Predicting presumed serious infection among hospitalized children on central venous lines with machine learning
Azade Tabaie1, Evan W Orenstein2, Shamim Nemati3
1Department of Biomedical Informatics, Emory School of Medicine, Atlanta, GA, USA.
Insights
Machine learning accurately predicts serious infections in pediatric patients with central venous lines up to 8 hours earlier. This early detection of presumed serious infection (PSI) aids timely intervention and improves care quality.
Area of Science:
- Pediatric critical care medicine
- Medical informatics
- Machine learning in healthcare
Background:
- Presumed serious infection (PSI) in pediatric patients with central venous lines (CVLs) requires early detection.
- Timely intervention for PSI can prevent adverse outcomes and enhance patient care quality.
Purpose of the Study:
- To develop and validate a machine learning model for early prediction of PSI in pediatric patients with CVLs.
- To identify key clinical features predictive of PSI onset.
Main Methods:
- Retrospective analysis of electronic medical records from a pediatric health system.
- Training XGBoost and ElasticNet machine learning models to predict PSI 8 hours in advance.
- Benchmarking model performance against the PRISM-III score.
Main Results:
- The machine learning model achieved an AUC of 0.84, with 73% sensitivity and 36% PPV.
- PRISM-III demonstrated lower performance (19% sensitivity, 30% PPV) at a cutoff of ≥10.
- Key predictive features included diastolic blood pressure, height, and temperature.
Conclusions:
- Machine learning models utilizing electronic health record data can effectively predict PSI onset in pediatric patients with CVLs.
- Early prediction allows for proactive medical interventions, potentially improving patient outcomes.
Background:
Presumed serious infection (PSI) is defined as a blood culture drawn and new antibiotic course of at least 4 days among pediatric patients with Central Venous Lines (CVLs). Early PSI prediction and use of medical interventions can prevent adverse outcomes and improve the quality of care.
Methods:
Clinical features including demographics, laboratory results, vital signs, characteristics of the CVLs and medications used were extracted retrospectively from electronic medical records. Data were aggregated across all hospitals within a single pediatric health system and used to train machine learning models (XGBoost and ElasticNet) to predict the occurrence of PSI 8 h prior to clinical suspicion. Prediction for PSI was benchmarked against PRISM-III.
Results:
Our model achieved an area under the receiver operating characteristic curve of 0.84 (95% CI = [0.82, 0.85]), sensitivity of 0.73 [0.69, 0.74], and positive predictive value (PPV) of 0.36 [0.34, 0.36]. The PRISM-III conversely achieved a lower sensitivity of 0.19 [0.16, 0.22] and PPV of 0.30 [0.26, 0.34] at a cut-off of ≥ 10. The features with the most impact on the PSI prediction were maximum diastolic blood pressure prior to PSI prediction (mean SHAP = 3.4), height (mean SHAP = 3.2), and maximum temperature prior to PSI prediction (mean SHAP = 2.6).
Conclusion:
A machine learning model using common features in the electronic medical records can predict the onset of serious infections in children with central venous lines at least 8 h prior to when a clinical team drew a blood culture.
Related Concept Videos
01:27Parentral Nutrition: Centeral and Peripheral Parental Nutrition
PN can be administered through two primary routes:
1. Central Parenteral Nutrition (CPN):
CPN involves delivering a high concentration of nutrients through a large vein. This is typically achieved using a Peripherally Inserted Central Catheter (PICC) or,...
01:18Steps in Outbreak Investigation
01:29Endocarditis IV: Nursing Management
01:22Healthcare Associated Infections II: Preventive Measures
The best practices for preventing healthcare-associated infections include hand hygiene, patient risk...