An observational study in Extubation Failure and Critical Illness, sponsored by Universitas Indonesia. Recruiting at 2 sites in Indonesia. Open to participants aged 18 Years and older. Per ClinicalTrials.gov, last updated 2026-11-05.
Sponsored by Universitas Indonesia · Observational
The goal of this observational study is to predict who will have problems after the breathing tube is removed. Taking out the tube is called extubation, and having these problems is called extubation failure. Participants are adults in the intensive care unit (ICU) who use a breathing machine (ventilator).
The main questions it aims to answer are:
Extubation failure means the participant needs the tube put back in, needs a surgical opening in the neck to help breathing, or dies.
Researchers will use machine learning to build the computer tools. Machine learning is a type of artificial intelligence that finds patterns in data. Researchers will compare five types of machine learning to see which works best. They will build the tools with participants from Dr. Cipto Mangunkusumo National General Hospital in Jakarta. They will test the tools with participants from Universitas Indonesia Hospital in Depok.
Participants get their usual care. Doctors do not use the tools to make treatment decisions. Researchers will:
This is a prospective, multicenter, observational cohort study. Extubation and reintubation decisions are made by the attending anesthesiologists as part of routine care and are not informed by the models. The models will be developed only after data collection is complete.
Data collection:
Candidate predictors are chosen from previous studies and clinical relevance. They fall into four groups: patient characteristics, clinical condition, mechanical ventilation parameters, and laboratory results. Trained ICU staff and research assistants record them from bedside monitors, ventilator displays, paper and electronic medical records, and family interviews. Patient characteristics and severity scores (APACHE II and SAPS II) are recorded at ICU admission. Before extubation, ventilator parameters are taken within 4 hours, clinical condition within 24 hours, and laboratory results within 48 hours, using the measurement closest to extubation. Patients are followed for 7 days after planned extubation through bedside monitoring and review of medical records after transfer to the ward.
Participants:
Participants are enrolled when they receive invasive mechanical ventilation in the ICU, before it is known whether they will have a planned extubation. After enrollment, participants are excluded from the analysis if they die or have a tracheostomy without the breathing tube being removed; stay on a breathing machine for more than 2 weeks without the tube being removed; have an unplanned extubation (the tube was pulled out by the participant or by accident); are transferred to another hospital before extubation or within 7 days after; or if they or their family decline further intensive care. The number of participants enrolled is therefore larger than the number analyzed. The number analyzed will be reported in the results article.
Data management:
Data are entered into REDCap, a secure web-based data management application, using study codes without names, medical record numbers, or dates of birth. Data are stored at the Big Data Center, Indonesia Medical Education and Research Institute (IMERI), Universitas Indonesia. Research assistants periodically check the data for completeness and implausible values. Values that are not measured are left blank during data collection and will be handled in the analysis.
Sample size:
The development cohort target was calculated to estimate sensitivity (80%, precision of 10 percentage points, 95% confidence), assuming a 10% extubation failure rate and 10% dropout. The external validation target was calculated to estimate the extubation failure rate with a 5 percentage point margin of error. The development sample is smaller than formal prediction model sample size criteria would require. The study is therefore exploratory, and the external validation is preliminary.
Planned analysis:
Five supervised machine learning algorithms will be developed: LightGBM, CatBoost, XGBoost, random forest, and support vector machine. The development cohort will be randomly split, stratified by outcome, into training, validation, and holdout test sets. Missing values will be imputed, and class imbalance will be handled during model training. The classification threshold will be chosen on the validation set. Performance will be assessed on the holdout test set and on the external validation cohort, using the final models without modification. Predictor importance will be assessed with SHAP values for tree-based models and permutation importance for the support vector machine.
Reporting:
This study follows the TRIPOD+AI reporting guideline (Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis, updated for artificial intelligence).
1,780 studies on the registry are indexed under Critical Illness; 442 are open to participants now.
This study's planned enrollment of 1,065 is above the median of 140 across 814 observational studies indexed under Critical Illness.
Browse Critical Illness studies →Universitas Indonesia is the lead sponsor of 35 studies on the registry; 13 are open to participants now.
Counted across the registry records on this site, refreshed daily.
Adults treated in the intensive care units (ICUs) of two tertiary referral hospitals in Indonesia: Dr. Cipto Mangunkusumo National General Hospital in Jakarta (development cohort) and Universitas Indonesia Hospital in Depok (external validation cohort).
Go through the criteria one by one — yes, no or not sure — and take the list to your doctor or the study team. Kept in this browser only.
