Department of Pharmacy Services and Clinical Pharmacy, Michigan Medicine, Ann Arbor, MI.
University of Michigan College of Pharmacy, Ann Arbor, MI, USA.
Am J Health Syst Pharm. 2022 Jul 8;79(14):1198-1204. doi: 10.1093/ajhp/zxac092.
To evaluate the effectiveness of clinical decision support (CDS) alerts tied to high-risk medications at a Michigan health system by determining the true prescriber action rate in response to select "do not give" (DNG) alerts.
A retrospective review of prescriber actions in response to CDS alerts was conducted to evaluate the effectiveness of alerts designed to prevent prescribing of high-risk medications to patients with concurrent DNG orders. The primary endpoint was the overall action rate, determined by totaling orders cancelled within the alert display and orders modified shortly after an alert. The overall action rate was hypothesized to significantly exceed the action rate estimated on the basis of alert overrides alone. Following the initial review, changes were made to the alert format and preset documentation choices ("acknowledgement comments"), and it was hypothesized that these changes would increase the overall action rate. A repeat analysis was conducted to evaluate the impact of these changes.
Across a total of 506 CDS alerts over 14 months, 78% resulted in prescribers modifying orders to comply with alert recommendations. Prescribers cancelled orders in response to only 26% of alerts, often overriding alerts prior to modifying orders. Documentation of rationale or approval for overrides was inconsistent, and while requiring acknowledgement comments facilitated documentation of prescriber rationale, it did not consistently improve overall action rates.
These findings demonstrate that override rates alone are not good markers for the true effectiveness of CDS alerts and support the need for frequent evaluation of alerts at the institutional level. CDS alerts remain a valuable tool to prevent inappropriate prescribing of high-risk medications and for promoting patient safety.
通过确定针对特定“不予给药”（DNG）警报的“真实”医嘱执行率，来评估密歇根州卫生系统中与高危药物相关的临床决策支持（CDS）警报的有效性。
对医嘱执行人员针对 CDS 警报的操作进行回顾性审查，以评估旨在防止向同时存在 DNG 医嘱的患者开具高危药物的警报的有效性。主要终点是总体操作率，通过将警报显示期间取消的医嘱和警报后不久修改的医嘱进行汇总来确定。总体操作率预计将显著超过仅基于警报覆盖的操作率。在首次审查后，对警报格式和预设文档选择（“确认评论”）进行了更改，并假设这些更改将提高总体操作率。重复进行了分析以评估这些更改的影响。
在总共 14 个月的 506 次 CDS 警报中，有 78%的警报导致医嘱执行者修改医嘱以符合警报建议。仅 26%的医嘱执行者因警报而取消医嘱，他们经常在修改医嘱前覆盖警报。对于覆盖的理由或批准的记录并不一致，虽然要求确认评论有助于记录医嘱执行者的理由，但并不总能提高总体操作率。
这些发现表明，仅覆盖率不能很好地标记 CDS 警报的真实有效性，并支持在机构层面上需要频繁评估警报。CDS 警报仍然是防止高危药物不当开具和促进患者安全的有效工具。
