CoChat

Insights into the Problem of Alarm Fatigue with Physiologic Monitor Devices: A Comprehensive Observational Study of Consecutive Intensive Care Unit Patients

Open access

The DOI matches its Crossref record; 23 of the 23 references it cites were found in a registry or an index.

Abstract

PURPOSE: Physiologic monitors are plagued with alarms that create a cacophony of sounds and visual alerts causing "alarm fatigue" which creates an unsafe patient environment because a life-threatening event may be missed in this milieu of sensory overload. Using a state-of-the-art technology acquisition infrastructure, all monitor data including 7 ECG leads, all pressure, SpO(2), and respiration waveforms as well as user settings and alarms were stored on 461 adults treated in intensive care units. Using a well-defined alarm annotation protocol, nurse scientists with 95% inter-rater reliability annotated 12,671 arrhythmia alarms. RESULTS: A total of 2,558,760 unique alarms occurred in the 31-day study period: arrhythmia, 1,154,201; parameter, 612,927; technical, 791,632. There were 381,560 audible alarms for an audible alarm burden of 187/bed/day. 88.8% of the 12,671 annotated arrhythmia alarms were false positives. Conditions causing excessive alarms included inappropriate alarm settings, persistent atrial fibrillation, and non-actionable events such as PVC's and brief spikes in ST segments. Low amplitude QRS complexes in some, but not all available ECG leads caused undercounting and false arrhythmia alarms. Wide QRS complexes due to bundle branch block or ventricular pacemaker rhythm caused false alarms. 93% of the 168 true ventricular tachycardia alarms were not sustained long enough to warrant treatment. DISCUSSION: The excessive number of physiologic monitor alarms is a complex interplay of inappropriate user settings, patient conditions, and algorithm deficiencies. Device solutions should focus on use of all available ECG leads to identify non-artifact leads and leads with adequate QRS amplitude. Devices should provide prompts to aide in more appropriate tailoring of alarm settings to individual patients. Atrial fibrillation alarms should be limited to new onset and termination of the arrhythmia and delays for ST-segment and other parameter alarms should be configurable. Because computer devices are more reliable than humans, an opportunity exists to improve physiologic monitoring and reduce alarm fatigue.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1371/journal.pone.0110274