Exclusion Criteria:
- Used a breathing machine through a tracheostomy (a surgical opening in the neck) when admitted to the ICU
Adult ICU patients (≥18 years) on invasive mechanical ventilation via endotracheal tube at Dr. Cipto Mangunkusumo National General Hospital (RSCM), Jakarta, who underwent planned extubation. Enrolled between Sept 2024-Sept 2026. Pre-extubation predictors (patient characteristics, clinical condition, ventilator parameters, laboratory results) are recorded prospectively. Outcome: extubation failure, defined as reintubation, tracheostomy, or death from any cause within 7 days after planned extubation. Data are used to develop and internally validate machine learning models (LightGBM, CatBoost, XGBoost, random forest, SVM). No intervention is assigned; extubation follows routine care.
Adult ICU patients (≥18 years) on invasive mechanical ventilation via endotracheal tube at Universitas Indonesia Hospital (RS UI) who underwent planned extubation, using the same eligibility criteria, predictors, and 7-day outcome definition as the development cohort. Recruited during a period overlapping with the development cohort (Feb 2025-Sept 2026). Data are used only to externally validate the final models developed in the RSCM cohort, without modification. No intervention is assigned; extubation follows routine care.
Area under the receiver operating characteristic curve (AUROC) of machine learning models for predicting extubation failure
Discrimination of five supervised machine learning models (LightGBM, CatBoost, XGBoost, random forest, SVM) for predicting extubation failure in adult ICU patients on mechanical ventilation, including the difference in AUROC among the five algorithms. Extubation failure is defined as reintubation, tracheostomy, or death from any cause within 7 days after planned extubation. AUROC (unitless, range 0-1) is reported with 95% CI in the internal testing set (RSCM) and the external validation cohort (RS UI).
Time frame: up to 7 days after planned extubation
Accuracy of machine learning models for predicting extubation failure
Proportion of patients correctly classified (true positives plus true negatives, divided by all patients) by each of the five machine learning models for predicting extubation failure, at a probability threshold determined in the validation set by maximizing the F1-score. Reported as a percentage with 95% CI in the internal testing set and the external validation cohort.
Time frame: up to 7 days after planned extubation
Sensitivity (recall) of machine learning models for predicting extubation failure
Proportion of patients with extubation failure who are correctly identified (true positives divided by true positives plus false negatives) by each of the five machine learning models, at a probability threshold determined in the validation set by maximizing the F1-score. Reported as a percentage with 95% CI in the internal testing set and the external validation cohort.
Time frame: up to 7 days after planned extubation
Positive predictive value (precision) of machine learning models for predicting extubation failure
Proportion of patients predicted to have extubation failure who actually have extubation failure (true positives divided by true positives plus false positives) for each of the five machine learning models, at a probability threshold determined in the validation set by maximizing the F1-score. Reported as a percentage with 95% CI in the internal testing set and the external validation cohort.
Time frame: up to 7 days after planned extubation
Mean absolute SHAP value of candidate predictors for extubation failure (tree-based models)
Contribution of candidate predictors measured before extubation (patient characteristics, clinical condition, mechanical ventilation parameters, and laboratory results) to the predictions of each tree-based model (LightGBM, CatBoost, XGBoost, random forest), expressed as mean absolute SHAP value on the predicted probability scale (unitless, range 0-1).
Time frame: up to 7 days after planned extubation
Permutation importance of candidate predictors for extubation failure (SVM model)
Contribution of candidate predictors measured before extubation to the predictions of the support vector machine model, expressed as the mean decrease in AUROC (unitless) when the values of each predictor are randomly permuted in the internal testing set.
Time frame: up to 7 days after planned extubation
Not enrolling at this site.
Plan to share: No
Started Sep 2024 (actual). Primary completion is estimated for Jan 2027, 28 months after the start.
Of 439 non-industry observational studies under Critical Illness, half reached primary completion within 23 months of starting (the median), counting those still running; a quarter within 12 and three quarters within 45. This study's planned 28 months is above that.
Counting every sponsor, half of 442 studies reached it within 23 months; a quarter within 12 and three quarters within 46.
Critical Illness covers many different conditions, not one, so the studies compared here vary widely. Read these figures as a rough guide.
How these figures are worked out, and which studies are left out of them.
From the registry record's own update history. This site started tracking changes on Sep 25, 2026; for anything earlier, see the record history on ClinicalTrials.gov ↗
Eligibility is decided by the study team. Share this record with your doctor or contact the team directly.
Contact study teamGet an email when the registry record changes — status, dates, results — or when someone posts here.
Sign in to followQuestions and observations about this study, from anyone following it. Not medical advice, and not a channel to the study team — their contact details are on the registry record.
Sign in to join the discussion. Reading takes no account; posting does. You choose a display name, and a pseudonym is the default.
Nothing here yet. If you are running this trial, taking part in it, or weighing whether to, this is the place to say so.
Universitas Indonesia