Related Experiment Video
Updated: Jan 16, 2026
Manual Muscle Testing: A Method of Measuring Extremity Muscle Strength Applied to Critically Ill Patients
Published on: April 12, 2011
Development and validation of a machine learning model for early prediction of intensive care unit acquired weakness
Felipe Kenji Nakano1,2, Nathalie Van Aerde3, Grégoire Coppens3
1Department of Public Health and Primary Care, KU Leuven KULAK, Etienne Sabbelaan 53, 8500, Kortrijk, Belgium. felipekenji.nakano@kuleuven.be.
Insights
Machine learning models can predict intensive care unit (ICU) acquired weakness (ICU-AW) using early patient data. This helps identify high-need patients for tailored care protocols.
Area of Science:
- Critical Care Medicine
- Machine Learning in Healthcare
- Patient Outcomes
Background:
- Early identification of intensive care unit (ICU) patients at risk for high costs and needs is crucial for targeted care.
- ICU-acquired weakness (ICU-AW) within the first week is a predictor of adverse outcomes, but early prediction is challenging.
- Developing predictive models using data available within 24 hours of ICU admission is essential for timely intervention.
Purpose of the Study:
- To develop and validate a machine learning model for predicting ICU-acquired weakness (ICU-AW).
- To utilize data readily available within the first 24 hours of ICU admission for early risk assessment.
- To identify high-need patients who may benefit from specific care protocols.
Main Methods:
- A subset of 600 patients from the EPaNIC trial was used, with ICU-AW diagnosed if the Medical Research Council (MRC) sum score was higher than 48 at day 9.
- Three predictive models were compared: a random forest and a logistic regression model using day 1 data, and a random forest using only APACHE II score.
- Models were internally validated using 100 repetitions of fivefold cross-validation.
Main Results:
- The incidence of ICU-AW in the training set was 38.6%.
- The random forest model achieved an Area Under the Receiver Operating Characteristic Curve (AUROC) of 76%, outperforming logistic regression (74%).
- The random forest model demonstrated good calibration and clinical usefulness, identifying key predictors such as APACHE II, creatinine, and SOFA score.
Conclusions:
- Machine learning models, particularly random forests, can effectively predict ICU-AW risk using data from the first 24 hours of ICU admission.
- This predictive tool enables early prognostication in critically ill adult patients.
- The model has the potential to identify high-cost, high-need patients requiring different levels of care.
Background:
Early identification of potential high cost and high need patients on the ICU may assist in the development of targeted protocols, which allows proper resource utilization and initialization of preventive care. Weakness acquired in the ICU developed within the first week is an independent predictor of both short and long-term adverse outcomes, nonetheless early prediction is challenging. We aimed to develop and validate a machine learning model for ICU acquired-weakness (ICU-AW), using data readily available within the first 24 h of ICU admission.
Methods:
Patients from the EPaNIC trial (NCT00512122, N = 4640) who were assessed for muscle weakness at day 9 (IQR 8-13), after ICU-admission, using the Medical Research Council (MRC) sum. Patients are diagnosed with ICU-AW if their MRC is higher than 48. The final subset contains N = 600. Our models were internally validated using 100 repetitions of fivefold cross validation. We compared three predictive models: (i) a random forest and (ii) a logistic regression model built using descriptors available at day 1, (iii) a random forest using only APACHE II as a descriptor. Both random forests contain 150 trees.
Results:
The training set comprised 600 patients where the incidence of ICU-AW was 38.6% (232/600). The AUROC of the random forest with all descriptors and the logistic regression were 76% and 74%, respectively. The random forest (RF) achieved a specificity of 62% and a sensitivity 79%, whereas the logistic regression yielded 69% and 68%, respectively. The RF identified APACHE II, creatinine, SOFA PaO2/FiO2, bilirubin, BMI, age, glycemia upon admission, morning glycemia and sepsis as the most relevant descriptors. Lastly, the RF also presented very good calibration and clinical usefulness for a wide range of risk thresholds.
Conclusions:
Machine learning models, especially random forests, can be used to predict if patients are at risk of developing ICU-AW, using data available within 24 h of admission. This tool allows prognostication early in an adult general critically ill patient population, with the potential to detect high cost and high need patients who benefit from different levels of care.
