Continuous ECG monitoring should be the heart of bedside AI-based predictive analytics monitoring for early detection of clinical deterioration

Monfredi, Oliver; Moore, Christopher C.; Sullivan, Brynne A.; Keim‐Malpass, Jessica; Fairchild, Karen D.; Loftus, Tyler J.; Krahn, Katy N.; Dubrawski, Artur; Lake, Douglas E.; Moorman, J. Randall; Clermont, Gilles

doi:10.1016/j.jelectrocard.2022.10.011

Cited by 7 publications

(4 citation statements)

References 68 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Physiologic data is continuous, objective, and contains the earliest signs of the inflammatory response to infection via the autonomic nervous system 51 . An algorithm that runs on continuously patient-generated data allows for automated detection of the infant who is getting sick, rather than relying on clinicians with mixed skill levels to recognize early, subtle changes in a busy unit 52 .…”

Section: Discussionmentioning

confidence: 99%

Cardiorespiratory signature of neonatal sepsis: development and validation of prediction models in 3 NICUs

et al. 2023

Self Cite

View full text Add to dashboard Cite

Background: Heart rate characteristics aid early detection of late-onset sepsis (LOS), but respiratory data contain additional signatures of illness due to infection. Predictive models using cardiorespiratory data may improve early sepsis detection. We hypothesized that heart rate (HR) and oxygenation (SpO 2 ) data contain signatures that improve sepsis risk prediction over HR or demographics alone. Methods: We analyzed cardiorespiratory data from very low birth weight (VLBW, <1500g) infants admitted to three NICUs. We developed and externally validated four machine learning models to predict LOS using features calculated every 10m: mean, standard deviation, skewness, kurtosis of HR and SpO 2 , and cross-correlation. We compared feature importance, discrimination, calibration, and dynamic prediction across models and cohorts. We built models of demographics and HR or SpO 2 features alone for comparison with HR-SpO 2 models. Results: Performance, feature importance, and calibration were similar among modeling methods. All models had favorable external validation performance. The HR-SpO 2 model performed better than models using either HR or SpO 2 alone. Demographics improved the discrimination of all physiologic data models but dampened dynamic performance. Conclusions: Cardiorespiratory signatures detect LOS in VLBW infants at 3 NICUs. Demographics risk-stratify, but predictive modeling with both HR and SpO 2 features provides the best dynamic risk prediction.

show abstract

Section: Discussionmentioning

confidence: 99%

Cardiorespiratory signature of neonatal sepsis: development and validation of prediction models in 3 NICUs

et al. 2023

Self Cite

View full text Add to dashboard Cite

show abstract

“…However it is important to remember that detection of AF based on PPG currently requires confirmation of AF by ECG ( 29 ). Finally, continuous ECG monitoring is taking part of bedside AI-based predictive analytics monitoring ( 38 ) that could be useful for post stroke patient management in the future.…”

Section: Digital Devices As Alternative To Standard Equipment At Each...mentioning

confidence: 99%

Searching for atrial fibrillation post stroke: is it time for digital devices?

Piot

Guidoux

2023

Front. Cardiovasc. Med.

View full text Add to dashboard Cite

The detection of atrial fibrillation (AF) in patients with cryptogenic stroke (CS) is an essential part of management to limit the risk of recurrence. However, in practice, not all patients who need AF screening are screened, or are screened with significant delays. The disparities of access to examinations, their costs as well as the increasing workload require an evolution of practices both in terms of organization and the type of equipment used. The ubiquity and ease of use of digital devices, together with their evaluation in large population and their expected lower cost, make them attractive as potential alternatives to current equipment at all stages of patient management. However, reliability and accuracy of each digital device for the detection of paroxysmal AF in CS patients should be established before consideration for inclusion in clinical practice. The aim of this short analysis is therefore to review the current practical issues for AF detection in post stroke patients, the potential benefits and issues using digital devices in stroke patients and to position the different digital devices as alternative to standard equipment at each stage of stroke patient pathway. This may help to design future studies for the evaluation of these devices in this context. Under this condition, the time for digital devices to detect AF after stroke seems very close.

show abstract

“…We have argued in favor of the use of models based on cardiorespiratory physiology as a means to circumvent data drift. (Monfredi et al 2023) Our reasoning is that such observations as heart rate, respiratory rate, blood pressure, and the mathematical properties of continuous cardiorespiratory monitoring ( e . g ., physiological time series) are not subject to the ordering practices of clinicians and their association with clinical deterioration should be stable with respect to time.…”

Section: Introductionmentioning

confidence: 99%

Prospective validation of clinical deterioration predictive models prior to intensive care unit transfer among patients admitted to acute care cardiology wards

Keim-Malpass,

Moorman,

Moorman

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

Very few predictive models have been externally validated in a prospective cohort following the implementation of an artificial intelligence analytic system. This type of real-world validation is critically important due to the risk of data drift, or changes in data definitions or clinical practices over time, that could impact model performance in contemporaneous real-world cohorts. In this work, we report the model performance of a predictive analytics tool that was developed prior to COVID-19 and demonstrates model performance during the COVID-19 pandemic. The analytic system (CoMET®, Nihon Kohden Digital Health Solutions LLC, Irvine, CA) was implemented in a randomized controlled trial that enrolled 10,422 patient visits in a 1:1 display-on display-off design. The CoMET scores were calculated for all patients but only displayed in the display-on arm. Only the control/display-off group is reported here because the scores could not alter care patterns. Of the 5184 visits in the display-off arm, 311 experienced clinical deterioration and care escalation, resulting in transfer to the intensive care unit (ICU), primarily due to respiratory distress. The model performance of CoMET was assessed based on areas under the receiver operating characteristic curve, which ranged from 0.732 to 0.745. The models were well-calibrated, and there were dynamic increases in the model scores in the hours preceding the clinical deterioration events. A hypothetical alerting strategy based on a rise in score and duration of the rise would have had good performance, with a positive predictive value more than 10-fold the event rate. We conclude that predictive statistical models developed five years before study initiation had good model performance despite the passage of time and the impact of the COVID-19 pandemic. We speculate that some of the model performance’s stability is due to continuous cardiorespiratory monitoring, which should not drift as practices, policies, and patient populations change.Clinical Trial registrationClinicalTrials.govNCT04359641;https://clinicaltrials.gov/ct2/show/NCT04359641.

show abstract

Continuous ECG monitoring should be the heart of bedside AI-based predictive analytics monitoring for early detection of clinical deterioration

Cited by 7 publications

References 68 publications

Cardiorespiratory signature of neonatal sepsis: development and validation of prediction models in 3 NICUs

Cardiorespiratory signature of neonatal sepsis: development and validation of prediction models in 3 NICUs

Searching for atrial fibrillation post stroke: is it time for digital devices?

Prospective validation of clinical deterioration predictive models prior to intensive care unit transfer among patients admitted to acute care cardiology wards

Contact Info

Product

Resources

About