Related Experiment Video
Updated: Nov 25, 2025
A Machine Learning Approach to Design an Efficient Selective Screening of Mild Cognitive Impairment
Published on: January 11, 2020
Avoidable Serum Potassium Testing in the Cardiac ICU: Development and Testing of a Machine-Learning Model
Bhaven B Patel1,2, Francesca Sperotto1,2,3,4,5,6, Mathieu Molina2
1Harvard Institute for Applied Computational Science, Harvard University, Cambridge, MA.
Insights
Machine learning models can predict which serum potassium blood tests are unnecessary for pediatric cardiac surgery patients. This approach could avoid 27.2% of tests, reducing costs and patient risks.
Area of Science:
- Pediatric cardiac surgery
- Clinical informatics
- Machine learning in healthcare
Background:
- Serum potassium monitoring is crucial for pediatric patients post-cardiac surgery.
- Frequent blood draws for potassium can lead to complications like anemia and infection.
- Optimizing blood draw frequency is essential for patient safety and resource management.
Purpose of the Study:
- To develop and validate a machine learning model to identify potentially avoidable serum potassium blood draws in pediatric cardiac surgery patients.
- To assess the accuracy and potential impact of such a model in a real-world clinical setting.
Main Methods:
- Retrospective cohort study at a tertiary-care center.
- Utilized machine learning (random forest classifiers) to predict normal potassium levels.
- Included variables such as serum chemistry, fluid balance, and renal function markers.
- Models were developed and validated on a large dataset of pediatric patients.
Main Results:
- The study included 7,269 admissions (6,196 patients) from January 2010 to December 2018.
- Machine learning models achieved a median positive predictive value of 0.900 for predicting normal potassium.
- A median of 27.2% of serum potassium samples were predicted as normal and potentially avoidable.
- The models demonstrated a low rate of incorrectly predicting abnormal potassium as normal.
Conclusions:
- Machine learning models can accurately predict potentially avoidable serum potassium blood tests in critically ill pediatric patients.
- Implementing such models can lead to significant reductions in unnecessary blood draws, decreasing costs and patient risks.
- This technology offers a promising approach to optimize laboratory testing in intensive care settings.
Objectives:
To create a machine-learning model identifying potentially avoidable blood draws for serum potassium among pediatric patients following cardiac surgery.
Design:
Retrospective cohort study.
Setting:
Tertiary-care center.
Patients:
All patients admitted to the cardiac ICU at Boston Children's Hospital between January 2010 and December 2018 with a length of stay greater than or equal to 4 days and greater than or equal to two recorded serum potassium measurements.
Interventions:
None.
Measurements And Main Results:
We collected variables related to potassium homeostasis, including serum chemistry, hourly potassium intake, diuretics, and urine output. Using established machine-learning techniques, including random forest classifiers, and hyperparameter tuning, we created models predicting whether a patient's potassium would be normal or abnormal based on the most recent potassium level, medications administered, urine output, and markers of renal function. We developed multiple models based on different age-categories and temporal proximity of the most recent potassium measurement. We assessed the predictive performance of the models using an independent test set. Of the 7,269 admissions (6,196 patients) included, serum potassium was measured on average of 1 (interquartile range, 0-1) time per day. Approximately 96% of patients received at least one dose of IV diuretic and 83% received a form of potassium supplementation. Our models predicted a normal potassium value with a median positive predictive value of 0.900. A median percentage of 2.1% measurements (mean 2.5%; interquartile range, 1.3-3.7%) was incorrectly predicted as normal when they were abnormal. A median percentage of 0.0% (interquartile range, 0.0-0.4%) critically low or high measurements was incorrectly predicted as normal. A median of 27.2% (interquartile range, 7.8-32.4%) of samples was correctly predicted to be normal and could have been potentially avoided.
Conclusions:
Machine-learning methods can be used to predict avoidable blood tests accurately for serum potassium in critically ill pediatric patients. A median of 27.2% of samples could have been saved, with decreased costs and risk of infection or anemia.
