If your cell phone had the same ring tone as everyone else’s, would you always answer your phone? In a crowd of people with many phones going off at the same time, you would probably become conditioned to the sound and possibly ignore it. If technology had not leaped forward to allow people to personalize their cell phone ring tones, it is likely people would have developed one of two responses: If a phone in the vicinity were to ring, some would reach for their device every time, while others would never even hear it.
The latter phenomenon is quite common—think car alarms, snooze alarms, and the backup beeping of trucks. Even smoke alarms, fire drills, and boys who cry wolf can reach a point of not requiring attention. The human ability to tune out enables us to go about living and working even when the environment may not be ideal sound-wise. However, this same ability can lead to situations termed “alarm fatigue,” which are rare in the average individual’s life but are a serious concern for the medical professional, particularly nurses and other clinicians.
There are a multitude of alarms in the medical setting, each warning about something different and requiring a unique response. Unfortunately, those alarms that require no clinical action (either immediately or at all) run the risk of falling into the “alarm fatigue” category where they are no longer heard.
“Sometimes, those low priorities will eventually lead to a critical alarm and could be an indicator that something bigger is coming in the next 6 to 12 hours,” says Jillyan Morano, a clinical engineer with Linc Health at Winchester Hospital in Winchester, Mass. If missed, patients are potentially at risk for serious harm, and, in fact, such instances have occurred in hospitals throughout the country.
Clinical/biomedical engineering can help to combat alarm fatigue, and thereby improve patient care and safety, by guiding clinicians in the use of these systems, particularly the methods by which they can tailor alarm parameters to specific patient data. Ideally, this can serve to eliminate some of the alarm noise that does not require action, allowing clinicians to focus on the signals that do. “If you cater your alarms [to the patient], when you do hear a low priority, it’s more clinically significant than it was before catering,” Morano says.
To provide this assistance, biomeds have four roles, according to Jim Welch, CCE, vice president for the patient safety initiative of Masimo, Irvine, Calif.
- First: Develop familiarity with the challenges concerning alarm management.
- Second: Catalog the available alarm configurations within their inventory.
- Third: Reduce false alarms that result from wear and tear by maintaining the equipment.
- Fourth and Final—and most important, according to Welch: Participate with clinicians to optimize alarm configurations.
“[Clinicians and biomeds] should keep in mind that alarm management is not just about measuring how loud alarms sound,” Welch says. “It is about reducing the number of clinically insignificant alarms so that when a true alarm occurs, there is an appropriate clinical response.”
An Alarming Number of Alarms
Part of the challenge is the sheer number of alarms present in a hospital. “Almost every medical device found in the patient environment has one or more alarms,” Welch says.
These range from physiological monitoring or telemetry systems to ventilators and infusion pumps to beds and hypo/hyperthermia blankets. “There’s binging and bonging all over the place on a hospital floor,” Morano says.
How many devices are present depends on the environment of care. For instance, an ICU bed will have more alarms than one in a general care environment. And overall, everyone is seeing more alarms. “The number of medical devices at bedside is expanding due to a generally sicker patient population than was the norm a decade ago and the expectation that no patient be harmed during their hospital stay,” Welch says.
Alarms are used to alert care providers to potentially dangerous patient situations. “However, most medical devices are oversensitive to a potentially harmful situation at the expense of specificity—that is, a truly harmful situation actually exists,” Welch says. “Imagine your company’s fire alarm going off every time the heat rises from a passerby with a cup of hot coffee in hand. How useful would your fire alarm be?”
To combat this, some manufacturers try to build hierarchy into the alarms, such as different noises for different alerts: the less serious alarms get a less attention-getting sound, and vice versa. “Some alarms, such as those on some IV pumps and feeding pumps, will cause the device to stop functioning until the situation is corrected,” says Maria M. Cvach, MS, RN, CCRN, assistant director of nursing, clinical standards at The Johns Hopkins Hospital in Baltimore.
In general, alarms, even the less significant ones, require some action; the problem is that not all actions are clinically significant—some alarms just need to be acknowledged and turned off. These are the ones that tend to be ignored. As a result, alarms are classified into different categories.
True alarms require a clinical action to avoid harm to the patient. A true critical alarm will require immediate action (such as a physician’s response to a Code Blue hospital emergency code). False alarms are produced when the alarm is inaccurate. “For example, an alarm could sound indicating that a patient is in an arrhythmia but the patient could just be brushing their teeth,” Morano says. “This would be a false alarm. These false alarms contribute to alarm fatigue.
“Technical alarms indicate something is wrong with the system,” she continues. “For instance, it says ‘Leads Off’ or ‘SpO2 Probe Off.’ Because of alarm fatigue, these low-priority alarms can go unnoticed. If the leads are off the patient, then the patient is not being monitored, and that could lead to a missed event.”
Nuisance alarms, which are those most often responsible for alarm fatigue, are defined by IEC 60601-1-8 and are true, but do not require clinical action on the part of the care provider. “Nuisance alarms indicate a physiological event when no true clinically actionable event occurs,” Morano says.
Literature suggests that true actionable alarms account for less than a few percent of all patient alarms. “A survey published by the Healthcare Technology Foundation in 2006 identified physiologic monitors as the major contributor to nuisance alarms,” Welch says. “However, other devices, such as infusion pumps, ventilators, out-of-bed detectors, and other devices, all contribute to the desensitization of nurses to true patient safety events when they occur.”
