The Johns Hopkins Hospital (Baltimore, MD) was selected as the winner of ECRI Institute's 7th Health Devices Achievement Award for its comprehensive study of alarm desensitization, or "alarm fatigue." This patient safety initiative has led to demonstrable improvements in the management of clinical alarms within many of the hospital's care units, reducing the potential for missed alarms and also reducing the noise levels, thereby providing a quieter working environment for caregivers and a safer, more restful environment for patients. James P. Keller, Jr., ECRI Institute's vice president for health technology evaluation and safety, formally presented the award at The Johns Hopkins Hospital on October 25, 2012.

The Health Devices Achievement Award recognizes outstanding initiatives undertaken by member healthcare institutions to improve patient safety, reduce costs, or otherwise facilitate better strategic management of health technology. ECRI Institute announced the winner and five other finalists for the 7th award in October 2012. Learn about the other submissions that achieved recognition.

ECRI Institute congratulates the project team: Maria Cvach; Peter Doyle, PhD; Adam Sapirstein, MD; and Andrew Currie. 

The Concept

Medical device alarms clearly perform an essential patient safety function. However, the sheer number of alarms in some care areas can itself become a patient safety concern. For example, when capturing baseline alarm data, The Johns Hopkins Hospital recorded an average of 317 alarms per bed per day in one of its intensive care units, and 771 alarms per bed per day in another (during analyses conducted over seven days). With so many medical devices sounding alarms and issuing alerts, caregivers can too easily become overwhelmed trying to respond to the alarms, or they can become desensitized, which can lead to missed alarms or delayed response. The problem is so pervasive, and the potential consequences so severe, that we have included alarm hazards at or near the top of our annual list of the Top 10 Health Technology Hazards every year since the list's inception.

To address problems associated with alarm fatigue, The Johns Hopkins Hospital formed a multidisciplinary Alarm Management Committee co-chaired by Maria Cvach, assistant director of nursing; Andrew Currie, director of clinical engineering; and Dr. Adam Sapirstein, a faculty member in Johns Hopkins's Armstrong Institute for Patient Safety and Quality. The committee studied the type, frequency, and duration of alarms that occur within particular care units and then used this information to develop and apply changes that would positively affect patient care. A key target of the initiative was to reduce the number of nonactionable audible alarms—alarms that sound for events that do not require staff to intervene. Reducing the number of nonactionable alarms not only reduces the overall alarm load, but also increases the percentage of actionable alarms; thus, nurses are more likely to respond promptly and correctly to the alarms that do sound.

The work of the committee included the following:

• Developing a fault tree analysis to identify all possible failure modes associated with a missed alarm. Analyzing the failure modes created an understanding of the complexity of the alarm management function and the need for redundancy (e.g., secondary notification methods) to prevent a failure.

• Collecting and analyzing baseline data with respect to the number and types of alarms occurring in each monitored care unit, and using this data to inform committee decisions on alarm-reduction strategies. This was no trivial task, as Maria Cvach notes: "It took two years to figure out how to get the data and, more importantly, to understand what it was telling us." One key strategy for minimizing recurring alarms that desensitize or distract clinicians involved making modest changes to alarm default settings. Working on a unit-by-unit basis, the committee identified safe default setting changes that could help reduce the number of nuisance alarms and other clinically insignificant alarms.

Additionally, software was purchased that allows the facility to build alarm-escalation algorithms specifying how and when an alarm should be communicated. For a noncrisis alarm condition, for example, a delay could be instituted so that caregivers are notified only if the alarm persists for a specified period of time (e.g., one minute); thus, conditions that quickly self-correct would not send an alert to the care provider.

• Developing an alarm policy and, where practical, standardizing equipment and methods. Consistent with the committee's goal of simplifying the alarm management approach, the hospital standardized monitoring equipment across the facility and implemented a uniform approach to alarm management. The committee even extended its standardization efforts down to the feature level—for example, alarm tones were standardized whenever possible so that similar conditions would sound similar tones in different care units. Thus, nurses and physicians who float among various care areas can more readily identify the meaning of the alarms they hear.

• Supporting the alarm policy through reeducation and competency training. As part of this effort, the committee assessed the educational needs of clinicians with respect to bedside monitoring.

• Analyzing and implementing methods for adjunct alarm notification. Each care unit was required to establish at least one secondary notification method to ensure that alarms are audible and appropriately managed. The committee recommended a handful of solutions that it had investigated and found to be effective; then, it allowed the individual care units to choose the notification systems that made the most sense for them, considering factors such as the unit's size and workflow. Thus, the system selected could vary from one care unit to the next.

The committee also collected and analyzed monitor-alarm data after interventions were made in order to confirm and quantify improvements. For example, the pilot program for this initiative, which was conducted in a 15-bed medical progressive care unit, yielded "a 43% reduction in critical physiologic monitor alarms from the baseline data collected approximately one year prior" (Graham and Cvach 2010). Success in that care area prompted the facility to expand its initiative, looking at other ways to improve alarm management and extending its efforts to other care areas. According to Ms. Cvach, "The beauty of the Alarm Management Committee is that it engages everybody. Members take the lessons that have been learned in other care units and apply them in their own units." 

Best Practices

The Johns Hopkins alarm management initiative demonstrates that the number of nonactionable alarms can be reduced—thereby decreasing caregivers' alarm burden without compromising patient safety—by making modest default parameter changes, standardizing care policies and equipment, and providing reliable secondary alarm notification. According to Ms. Cvach, "The things we did at Johns Hopkins really could be done at any hospital." All that was required was some hard work and desire: The organization invested the time to understand the problem, studied and tested various solutions, and then shared knowledge among various staff and departments. The project truly was a collaborative effort, involving contributions from nurses, physicians, clinical engineers, and IT personnel, as well as the cooperation of the hospital's monitor vendor.

Keys to the success of the program, as noted by Ms. Cvach, included:

• Taking the time to understand how the facility's monitoring systems worked
• Focusing on adjusting alarms to actionable levels
• Having "a great clinical engineering department"; clinical engineering staff were vital in helping committee members gain a true understanding of the facility's monitors and the implications of making any change
• Establishing an alarm committee "that allowed us to make decisions based on each other's experiences"
• Involving the frontline caregivers in decisions, allowing care units to select alternatives "that made sense for their particular workflow"

Also, the committee's thorough approach to collecting and analyzing data, both before and after interventions, allowed Johns Hopkins to measure the effectiveness of its efforts.