Predictive models of prolonged mechanical ventilation yield moderate accuracy

Figueroa-Casas, Juan B.; Dwivedi, Alok; Connery, Sean; Quansah, Raphael; Ellerbrook, Lowell; Galvis, Juan

doi:10.1016/j.jcrc.2015.01.020

Cited by 29 publications

(38 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Sharma et al focused on predicting which patients would require intubation exceeding 48 hours after having cardiac surgery [9]. Figueroacasas et al identified three different regression models for predicting within the first two days whether a patient would require prolonged intubation [10]. A final study conducted by Walgaard et al focused on identifying which factors were most predictive of prolonged MV in patients with Guillain-Barre syndrome [11].…”

Section: Related Workmentioning

confidence: 99%

Predicting and Understanding Unexpected Respiratory Decompensation in Critical Care Using Sparse and Heterogeneous Clinical Data

Ren

Johnson

Lehman

et al. 2018

2018 IEEE International Conference on Healthcare Informatics (ICHI)

View full text Add to dashboard Cite

Hospital intensive care units (ICUs) care for severely ill patients, many of whom require some form of organ support. Clinicians in ICUs are often challenged with integrating large volumes of continuously recorded physiological and clinical data in order to diagnose and treat patients. In this work, we focus on developing interpretable models for predicting unexpected respiratory decompensation requiring intubation in ICU patients. Predicting need for intubation could have important implications for the patient and medical staff and potentially enable timely interventions for improved patient outcome. Using data from adult ICU patients from the Medical Information Mart for Intensive Care (MIMIC)-III database, we developed gradient boosting models for predicting intubation onset. In a cohort of 12,470 patients, of whom 1,067 were intubated (8.55%), we achieved an area under the receiver operating characteristic curve (AUROC) of 0.89, with 95% confidence interval (CI) 0.87-0.91, when predicting intubation 3 hours ahead of time, a significant increase (p<0.001) over the AUROC achieved using several baselines, including logistic regression (0.81, 95% CI 0.78-0.84) and neural networks (0.80, 95% CI 0.77-0.83]). Finally, we conducted feature importance analysis using gradient boosting and derived useful insights in understanding the relative importance of clinical vs. biological variables in predicting impending respiratory decompensation in ICUs.

show abstract

Section: Related Workmentioning

confidence: 99%

Predicting and Understanding Unexpected Respiratory Decompensation in Critical Care Using Sparse and Heterogeneous Clinical Data

Ren

Johnson

Lehman

et al. 2018

2018 IEEE International Conference on Healthcare Informatics (ICHI)

View full text Add to dashboard Cite

show abstract

“…Many studies have attempted to identify predictors of prolonged mechanical ventilation (PMV) [3][4][5] as well as development of mortality risk prediction models for patients requiring 21 6 and 14 days of mechanical ventilation 7 .…”

Section: Introductionmentioning

confidence: 99%

Variation in Definition of Prolonged Mechanical Ventilation

et al. 2017

View full text Add to dashboard Cite

“…As no accurate method to predict the need for prolonged mechanical ventilation exists, unnecessary tracheostomy could not completely be avoided if early tracheostomy had been applied in all cases. It would be considered that 40% of tracheostomy in the early tracheostomy arm could be avoidable in the late tracheostomy arm [11]. Therefore, it cannot be determined that the benefits of performing early tracheostomy outweigh the harms of performing it in patients with ARDS.…”

Section: Resultsmentioning

confidence: 99%

The clinical practice guideline for the management of ARDS in Japan

et al. 2017

View full text Add to dashboard Cite

BackgroundThe Japanese Society of Respiratory Care Medicine and the Japanese Society of Intensive Care Medicine provide here a clinical practice guideline for the management of adult patients with ARDS in the ICU.MethodThe guideline was developed applying the GRADE system for performing robust systematic reviews with plausible recommendations. The guideline consists of 13 clinical questions mainly regarding ventilator settings and drug therapies (the last question includes 11 medications that are not approved for clinical use in Japan).ResultsThe recommendations for adult patients with ARDS include: we suggest against early tracheostomy (GRADE 2C), we suggest using NPPV for early respiratory management (GRADE 2C), we recommend the use of low tidal volumes at 6-8 mL/kg (GRADE 1B), we suggest setting the plateau pressure at 30cmH20 or less (GRADE2B), we suggest using PEEP within the range of plateau pressures less than or equal to 30cmH2O, without compromising hemodynamics (Grade 2B), and using higher PEEP levels in patients with moderate to severe ARDS (Grade 2B), we suggest using protocolized methods for liberation from mechanical ventilation (Grade 2D), we suggest prone positioning especially in patients with moderate to severe respiratory dysfunction (GRADE 2C), we suggest against the use of high frequency oscillation (GRADE 2C), we suggest the use of neuromuscular blocking agents in patients requiring mechanical ventilation under certain circumstances (GRADE 2B), we suggest fluid restriction in the management of ARDS (GRADE 2A), we do not suggest the use of neutrophil elastase inhibitors (GRADE 2D), we suggest the administration of steroids, equivalent to methylprednisolone 1-2mg/kg/ day (GRADE 2A), and we do not recommend other medications for the treatment of adult patients with ARDS (GRADE1B; inhaled/intravenous β2 stimulants, prostaglandin E1, activated protein C, ketoconazole, and lisofylline, GRADE 1C; inhaled nitric oxide, GRADE 1D; surfactant, GRADE 2B; granulocyte macrophage colony-stimulating factor, N-acetylcysteine, GRADE 2C; Statin.)ConclusionsThis article was translated from the Japanese version originally published as the ARDS clinical practice guidelines 2016 by the committee of ARDS clinical practice guideline (Tokyo, 2016, 293p, available from http://www.jsicm.org/ARDSGL/ARDSGL2016.pdf). The original article, written for Japanese healthcare providers, provides points of view that are different from those in other countries.Electronic supplementary materialThe online version of this article (doi:10.1186/s40560-017-0222-3) contains supplementary material, which is available to authorized users.

show abstract

Predictive models of prolonged mechanical ventilation yield moderate accuracy

Cited by 29 publications

References 23 publications

Predicting and Understanding Unexpected Respiratory Decompensation in Critical Care Using Sparse and Heterogeneous Clinical Data

Predicting and Understanding Unexpected Respiratory Decompensation in Critical Care Using Sparse and Heterogeneous Clinical Data

Variation in Definition of Prolonged Mechanical Ventilation

The clinical practice guideline for the management of ARDS in Japan

Contact Info

Product

Resources

About