dispensing outcomes | |||
| Discontinued e-prescriptions with CancelRx, No. (%) | 0 | 22 443 (85.9) | 22 443 (42.1) |
| Discontinued e-prescriptions filled or sold within 6 mo, No. (%) | 2162 (8.0) | 369 (1.4) | 2531 (4.8) |
| Discontinued e-prescriptions reordered within 120 d, No. (%) | 2830 (10.4) | 2820 (10.8) | 5650 (10.6) |
| Patient characteristics, No./total No. (%) | |||
| Sexd | |||
| Female | 5532/10 533 (52.5) | 5848/10 783 (54.2) | 9332/17 451 (53.5) |
| Male | 5001/10 533 (47.5) | 4934/10 783 (45.8) | 8118/17 451 (46.5) |
| Race | |||
| Asian | 446/10 533 (4.2) | 452/10 783 (4.2) | 757/17 451 (4.3) |
| Black | 4789/10 533 (45.5) | 4808/10 783 (44.6) | 7648/17 451 (43.8) |
| White | 4441/10 533 (42.2) | 4579/10 783 (42.5) | 7510/17 451 (43.0) |
| Othere | 857/10 533 (8.1) | 944/10 783 (8.8) | 1536/17 451 (8.8) |
| Discontinuation reason | |||
| Blank | 11 607 (42.7) | 9192 (35.2) | 20 799 (39.0) |
| Hospital discharge | 4607 (17.0) | 4853 (18.6) | 9460 (17.8) |
| Completed | 3253 (12.0) | 4722 (18.1) | 7975 (15.0) |
| Dose or form adjustment | 2103 (7.7) | 2095 (8.0) | 4198 (7.9) |
| Alternate therapy | 1023 (3.8) | 2114 (8.1) | 3137 (5.9) |
| Duplicate | 1040 (3.8) | 1634 (6.3) | 2674 (5.0) |
| Error | 1367 (5.0) | 944 (3.6) | 2311 (4.3) |
| Unknown | 1255 (4.6) | 0 | 1255 (2.4) |
| Other | 622 (2.3) | 390 (1.5) | 1012 (1.9) |
| ADE or contraindicated | 294 (1.1) | 183 (0.7) | 477 (0.9) |
| Patient encounter type at discontinuation | |||
| Outpatient prescriber visit | 13 219 (48.7) | 13 320 (51.0) | 26 539 (49.8) |
| Hospital encounter | 9104 (33.5) | 7737 (29.6) | 16 841 (31.6) |
| Orders only | 2421 (8.9) | 3010 (11.5) | 5431 (10.2) |
| Telephone | 1297 (4.8) | 835 (3.2) | 2132 (4.0) |
| Refill | 718 (2.6) | 771 (3.0) | 1489 (2.8) |
| Other | 412 (1.5) | 454 (1.7) | 866 (1.6) |
Abbreviation: ADE, adverse drug event.
Before CancelRx includes discontinued orders from January 21, 2018, to January 12, 2019.
After CancelRx includes discontinued orders from January 20 to December 7, 2019.
All discontinued orders between January 21, 2018, and December 7, 2019, excluding week of CancelRx implementation (January 13 to 19, 2019).
Data missing for 1 patient in the post-CancelRx period, which is why the numbers of male and female patients do not sum to the total number of unique patients in the post-CancelRx period or overall.
Other includes American Indian or Alaska Native, Native Hawaiian or Other Pacific Islander, other race, and unknown.
In the interrupted time series analysis, prior to CancelRx implementation, there was no significant week-to-week trend within the proportion of discontinued orders that was dispensed after discontinuation (β = −0.000203; P = .07). After CancelRx implementation, there was a significant reduction in the proportion of prescriptions dispensed after discontinuation 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) (Figure 1). The proportion of discontinued medications that was reordered within 120 days was not significantly different between the pre- and post-CancelRx implementation periods (10.4% [2820 of 27 171] vs 10.8% [2830 of 26 127]; P = .15).
Figure 1. Interrupted Time Series Analysis of the Proportion of Discontinued Medications Dispensed in 6 Months Before and After CancelRx Implementation.
The dotted vertical line at week 52 indicates CancelRx implementation. The dashed horizontal lines indicate regression lines depicting the trend in the proportion of discontinued medications dispensed and sold before CancelRx and after CancelRx.
Prior to CancelRx implementation, there was wide variation by pharmacy and medication class in the proportion of medications dispensed after discontinuation (Figure 2 and Figure 3). Across pharmaceutical classes, prior to CancelRx implementation, the mean (SD) proportion of medications dispensed after discontinuation ranged from 3.9% (12.1%) to 12.5% (5.5%). Three pharmaceutical classes had rates of dispensing after discontinuation that exceeded the mean, including immunosuppressants, anticoagulants and antiplatelet agents, and cardiovascular medications. Across the 11 included pharmacies, prior to CancelRx, the mean (SD) proportion of medications dispensed after discontinuation ranged from 4.5% (7.1%) to 12.4% (4.3%), and 2 pharmacies had rates of dispensing after discontinuation that exceeded the mean.
