Frequency of Clinical Alarms in Intensive Care Units and Nurses' Sensitivity to Them

An Observational Study

Burcu Ceylan, PhD, RN; Leyla Baran, PhD, RN; Ülkü Yapucu Güneş, PhD, RN


Am J Crit Care. 2021;30(3):186-192. 

In This Article


In this study involving a total observation time of 80 hours, the mean alarm frequency per hour per bed across all ICUs was 1.8. Considering that a nurse is responsible for caring for 4 patients, the nurse must respond to 7.2 alarms per hour. In a study conducted in the internal medicine and surgical units of a medical center in the United States,[8] observations totaling 54 hours yielded an average frequency of alarms per hour of 3.79. Other observational studies conducted in ICUs[7,15] indicated a mean alarm frequency per hour of 6 to 10.9. The difference in results between previous research and the current study may be due in part to differences in the number and type of variables monitored by bedside devices. Another factor may be whether the counting method was manual or automatic. Such differences in counting methods and the observation period should be taken into account when interpreting findings. Individual adjustments made to patient alarms can also affect alarm frequency.

In the present study, false alarms constituted 15.3% of all the alarms. This finding was similar to the proportion of 13% found by Baillargeon[14] in a study involving 2 medical-surgical critical care clinics of a hospital with observations totaling 362 minutes. However, other studies have indicated a higher frequency of false or technical alarms than of valid alarms.[7,16] These differences in results may be related to differences in how the alarms were categorized. Moreover, previous research may have used an automatic counting system with a wider range of upper and lower limits compared with the current study. In the current study, most of the false alarms were due to changes in the position of the sensor, resulting in deviations from normal values. Cho et al[16] reported that false alarms stemmed from setting a single alarm interval for all patients.

In this study, nurses did not respond to 38.2% of valid alarms, 40.9% of false alarms, and 77.8% of technical alarms. In another study, nurses did not respond to 58.9% of technical alerts, and they responded to only 46.8% of valid alarms.[8] These results suggest that nurses had developed insensitivity to alarms.[15,16] In our study, we observed that over the course of 1 hour, false alarms were active for a mean of 14 minutes and technical alarms for a mean of 53 minutes. Although the average time to respond to valid alarms was 8 minutes, this time ranged widely, from 1 to 25 minutes. Nurses responded by initiating an intervention to almost all of the valid alarms sounding from ventilators, pulse oximeters, and enteral feeding pumps. Nurses generally did not respond to alarms from bedside monitors, and as the number of alarms from such monitors increased over the hour, the frequency of responses declined.

These findings indicate that nurse sensitivity to alarms depends greatly on the type of alarm— that is, the seriousness of the alarm and the type of device emitting it. Nurses take longer to respond to alarms that do not signal a need for an emergency intervention. Another study indicated that nurses took an average of 7 minutes to respond to alarms, with a response time of more than 10 minutes in 5 out of 8 responses.[14] The number of patients assigned to nurses influences their response as much as the nurses' sensitivity to alarms. In the ICUs included in this study, each nurse was in charge of 3 or 4 patients. We found that a mean of 2 alarms sounded per hour per bed. Considering that these alarms were active for at least 8 minutes, with technical alerts and false alarms sounding at intervals, the nurses were exposed to alarms for almost 50 minutes of every hour. We found a low frequency (13.9%) of technical alarms stemming from factors such as the device's failure to sense the probe or the probe not being connected to the patient, but the duration of the alarm activity was quite long (53 minutes). Technical alarms do not necessarily require a prompt response because they do not threaten patient safety. However, responding late or not at all to technical alarms may prevent nurses from becoming aware of valid alarms, putting patient safety at risk.[10]

In this study, considering that there were 12 beds in each ICU and a mean of 1.8 alarms sounding for each bed per hour, the mean number of alarms sounding in the ICU within 1 hour was 21.6. Thus, during an 8-hour workday, the nurse would be exposed to 173 alarms. This frequency of alarms results in extreme sensory overload in nurses working in these units, leading to desensitization or alarm fatigue. Alarm fatigue in the hospital setting is a serious issue given its negative impact on both caregivers and patients and its potential to lead to delayed treatment or poor outcomes, including severe injury or even death.[17,18] In a US study based on a national online survey, many nurses reported exposure to a high frequency of alarms and suggested that the resulting sensory overload could be a barrier to high-quality patient care, could diminish nurses' trust in the monitoring devices, and could cause employees to deactivate the devices.[5] Every individual has a limit to their capacity to perceive sensory stimuli.[8] In one study, the researchers reported that the sounding of 6 or more alarms per patient on a single shift increased the risk of confusion among employees.[2]