This means that patients can be put at risk because an alarm could be unintentionally ignored, and unfortunately, cases of this nature have been reported. “The Joint Commission released a Sentinel Event Alert in 2002 that listed alarm fatigue as a significant factor for missed alarms that resulted in ventilator-related deaths and injuries,” Cvach says.
Alarm fatigue has, therefore, become an issue that health care institutions are grappling with. The problem is widespread and is the basis for the upcoming Medical Device Alarms Summit (to take place October 4 and 5) co-convened by the Association for the Advancement of Medical Instrumentation (AAMI), the FDA, the American College of Clinical Engineering, the ECRI Institute, and The Joint Commission. Its purpose is to develop priorities on the key issues surrounding the safety and effectiveness of medical device and system alarms.
Many health care institutions have developed, or are developing, alarm management policies designed to eliminate alarm fatigue and improve the use of alerting tools, but the process is difficult and diverse depending on the department. “Each unit has to develop an alarm management plan that works for their patients, because the response varies based on the patient population, the types of alarms, and the configuration of the unit,” Cvach says.
Yet, each unit shares the same goal. “The real challenge for nurses and clinicians is how to set alarm configurations—thresholds, prioritization, delays, etc—such that alarm specificity is high while nuisance alarms are suppressed as much as possible,” Welch says. This is where clinical engineering can help.
Cvach notes that units need help with knowing what types of alarms occur, how to standardize alarm defaults so they are actionable, and how to tailor alarms to meet patient needs. Much of this help can come in the form of data. “Clinical engineering can really help by collecting data for the units so the clinicians can make educated decisions about how to set the parameters,” Cvach says.
When Cvach first began focusing on alarm management, the nursing team had not even been aware that data could be collected from patient monitors. “It wasn’t until 2007 when our monitor group formed that we learned from clinical engineering there was a way to collect this data,” Cvach said. Now, she can e-mail clinical engineering to request specific data, eg, “12 days of alarm data for x unit from this date to this date,” and receive the information within 24 hours.
Cvach can then analyze the data for the types of alarms sounding, the duration, the time, and other parameters, and produce information that can help guide alarm adjustment. “We then do a before-and-after comparison to see whether or not we made a difference when we adjusted the parameters,” Cvach says.
What Cvach has found most interesting about the exercise is the large number of alarms that the data reveals do not prompt an action. However, she also understands the philosophy that has led to this common situation. “If I was not educated about this, I would say, ‘Turn all the alarms on,’ because that seems like the good thing to do,” Cvach says.
Alarm settings are often carried over from historical settings or taken from another area of the hospital and applied to a different environment of care (which may not be optimal). Or, clinicians set alarm configuration policies that have high sensitivity to true nonactionable alarms, which results in too many nuisance alarms. “The impact is that an illusion of patient safety is set, but too many nuisance alarms create an unsafe condition where frequent alarms have desensitized clinicians to actionable events,” Welch says.
Biomeds cannot tell clinicians how to set the alarms to avoid these situations, but they can provide guidance. Morano illustrates this fine line with an example of a patient whose default bradycardia alarm is set at 50 beats per minute but whose normal rhythm is 48 beats per minute. The patient’s alarm will constantly sound but may never require actual clinical attention.
“I can’t tell the clinicians to change the alarm setting to 40 beats per minute, but I can point out the situation and help educate them on how to use the system more effectively,” Morano says. Morano’s team recently acquired software that will allow them to measure alarm response so they can collect that relevant data for clinicians to help them tailor defaults to the unit and the patient.
Small changes can make a big difference. For example, some physiologic monitoring systems have the ability to hold off or delay audio alarm activation for a few seconds, which can substantially reduce the number of nonactionable alarms. “In the case of Masimo pulse oximeters, SpO2 threshold alarms can be delayed up to 15 seconds. At a low alarm threshold setting of 90% SpO2, an alarm delay of 15 seconds will reduce the number of alarms in a general care area by 70%,” Welch says. Unfortunately, Welch adds, not all multiparameter monitors that embed this technology adopt alarm delays.
However, other devices will likely have other ways to adjust settings. It is important for biomeds to maintain control over some of these adjustments, particularly those alarms that should never shut off, but clinicians should be taught how to adjust those alarms that can be adjusted.
Similarly, biomeds can also work with clinicians to develop a valuable notification system. “If you have 13 devices with 13 alarms, you have got to do a risk assessment to decide which ones require a high-priority notification,” Cvach says.
Integration poses a complex challenge, Welch concurs. “Bringing all alarms to a single device, such as a tablet or iPhone, increases the likelihood of alarm fatigue unless there is a well-thought-out hierarchy in alarm management,” he says. “For example, a true apnea alarm should have priority over an end-of-infusion alarm. The challenge is not just a priority structure; it’s how data gets integrated into actionable information.”
New technologies are emerging to help with these aims. And obtaining buy-in from administration is typically fairly easy because of the high-profile concern. A missed alarm can provide solid ground for more costly litigation. What is important is that medical professionals have realized more alarms are not the answer, but that smart alarming is.
Renee Diiulio is a contributing writer for 24×7. For more information contact .