Figure 2. Mean Proportion of Discontinued Orders Dispensed in 6 Months by Medication Class.
Error bars indicate the 95% CIs. The horizontal dotted lines reflect the grand mean before CancelRx and the grand mean after CancelRx. The dark blue circles indicate before CancelRx, and the orange circles indicate after CancelRx. CNS indicates central nervous system.
Figure 3. Mean Proportion of Discontinued Orders Dispensed in 6 Months by Pharmacy.
Error bars indicate the 95% CIs. The horizontal dotted lines reflect the grand mean before CancelRx and the grand mean after CancelRx. The dark blue circles indicate before CancelRx, and the orange circles indicate after CancelRx.
aNo reliable estimate.
After CancelRx implementation, variation in the mean (SD) proportion of medications dispensed after discontinuation was reduced, ranging from 0.7% (1.9%) to 3.7% (4.6%) across pharmaceutical classes excluding dermatologic medications, which had variable estimates because of a small sample size (110 during the study), with a mean (SD) proportion of medications dispensed after discontinuation of 0.4% (1.2%) to 2.2% (2.1%) across pharmacies; the proportion of immunosuppressants and anticoagulants or antiplatelet medications dispensed after discontinuation was still above the mean for all medications.
Discussion
After CancelRx implementation, there was an immediate and persistent reduction in the proportion of e-prescriptions that was dispensed after discontinuation in the EHR from a baseline of 8.0% to 1.4%. Implementation of CancelRx also reduced variation by pharmacy and by medication class in the proportion of medications dispensed after discontinuation in the EHR, with the greatest reductions among immunosuppressants, anticoagulants and antiplatelet drugs, and cardiovascular medications. There was no change in the proportion of medications reordered in the 120 days after discontinuation.
This significant reduction in dispensing of prescriptions after discontinuation in the EHR suggests that CancelRx implementation may reduce medication dispensing errors by communicating intended medication changes to pharmacies. A prior study by Watterson et al8 demonstrated that CancelRx implementation reduced discrepancies between EHR and pharmacy management systems but did not evaluate the association with pharmacy dispensing. This study builds on the study by Watterson et al8 and a pilot study9 to demonstrate a reduction in dispensing of discontinued prescriptions.
However, CancelRx was not associated with the complete elimination of dispensing after discontinuation in the EHR. Most of the continued dispensing was likely the result of the EHR not sending a CancelRx transaction. Our health system implementation of CancelRx did not send a cancellation when a medication was discontinued as a duplicate therapy due to a perceived lower risk of harm with dispensing of a duplicate prescription and concern that the active prescription might erroneously be discontinued in the EHR (ie, leaving either an expired prescription or patient-reported medication on the medication list), resulting in no prescription at the pharmacy. We could not identify a reason for the absence of a CancelRx in 7.8% of discontinued e-prescriptions. In our EHR, the ability to send a CancelRx is tied to the permissions of the authorizing prescriber rather than discontinuing prescriber. Based on performance in the pilot study,9 if at the time of discontinuation the authorizing prescriber did not have e-prescribing permissions (ie, was removed from the system, such as a former resident), a CancelRx would not be sent.7 Further evaluation is needed to determine if this accounts for the absence of CancelRx transactions.
Pilot implementation previously identified a gap that allowed a prescription to be dispensed after a CancelRx transaction.7 In the health system’s pharmacy management software, once the CancelRx is matched to a prescription, the pharmacy cannot sell the prescription. If the CancelRx did not automatically match to a prescription, however, it could continue to be processed and sold until it was manually matched. To limit this vulnerability, pharmacy staff closely monitored the queue of CancelRx transactions. CancelRx messages may also not result in cancellation at the pharmacy if the prescription cannot be identified in the system, for example, if the prescription has been transferred to another pharmacy. We would not identify these events when prescriptions were transferred to pharmacies outside of our health system.
Our baseline rate of dispensing after discontinuation exceeded those in the previously published literature because we included in our analysis any discontinued e-prescription, even if a duplicate prescription existed at the pharmacy.2,3 As a result, not all medications that were dispensed after discontinuation resulted in a medication error. Long-term medications may be more likely to have duplicate prescriptions at the pharmacy, which could be associated with the higher proportion of immunosuppressants, anticoagulants and antiplatelet drugs, and cardiovascular medications that was dispensed after discontinuation prior to CancelRx implementation. Nevertheless, these long-term medications often require dose adjustment, and failure to communicate discontinuation of previous prescriptions may result in the wrong dose dispensed.5 We did not identify an increase in reordered medications after CancelRx implementation. This was used as a balancing measure for an increase in unintended discontinuation of e-prescriptions at the pharmacy (eg, error in medication reconciliation), which had been identified as a potential risk in the pilot analysis.
