Eschmann Emmanuel, Beeler Patrick E, Zünd Gregor, Blaser Jürg
Research Centre for Medical Informatics.
Stud Health Technol Inform. 2013;192:1056.
Electronic alerts for preventing hyperkalaemia during potassium-increasing drug-drug-interactions (DDIs) are often overridden due to their low specificity. Treatments of 76,467 inpatients were retrospectively analysed to establish more specific alerts. Alerting concepts for identifying DDIs that induced hyperkalaemia (serum potassium ≥5.5 mEq/l were compared. The positive predictive value (PPV) of alerts was 2.9% if they were triggered at onset of each potassium-increasing DDI. The PPV increased to 5.1% if alerts at onset were suppressed for serum potassium levels of <4.0 mEq/l. The PPV rose to 24.2% with a novel approach, triggering alerts whenever an elevated potassium level of >4.8 mEq/l was detected at onset or during the entire DDI period. Thus, triggering DDI alerts based on periodically monitored potassium levels may improve specificity of alerts and thereby reduce alert fatigue.
在增加钾的药物相互作用（DDIs）期间，用于预防高钾血症的电子警报由于其低特异性常常被忽略。对76467名住院患者的治疗进行回顾性分析，以建立更具特异性的警报。比较了用于识别导致高钾血症（血清钾≥5.5 mEq/L）的药物相互作用的警报概念。如果在每次增加钾的药物相互作用开始时触发警报，其阳性预测值（PPV）为2.9%。如果对于血清钾水平<4.0 mEq/L抑制开始时的警报，PPV增加到5.1%。采用一种新方法，每当在开始时或整个药物相互作用期间检测到钾水平>4.8 mEq/L升高时触发警报，PPV上升到24.2%。因此，基于定期监测的钾水平触发药物相互作用警报可能会提高警报的特异性，从而减少警报疲劳。