Cites

23 of 23 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1 Physiologic monitor device in Intensive Care Unit. Bedside patient monitor (GE Healthcare, Milwaukee, WI) displays multiple physiologic waveforms and vital sign measurements. The nurse pictured here gave written informed consent to publish this photograph supplied by the San Francisco Chronicle newspaper (with permission) for their story on alarm fatigue at: http://www.sfgate.com/health/article/Hospitals-look-to-reduce-danger-of-alarm-fatigue-4918018.php.
Figure 1 Physiologic monitor device in Intensive Care Unit. Bedside patient monitor (GE Healthcare, Milwaukee, WI) displays multiple physiologic waveforms and vital sign measurements. The nurse pictured here gave written informed consent to publish this photograph supplied by the San Francisco Chronicle newspaper (with permission) for their story on alarm fatigue at: http://www.sfgate.com/health/article/Hospitals-look-to-reduce-danger-of-alarm-fatigue-4918018.php.CC BY · PubMed Central
Figure 2 Hospital infrastructure to automatically store all physiologic monitor waveform and alarm data.
Figure 2 Hospital infrastructure to automatically store all physiologic monitor waveform and alarm data.CC BY · PubMed Central
Figure 3 Patient monitoring ECG lead configuration. A 5-electrode lead configuration was used in all study ICUs with Mason-Likar electrode placement of the limb leads on the torso and one chest electrode that is routinely placed in the V1 location.
Figure 3 Patient monitoring ECG lead configuration. A 5-electrode lead configuration was used in all study ICUs with Mason-Likar electrode placement of the limb leads on the torso and one chest electrode that is routinely placed in the V1 location.CC BY · PubMed Central
Figure 4 True positive ventricular tachycardia alarm using seven available ECG leads for diagnosis. Page one of the alarm annotation analysis tool shows a 10-second rhythm strip of all seven available ECG leads at the time that a ventricular tachycardia alarm was triggered. In this and subsequent Figures, ECG Leads are displayed from top to bottom in the following sequence: Lead I, II, III, V (typically V1), aVR, aVL, aVF. As evident at the beginning of the rhythm strip, the patient has an underlying rhythm of atrial fibrillation with a rapid ventricular rate of about 140. There is an isolated ventricular premature beat (4th beat from the end) and its QRS morphology is identical to the initial beat of the alarm event. Knowing that the event is initiated by a ventricular ectopic beat provides strong evidence that this event is a true ventricular tachycardia alarm.
Figure 4 True positive ventricular tachycardia alarm using seven available ECG leads for diagnosis. Page one of the alarm annotation analysis tool shows a 10-second rhythm strip of all seven available ECG leads at the time that a ventricular tachycardia alarm was triggered. In this and subsequent Figures, ECG Leads are displayed from top to bottom in the following sequence: Lead I, II, III, V (typically V1), aVR, aVL, aVF. As evident at the beginning of the rhythm strip, the patient has an underlying rhythm of atrial fibrillation with a rapid ventricular rate of about 140. There is an isolated ventricular premature beat (4th beat from the end) and its QRS morphology is identical to the initial beat of the alarm event. Knowing that the event is initiated by a ventricular ectopic beat provides strong evidence that this event is a true ventricular tachycardia alarm.CC BY · PubMed Central
Figure 5 True positive ventricular tachycardia alarm using non-ECG waveforms for diagnosis. Page 2 of the alarm annotation analysis tool depicts the same alarm event as in Figure 4 with all available non-ECG waveforms. Additional proof that this is a true ventricular tachycardia alarm is provided by observing cessation of the arterial blood pressure waveform that falls to near zero during the arrhythmia.
Figure 5 True positive ventricular tachycardia alarm using non-ECG waveforms for diagnosis. Page 2 of the alarm annotation analysis tool depicts the same alarm event as in Figure 4 with all available non-ECG waveforms. Additional proof that this is a true ventricular tachycardia alarm is provided by observing cessation of the arterial blood pressure waveform that falls to near zero during the arrhythmia.CC BY · PubMed Central
Figure 6 False positive ventricular tachycardia alarm using seven available ECG leads for diagnosis. Page one of the alarm annotation analysis tool in a second patient with a ventricular tachycardia alarm. Proof that this is a false positive alarm is provided by observing Lead III that shows clearly-visible P-QRS-T waveforms indicating normal sinus rhythm. All six remaining ECG leads show artifact that mimics ventricular tachycardia. It is important to point out that Lead III is not one of the two leads routinely displayed on the bedside monitor in our ICUs so unless all available leads are reviewed, a misdiagnosis would be made of rapid polymorphic ventricular tachycardia. This type of rapid, repetitive artifact on the ECG is often created during patient monitoring by motion artifact during activities of daily living. The non-artifact lead (Lead III) uses the left arm and left leg electrodes but not the right arm electrode. So, this is likely to be a right-handed patient doing something like brushing teeth.
