Dataset Information

Real-time machine learning-based intensive care unit alarm classification without prior knowledge of the underlying rhythm.

ABSTRACT:

Aims

This work attempts to develop a standalone heart rhythm alerting system for the intensive care unit (ICU), where life-threatening arrhythmias have to be identified/alerted more precisely and more instantaneously (i.e. with lower latency) than existing bedside monitors.

Methods and results

We use the dataset from the PhysioNet 2015 Challenge, which contains records that led to true and false arrhythmic alarms in the ICU. These records have been re-annotated as one of eight classes, namely (i) asystole, (ii) extreme bradycardia, (iii) extreme tachycardia, (iv) ventricular fibrillation (VF), (v) ventricular tachycardia (VT), (vi) normal sinus rhythm, (vii) sinus tachycardia, and (viii) noise/artefacts. Arrhythmia-specific features and features that measure the signal quality were extracted from all the records. To improve VF detection, an improved, over an existing, single-lead R-wave detection was developed that takes into account the R-waves detected in all electrocardiographic (ECG) leads. To avoid false R-wave detection due to pacing spikes, ECG signals were filtered with a low pass filter prior to R-wave detection, while the raw signals were used for feature extraction. Random forest was used as the classifier, and 10-time five-fold cross-validation, resulted in a macro-average sensitivity of 81.54%.

Conclusions

In conclusion, comparing with the bedside monitors used in the PhysioNet 2015 competition, we find that our method achieves higher positive predictive values for asystole, extreme bradycardia, VT, and VF; furthermore, our method is able to alert the presence of arrhythmia instantaneously, i.e. up to 4 s earlier.

SUBMITTER: Au-Yeung WM

PROVIDER: S-EPMC8482048 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundClinicians in intensive care units experience alarm fatigue related to frequent false and non-actionable alarms produced by physiologic monitors. To reduce non-actionable alarms, alarm settings may need to be customized for individual patients; however, nurses may not customize alarms because of competing demands and alarm fatigue.ObjectiveTo examine the effectiveness and acceptance of physiologic monitor software to support customization of alarms.MethodsThis pre/post intervention study was conducted in a 56-bed medical intensive care unit. IntelliVue® Alarm Advisor customization support software for alarm limit violations was installed on all monitors and education on its use provided. For 2 months before and after implementation of the software, data were collected on patient characteristics from the electronic health record, alarm counts and duration from the monitoring system, and nurses' experience of alarms from a survey.ResultsMedium-priority heart rate, respiratory rate, and arterial pressure alarms were significantly reduced after software implementation (9.3%, 11.8%, and 15.9% reduction respectively; p<0.001 for all). The duration of these alarms was also significantly shorter (7.8%, 13.3%, and 9.3% reduction respectively; p<0.05 for all). The number and duration of SpO2 alarms did not decrease (p>0.05 for both). Patients post-intervention had worse Glasgow Coma Scale scores (p = 0.014), but otherwise were comparable to those pre-intervention. Nurses reported less time spent on non-actionable alarms post-intervention than pre-intervention (p = 0.026). Also lower post-intervention were the proportions of nurses who reported that alarms disturbed their workflow (p = 0.027) and who encountered a situation where an important alarm was ignored (p = 0.043). The majority (>50%) agreed that the software supported setting appropriate alarm limits and was easy to use.ConclusionAlarm customization software was associated with a reduction in alarms. Use of software to support nurses' recognition of trends in patients' alarms and facilitate changes to alarm settings may add value to alarm reduction initiatives.