Skip to content

Examining the Impact of Machine Learning Algorithms on Extubation Failure Rate

Investigating the Impact of Machine Learning Algorithms on the Extubation Failure Rate in Mechanically Ventilated Patients Admitted to the Intensive Care Unit: A Randomized Clinical Trial

Status
Recruiting
Phases
Unknown
Study type
Interventional
Source
IRCT
Registry ID
IRCT20250503065574N1
Enrollment
60
Registered
2025-09-24
Start date
2025-09-23
Completion date
Unknown
Last updated
2025-10-13

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Dependence on enabling machines and devices, not elsewhere classified. Dependence on respirator

Interventions

Intervention 1: Intervention group: Intervention group: In this study, a machine learning model was first developed (data cleaning and preparation, data division into training and testing, machine lea

Sponsors

Mashhad University of Medical Sciences
Lead Sponsor

Eligibility

Sex/Gender
All
Age
18 Years to 75 Years

Inclusion criteria

Inclusion criteria: Intubated patients in the ICU for at least 12 hours who are ready for weaning from the ventilator undergoing a spontaneous breathing trial (SBT) with an oxygenation level (FiO2) of less than 50% positive end-expiratory pressure (PEEP) = 5 cmH2O pressure support (PS) = 5 cmH2O for 30 minutes to hours appropriate Glasgow Coma Scale (GCS) score (= 8–10) PaO2/FiO2 ratio > 150 ability to cough effectively and clear secretions adequate respiratory muscle tone stable hemodynamics and age range of 18–75 years

Exclusion criteria

Exclusion criteria: Patients with severe neurological disorders those unable to breathe due to neuromuscular diseases patients with a do-not-resuscitate (DNR) status Patients with airway obstruction

Design outcomes

Primary

MeasureTime frame
Extubation Failure Rates. Timepoint: Within 48 hours after Extubation. Method of measurement: Documentation of the need for re-intubation in the patient's medical record within 48 hours after extubation.

Secondary

MeasureTime frame
Length of ICU stay. Timepoint: From ICU admission to ICU discharge. Method of measurement: Calculation of the number of ICU hospitalization days from admission to discharge.;ICU mortality. Timepoint: up to the time of ICU discharge. Method of measurement: Documentation of the patient’s final status (alive or deceased) in the medical record.;Occurrence of post-extubation pulmonary edema. Timepoint: within 48 to 72 hours after extubation. Method of measurement: Clinical diagnosis and chest imaging.;Occurrence of aspiration after extubation. Timepoint: within 48 to 72 hours after extubation. Method of measurement: Clinical diagnosis or imaging evidence and documented.;Clinical assessment of respiratory muscle strength and inability to maintain effective ventilationOccurrence within 48 hours of extubationPost-extubation muscle fatigue. Timepoint: within 48 hours of extubation. Method of measurement: Clinical assessment of respiratory muscle strength and inability to maintain effective ventilationOccurrence.;Comparison of prediction accuracy of AI model and ICU physicians. Timepoint: After extubation and observing the final outcome (success or failure). Method of measurement: Comparison of sensitivity, specificity, positive and negative predictive value of AI model with physicians' decision in predicting extubation.

Countries

Iran (Islamic Republic of)

Contacts

Public ContactMohadeseh Rajabi

Mashhad University of Medical Sciences

mohadesehrajabi78@gmail.com+98 51 5229 4332

Outcome results

None listed

Source: IRCT (via WHO ICTRP) · Data processed: Feb 4, 2026