Figure 6 False positive ventricular tachycardia alarm using seven available ECG leads for diagnosis. Page one of the alarm annotation analysis tool in a second patient with a ventricular tachycardia alarm. Proof that this is a false positive alarm is provided by observing Lead III that shows clearly-visible P-QRS-T waveforms indicating normal sinus rhythm. All six remaining ECG leads show artifact that mimics ventricular tachycardia. It is important to point out that Lead III is not one of the two leads routinely displayed on the bedside monitor in our ICUs so unless all available leads are reviewed, a misdiagnosis would be made of rapid polymorphic ventricular tachycardia. This type of rapid, repetitive artifact on the ECG is often created during patient monitoring by motion artifact during activities of daily living. The non-artifact lead (Lead III) uses the left arm and left leg electrodes but not the right arm electrode. So, this is likely to be a right-handed patient doing something like brushing teeth.CC BY · PubMed Central
Figure 7 False positive ventricular tachycardia alarm using non-ECG waveforms for diagnosis. Page 2 of the alarm annotation analysis tool depicts the same alarm event as in Figure 6 showing all available non-ECG waveforms. Additional proof that this is a false ventricular tachycardia alarm is provided by the following: a.) no change in the arterial pressure waveform during the event, b.) arterial waveform pulsations match the normal sinus rhythm rate, and c.) SpO2 waveform pulsations match the normal sinus rhythm rate. Of interest, the same artifact that contaminates the ECG signal also contaminates the respiratory waveform, as evidenced by an erroneous device-measured respiratory rate of 162 breaths per minute.
Figure 7 False positive ventricular tachycardia alarm using non-ECG waveforms for diagnosis. Page 2 of the alarm annotation analysis tool depicts the same alarm event as in Figure 6 showing all available non-ECG waveforms. Additional proof that this is a false ventricular tachycardia alarm is provided by the following: a.) no change in the arterial pressure waveform during the event, b.) arterial waveform pulsations match the normal sinus rhythm rate, and c.) SpO2 waveform pulsations match the normal sinus rhythm rate. Of interest, the same artifact that contaminates the ECG signal also contaminates the respiratory waveform, as evidenced by an erroneous device-measured respiratory rate of 162 breaths per minute.CC BY · PubMed Central
Figure 8 Frequency of all unique alarms (N = 2,558,760) over a 31-day period.
Figure 8 Frequency of all unique alarms (N = 2,558,760) over a 31-day period.CC BY · PubMed Central
Figure 9 False apnea alarm in a patient breathing adequately on mechanical ventilation. The respiratory waveform (bottom tracing labelled “Resp”) has a flat line appearance. The detection of respirations from the ECG lead (impedance method) is inaccurate in this patient, displaying an erroneous respiratory rate of 4 per minute.
Figure 9 False apnea alarm in a patient breathing adequately on mechanical ventilation. The respiratory waveform (bottom tracing labelled “Resp”) has a flat line appearance. The detection of respirations from the ECG lead (impedance method) is inaccurate in this patient, displaying an erroneous respiratory rate of 4 per minute.CC BY · PubMed Central
Figure 10 False alarm with one non-artifact ECG lead that confirms artifact mimicking ventricular fibrillation. Six of the seven ECG leads show what looks like a rapid (>400) polymorphic ventricular arrhythmia. However, Lead II clearly shows sinus rhythm at a rate of 94. Without this single non-artifact lead, a misdiagnosis would be made of ventricular fibrillation.
Figure 10 False alarm with one non-artifact ECG lead that confirms artifact mimicking ventricular fibrillation. Six of the seven ECG leads show what looks like a rapid (>400) polymorphic ventricular arrhythmia. However, Lead II clearly shows sinus rhythm at a rate of 94. Without this single non-artifact lead, a misdiagnosis would be made of ventricular fibrillation.CC BY · PubMed Central
Figure 11 False accelerated ventricular rhythm alarm in a patient with left bundle branch block. Sinus rhythm at a rate in the 60′s is evident by observing P waves preceding each QRS complex with a consistent PR interval. P waves are visible in all seven leads (especially clear-cut in Leads I and II).
Figure 11 False accelerated ventricular rhythm alarm in a patient with left bundle branch block. Sinus rhythm at a rate in the 60′s is evident by observing P waves preceding each QRS complex with a consistent PR interval. P waves are visible in all seven leads (especially clear-cut in Leads I and II).CC BY · PubMed Central
Figure 12 True accelerated ventricular rhythm alarm showing why this arrhythmia is not considered an “actionable” alarm condition. Accelerated ventricular rhythm at a rate of 56 for the first 5 beats followed by 2 fusion beats; the last 2 beats are normal sinus rhythm. The invasive arterial pressure waveform (bottom tracing) shows no change between normal rhythm and accelerated ventricular rhythm which confirms the rationale for the published guidelines stating no treatment is indicated for this arrhythmia in hospital settings [5].
