Related Experiment Videos
Risk Factors for Pediatric Acute Chest Syndrome Utilizing Machine Learning Techniques
Kerry A Morrone1, Shweta Garg2, Boudewijn Aasman1
1Albert Einstein College of Medicine, Bronx, New York, United States.
Insights
Machine learning can predict acute chest syndrome (ACS) in children with sickle cell disease (SCD), identifying high-risk patients earlier. This tool aids in timely intervention for this serious complication.
Area of Science:
- Pediatric Hematology
- Medical Informatics
- Machine Learning in Medicine
Background:
- Acute Chest Syndrome (ACS) is a major cause of illness and death in pediatric sickle cell disease (SCD).
- ACS often presents unexpectedly despite known risk factors.
- Early prediction of ACS is crucial for improving patient outcomes.
Purpose of the Study:
- To develop and validate a machine learning model for predicting the progression to ACS in pediatric patients with SCD.
- To identify key clinical features associated with ACS development in this population.
Main Methods:
- A random forest machine learning model was developed using Scikit-Learn.
- The model utilized features including demographics, vital signs, lab values, medications, and comorbidities from 3223 patient encounters.
- Model performance was evaluated using C-statistic, sensitivity, specificity, and predictive values.
Main Results:
- The model achieved a sensitivity of 65.1% and specificity of 77.4% at a cutoff of 0.6.
- It predicted ACS a median of 18 hours before clinical diagnosis in positive cases.
- Important predictive features included pulse, respiratory rate, hypoxia, temperature, neutrophil count, and pain medication use.
Conclusions:
- This is the first machine learning model to predict ACS in children with SCD.
- Early detection of high-risk patients can potentially alter acute care management strategies.
- Further refinement of the model may improve prediction accuracy and clinical utility.
Abstract:
Acute chest syndrome (ACS) causes significant morbidity and mortality in both adult and pediatric patients with sickle cell disease (SCD). Despite known risk factors for developing ACS, this complication is often unexpected. We hypothesized that a machine learning model could predict progression to ACS in children. A random forest model was utilized using the Scikit-Learn Python library. Prediction performance was assessed using the C-statistic, sensitivity, and specificity. The features included are demographics, vital signs, laboratory values, medication orders, and additional comorbidities. The positive cohort developed ACS on diagnosis or during admission and the negative cohort did not. A total of 3223 encounters were included in the model for training and testing. 65 percent of the positive cohort developed ACS 24 hours after admission. The model had a sensitivity of 65.1% and specificity of 77.4% at the cutoff point of 0.6. For diagnosing ACS, the negative predictive value was 89.8% and the positive predictive value was 42.3%. For the correct positive cases the model predicted ACS a median of 18 hours before the diagnosis was determined in the electronic health record. The features that were important included pulse, respiratory rate, presence of hypoxia, temperature, neutrophil count and use of pain medications. Despite the model's limitations, it is the first model utilizing machine learning techniques to predict ACS in children. Detecting patients who are a higher risk for ACS could impact the approach to acute care management.