There are several next steps to maximize the safety benefits of CancelRx. First, health systems and remaining pharmacies must enable the CancelRx transaction. Although rates of adoption of CancelRx are increasing, only 57% of prescribers and 84% of pharmacies were estimated to have CancelRx enabled as of 2020.10 Second, to ensure that EHR users know whether a CancelRx resulted in successful inactivation of a prescription at the pharmacy, vendors should ensure that the outcome of the transaction is visible to prescribers in real time in their current workflows. While the status of the CancelRx message is visible in the order report, the CancelRx status is not visible in the medication list, which only indicates discontinuation in the EHR. Third, qualitative interviews with pharmacy staff suggest that additional information about the intent of the prescriber when discontinuing the e-prescription, the intended medication regimen, and clinical rationale for change would help identify which prescription(s) should be cancelled (this analysis is currently under review). Currently, the discontinuation reason selected by prescribers is not communicated to pharmacies as part of the CancelRx standard. Ideally, the information needed by pharmacy staff could be provided from the EHR with minimal impact on prescriber workflows. Fourth, to reduce noise in the system, vendors should identify strategies to transmit the highest-value messages and avoid low-value transactions, such as expired prescriptions. Receipt of low-value transactions was associated with negative perceptions of CancelRx at 1 institution.11 Fifth, EHR vendors should consider expansion of the functionality to medications reconciled from outside EHRs. In Epic, a large commercial vendor, a CancelRx is currently sent only for e-prescriptions and cannot be used to communicate discontinuation of medications that have been prescribed by a clinician outside of the local EHR.7 Expansion of CancelRx functionality to medications reconciled from outside the local EHR could further reduce the risk of medication errors.12
Limitations
Our study has some limitations. This study is limited to a single medical center with multiple outpatient pharmacies. We limited our analysis to dispensing up to 6 months after discontinuation, although a prescription could have been dispensed for up to 12 months. Pharmacy staff had access to the EHR, which might be expected to reduce the risk of medication errors compared with external pharmacies. Further studies across a larger network of clinical settings could further verify these findings.
Conclusions
This case series with interrupted time series analysis found that the implementation of CancelRx was associated with an immediate and persistent reduction in the proportion of e-prescriptions sold after discontinuation in the EHR. This study suggests that widespread implementation of CancelRx could significantly improve medication safety through the reduction of medication dispensing after discontinuation by prescribers.
Data Sharing Statement
References
- 1.Hodkinson A, Tyler N, Ashcroft DM, et al. Preventable medication harm across health care settings: a systematic review and meta-analysis. BMC Med. 2020;18(1):313. doi: 10.1186/s12916-020-01774-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Allen AS, Sequist TD. Pharmacy dispensing of electronically discontinued medications. Ann Intern Med. 2012;157(10):700-705. doi: 10.7326/0003-4819-157-10-201211200-00006 [DOI] [PubMed] [Google Scholar]
- 3.Copi EJ, Kelley LR, Fisher KK. Evaluation of the frequency of dispensing electronically discontinued medications and associated outcomes. J Am Pharm Assoc (2003). 2018;58(4S):S46-S50. doi: 10.1016/j.japh.2018.04.015 [DOI] [PubMed] [Google Scholar]
- 4.Shervani S, Madden W, Gleason LJ. Electronic health record interoperability—why electronically discontinued medications are still dispensed. JAMA Intern Med. 2021;181(10):1383-1384. doi: 10.1001/jamainternmed.2021.4881 [DOI] [PubMed] [Google Scholar]
- 5.Lourenco LM, Bursua A, Groo VL. Automatic errors: a case series on the errors inherent in electronic prescribing. J Gen Intern Med. 2016;31(7):808-811. doi: 10.1007/s11606-016-3606-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Nelson SD, Kumah-Crystal Y. Underuse of electronic health record features—the case for CancelRx. JAMA Intern Med. 2021;181(10):1384-1385. doi: 10.1001/jamainternmed.2021.4878 [DOI] [PubMed] [Google Scholar]
- 7.Pitts SI, Barasch N, Maslen AT, et al. Understanding CancelRx: results of end-to-end functional testing, proactive risk assessment, and pilot implementation. Appl Clin Inform. 2019;10(2):336-347. doi: 10.1055/s-0039-1688698 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Watterson TL, Stone JA, Brown R, et al. CancelRx: a health IT tool to reduce medication discrepancies in the outpatient setting. J Am Med Inform Assoc. 2021;28(7):1526-1533. doi: 10.1093/jamia/ocab038 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Pitts SI, Yang Y, Woodroof T, et al. The impact of electronic communication of medication discontinuation (CancelRx) on medication safety: a pilot study. J Patient Saf. 2022;18(6):e934-e937. doi: 10.1097/PTS.0000000000000998 [DOI] [PubMed] [Google Scholar]
- 10.Surescripts . National progress report 2020. Accessed July 25, 2023. https://surescripts.com/news-center/national-progress-report-2020
- 11.Watterson TL, Hernandez SE, Stone JA, Gilson AM, Ramly E, Chui MA. CancelRx implementation: observed changes to medication discontinuation workflows over time. Explor Res Clin Soc Pharm. 2022;5:100108. doi: 10.1016/j.rcsop.2022.100108 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Pitts SI, Yang Y, Thomas B, Chen AR. Discontinuation of outpatient medications: implications for electronic messaging to pharmacies using CancelRx. J Am Med Inform Assoc. 2022;29(12):2101-2104. doi: 10.1093/jamia/ocac181 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Sharing Statement