Figure 12 True accelerated ventricular rhythm alarm showing why this arrhythmia is not considered an “actionable” alarm condition. Accelerated ventricular rhythm at a rate of 56 for the first 5 beats followed by 2 fusion beats; the last 2 beats are normal sinus rhythm. The invasive arterial pressure waveform (bottom tracing) shows no change between normal rhythm and accelerated ventricular rhythm which confirms the rationale for the published guidelines stating no treatment is indicated for this arrhythmia in hospital settings [5].CC BY · PubMed Central
Figure 13 False accelerated ventricular rhythm alarm in a patient with ventricular pacing. Patient with atrial fibrillation and intermittent ventricular pacing does not have PaceMode activated. As a result, a period of ventricular pacing goes undetected by the algorithm (no pacemaker spikes are “painted” in) and a false alarm is generated of accelerated ventricular rhythm. The investigators determined this to be intermittent pacing (rather than accelerated ventricular rhythm) because the rate matched the pacemaker heart rate setting. Moreover, the QRS morphology across all 7 leads matched the QRS morphology of corresponding leads on a hospital-acquired standard “diagnostic” 12-lead ECG during a known period of pacing.
Figure 13 False accelerated ventricular rhythm alarm in a patient with ventricular pacing. Patient with atrial fibrillation and intermittent ventricular pacing does not have PaceMode activated. As a result, a period of ventricular pacing goes undetected by the algorithm (no pacemaker spikes are “painted” in) and a false alarm is generated of accelerated ventricular rhythm. The investigators determined this to be intermittent pacing (rather than accelerated ventricular rhythm) because the rate matched the pacemaker heart rate setting. Moreover, the QRS morphology across all 7 leads matched the QRS morphology of corresponding leads on a hospital-acquired standard “diagnostic” 12-lead ECG during a known period of pacing.CC BY · PubMed Central
Figure 14 Low amplitude QRS in a patient with an excessive number of alarms. Standard “diagnostic” 12-lead ECG recorded from the patient who contributed nearly half of the 12,671 arrhythmia alarms for annotation. The ECG shows left bundle branch block with low amplitude QRS complexes in the limb leads but not in the V leads. Since one of the available leads acquired with the physiologic patient monitoring device is a V lead, the arrhythmia algorithm could have avoided the excessive number of false alarms had all available leads been used for QRS detection.
Figure 14 Low amplitude QRS in a patient with an excessive number of alarms. Standard “diagnostic” 12-lead ECG recorded from the patient who contributed nearly half of the 12,671 arrhythmia alarms for annotation. The ECG shows left bundle branch block with low amplitude QRS complexes in the limb leads but not in the V leads. Since one of the available leads acquired with the physiologic patient monitoring device is a V lead, the arrhythmia algorithm could have avoided the excessive number of false alarms had all available leads been used for QRS detection.CC BY · PubMed Central
Figure 15 ECG signal quality in 12,671 annotated arrhythmia alarms. Good signal quality (green) was defined as a clearly visible P-QRS-T waveform across all available leads with little to no noise, baseline wander, or leads off. Fair signal quality (yellow) was defined as moderate noise or baseline wander but having identifiable QRS complexes for basic rhythm/rate detection. Poor signal quality (red) was defined as being unanalyzable because of excessive noise, baseline wander or leads off.
Figure 15 ECG signal quality in 12,671 annotated arrhythmia alarms. Good signal quality (green) was defined as a clearly visible P-QRS-T waveform across all available leads with little to no noise, baseline wander, or leads off. Fair signal quality (yellow) was defined as moderate noise or baseline wander but having identifiable QRS complexes for basic rhythm/rate detection. Poor signal quality (red) was defined as being unanalyzable because of excessive noise, baseline wander or leads off.CC BY · PubMed Central
Figure 16 Electrode failure causing artifact and a false ventricular tachycardia alarm. Electrocardiogram in 6 of the 7 available leads shows intermittent loss of signal (signal “squares off” on top and bottom of tracing) due to an electrode problem such as loss of skin contact or dried out electrode gel. One ECG lead (Lead II that uses the right arm and left leg electrodes) does not show electrode failure so the likely electrode that is malfunctioning is the left arm electrode. Failure to apply fresh electrodes in this case will result in numerous false alarms.
Figure 16 Electrode failure causing artifact and a false ventricular tachycardia alarm. Electrocardiogram in 6 of the 7 available leads shows intermittent loss of signal (signal “squares off” on top and bottom of tracing) due to an electrode problem such as loss of skin contact or dried out electrode gel. One ECG lead (Lead II that uses the right arm and left leg electrodes) does not show electrode failure so the likely electrode that is malfunctioning is the left arm electrode. Failure to apply fresh electrodes in this case will result in numerous false alarms.CC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 23 references

