Authenticity and Reliability of the Cardiac Monitor Alarms in the ICU: A Patient Safety Concern Due to Alarm Fatigue

Abstract

SESSION TITLE: Quality Improvement 1 SESSION TYPE: Original Investigation Poster PRESENTED ON: Wednesday, November 1, 2017 at 01:30 PM - 02:30 PM PURPOSE: It is desirable for the cardiac monitor (CM) to have a high sensitivity. Frequent alarms may cause alarm fatigue of the medical personnel leading to delayed or no response to the alarms, which could jeopardize patient’s safety. The impact of following standard processes/protocols such as pertinent skin preparation, changing electrocardiogram (EKG) leads daily, customization of alarm parameters, and education of the nursing staff to reduce the false alarms remains unclear. The aim of this study was to analyze the accuracy of the CM alarms in the intensive care unit (ICU) with utilization of the standard protocol. METHODS: This quality improvement project was conducted in the medical ICU of a teaching hospital in the United States from May 2016 to March 2017. Prior to initiation of this study, standard protocols for initial setting and maintenance of monitoring apparatus were taught to the nursing staff and were followed throughout the study. The alarm parameters were customized for each patient. Critical life threating cardiac alarms (red alarms) such as for asystole, tachycardia, bradycardia, ventricular tachycardia (VT) and ventricular fibrillation (VF) were included in the study. All the alarms were studied retrospectively for the occurrence of actual cardiac event suggested by the alarm. Positive predictive value was calculated for each arrhythmia. RESULTS: A total of 2408 alarms were analyzed on 350 patients. Overall, 557 (23.13%) were true positive (TP) and were clinically relevant, while majority (n=1851; 76.87%) were deemed to be false positive (FP). The most common type of alarms were for VT (~42%), followed by asystole (~23%) and tachycardia (20.5%). On detailed analysis, most of the tachycardia alarms (84.4%) were TP. On the contrary, majority of the alarms for VT (96.5%), asystole (~90%), VF (89.9%) and bradycardia (80.6%) were FP. Of 497 FP asystole alarms, 43 (8.6%) were FP due to EKG leads falling off the patient. The positive predictive value (PPV) of the alarm for tachycardia was the highest (84.4%), whereas the PPV of the alarm for VT (3.5%) was the lowest. CONCLUSIONS: This study demonstrated that significant number of alarms in the ICU are false without any clinical significance despite following the standard protocol to reduce the FP alarms. The alarm for tachycardia were more likely to be true in the majority of the patients and their higher accuracy resulted from customization of the alarm parameter settings. Though a higher sensitivity of CM alarms is desirable, false alarms are concerning and further study is warranted to increase the accuracy of CM alarms. CLINICAL IMPLICATIONS: Alarm fatigue ranked first on the list of “Top 10 Health Technology Hazards” released by Emergency Care Research Institute in 2012, 2013 and 2015. With this study, we highlight a pressing need of implementing an intelligent cardiac monitoring system in the ICUs. Besides being highly sensitive, the alarm system should also take into account other parameters before beeping. Reducing the number of FP alarms would lead to increased efficiency of the nursing staff and other medical personnel; hence would improve the patient safety. DISCLOSURE: The following authors have nothing to disclose: Parita Soni, Nidhi Aggarwal, Vivek Kumar, Peter Homel, Yizhak Kupfer No Product/Research Disclosure Information