Cited by

  • Scalable and accurate deep learning with electronic health records

    Rajkomar, A., Oren, E., Chen, K., Dai, A. M., Hajaj, N., Hardt, M., Liu, P. J., Liu, X., Marcus, J., Sun, M., et al. (2018) · npj Digital Medicine · doi:10.1038/s41746-018-0029-1 · Cited by 2,648

    “This is problematic because the resulting models may produce imprecise predictions: false-positive predictions can overwhelm physicians, nurses, and other providers with false alarms and subsequent alert fatigue [14], which the Joint Commission identified as a national patient safety priority in 2014 [10].”
    Background
  • Scalable and accurate deep learning with electronic health records.

    Rajkomar, A., Oren, E., Chen, K., Dai, A. M., Hajaj, N., Hardt, M., Liu, P. J., Liu, X., Marcus, J., Sun, M., et al. (2018) · eScholarship (California Digital Library) · Cited by 2,242

  • Ensuring fairness in machine learning to advance health equity.

    Rajkomar, A., Hardt, M., Howell, M. D., Corrado, G., & Chin, M. H. (2018) · Annals of Internal Medicine · doi:10.7326/m18-1990 · Cited by 1,341

    “If the model had a lower positive predictive value for African Americans, it might also disproportionately harm them through dismissal bias—a generalization of alert fatigue in which clinicians may learn to discount or “dismiss” alerts for African Americans because they are more likely to be false-positive (17).”
    Background
  • Big Data Analytics in Healthcare

    Belle, A., Thiagarajan, R., Soroushmehr, S. M. R., Navidi, F., Beard, D. A., & Najarian, K. (2015) · BioMed Research International · doi:10.1155/2015/370194 · Cited by 487

    “However, such uncompounded approaches towards development and implementation of alarm systems tend to be unreliable and their sheer numbers could cause “alarm fatigue” for both care givers and patients [10–12].”
    Background
  • Machine Learning and Decision Support in Critical Care

    Johnson, A. E. W., Ghassemi, M. M., Nemati, S., Niehaus, K. E., Clifton, D., & Clifford, G. D. (2016) · Proceedings of the IEEE · doi:10.1109/jproc.2015.2501978 · Cited by 432

    “[65] reviewed the issue of alarm fatigue associated with false alarms and suggested alarm algorithms should focus on: using all available ECG leads and extracting at least one lead with high quality data if available, providing contextual alarms based upon multiple features (e.”
    Background

See all 457

Metadata

Authors
Drew, B. J., Harris, P., Zègre-Hemsey, J. K., Mammone, T., Schindler, D., Salas-Boni, R., Bai, Y., Tinoco, A., Ding, Q., & Hu, X.
Affiliations
  • University of California, San Francisco (US)
  • University of California San Francisco Medical Center (US)
Corresponding author
Drew, Barbara J.
Venue
PLoS ONE
Published
2014-10-22
Volume
9
Issue
10
Pages
e110274
ISSN
1932-6203 (ISSN-L)
OpenAlex
W2001741247
Cited by
438
References
23
Citations by year
2026: 51 · 2025: 31 · 2024: 25 · 2023: 37 · 2022: 26 · 2021: 58 · 2020: 29 · 2019: 38 · 2018: 42 · 2017: 36 · 2016: 43 · 2015: 21 · 2014: 1
FWCI
33.15
Type
journal article
Language
en
Open access
Locations
  • PLoS ONE (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Carolina Digital Repository (University of North Carolina at Chapel Hill) (submittedVersion)
  • Dominican Scholar (Dominican University of California) (submittedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • eScholarship (California Digital Library) (submittedVersion)
  • UNC Libraries
Funders
University of California, San Francisco
Topics
  • Health Sciences › Medicine › Surgery › Healthcare Technology and Patient Monitoring
  • Health Sciences › Medicine › Public Health, Environmental and Occupational Health › Patient Safety and Medication Errors
  • Health Sciences › Health Professions › Health Information Management › Electronic Health Records Systems
Keywords
alarm fatigue, intensive care unit patients, false positive alerts, alarm settings, ECG leads, QRS amplitude, atrial fibrillation, ventricular tachycardia, bundle branch block, persistent atrial fibrillation, patient safety
MeSH terms
Adult, Aged, Arrhythmias, Cardiac, Atrial Fibrillation, Diagnostic Errors, Electrocardiography, Equipment Failure, False Positive Reactions, Female, Humans, Intensive Care Units, Male, Middle Aged, Monitoring, Physiologic, Critical Illness, Clinical Alarms

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Insights into the Problem of Alarm Fatigue with Physiologic Monitor Devices: A Comprehensive Observational Study of Consecutive Intensive Care Unit Patients. https://app.cochat.ai/papers/10.1371/journal.pone.0110274
The paper
Drew, B. J., Harris, P., Zègre-Hemsey, J. K., Mammone, T., Schindler, D., Salas-Boni, R., Bai, Y., Tinoco, A., Ding, Q., & Hu, X. (2014). Insights into the Problem of Alarm Fatigue with Physiologic Monitor Devices: A Comprehensive Observational Study of Consecutive Intensive Care Unit Patients. PLoS ONE. https://doi.org/10.1371/journal.pone.0110274

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.