Related Experiment Video
Updated: Sep 4, 2025
A Data-Driven Approach to Quantifying Immune States in Sepsis
Published on: February 7, 2025
Integrating structured and unstructured data for timely prediction of bloodstream infection among children
Azade Tabaie1, Evan W Orenstein2, Swaminathan Kandaswamy2
1Department of Biomedical Informatics, Emory School of Medicine, Atlanta, GA, USA. a.tabaie.87@gmail.com.
Insights
Hospitalized children with central venous lines (CVLs) face increased infection risks. Deep learning models integrating electronic health records (EHRs) and clinical notes can predict serious bloodstream infections more accurately.
Area of Science:
- Pediatric healthcare informatics
- Machine learning in medicine
- Infectious disease epidemiology
Background:
- Hospitalized children with central venous lines (CVLs) are susceptible to hospital-acquired infections.
- Electronic health records (EHRs) offer valuable data for predicting these infections.
- Serious bloodstream infections (SBSI) pose a significant risk in this vulnerable population.
Purpose of the Study:
- To evaluate the efficacy of deep learning models in predicting SBSI among pediatric patients with CVLs.
- To determine the added value of incorporating clinical notes alongside structured EHR data.
- To develop an advanced predictive model for rare events in complex healthcare settings.
Main Methods:
- A retrospective cohort study of pediatric patients with CVLs from 2013-2018.
- Extraction of structured EHR data and unstructured clinical notes.
- Training deep learning models to predict SBSI, comparing models with and without clinical note integration.
Main Results:
- The best model using only structured EHR data achieved a specificity of 0.951 and a positive predictive value (PPV) of 0.056 at 0.85 sensitivity.
- Incorporating clinical notes via contextualized word embeddings enhanced model performance, increasing specificity to 0.981 and PPV to 0.113.
- The study included 24,351 patient encounters meeting the inclusion criteria.
Conclusions:
- Integrating clinical notes with structured EHR data significantly improves the prediction of serious bloodstream infections in pediatric patients with CVLs.
- This approach offers a more accurate and timely method for identifying at-risk patients.
- The developed deep learning framework is applicable to predicting rare events in dynamic clinical environments.
Background:
Hospitalized children with central venous lines (CVLs) are at higher risk of hospital-acquired infections. Information in electronic health records (EHRs) can be employed in training deep learning models to predict the onset of these infections. We incorporated clinical notes in addition to structured EHR data to predict serious bloodstream infections, defined as positive blood culture followed by at least 4 days of new antimicrobial agent administration, among hospitalized children with CVLs.
Methods:
Structured EHR information and clinical notes were extracted for a retrospective cohort including all hospitalized patients with CVLs at a single tertiary care pediatric health system from 2013 to 2018. Deep learning models were trained to determine the added benefit of incorporating the information embedded in clinical notes in predicting serious bloodstream infection.
Results:
A total of 24,351 patient encounters met inclusion criteria. The best-performing model restricted to structured EHR data had a specificity of 0.951 and positive predictive value (PPV) of 0.056 when the sensitivity was set to 0.85. The addition of contextualized word embeddings improved the specificity to 0.981 and PPV to 0.113.
Conclusions:
Integrating clinical notes with structured EHR data improved the prediction of serious bloodstream infections among pediatric patients with CVLs.
Impact:
Developed an advanced infection prediction model in pediatrics that integrates the structured and unstructured EHRs. Extracted information from clinical notes to do timely prediction in a clinical setting. Developed a deep learning model framework that can be employed in predicting rare events in a complex and dynamic environment.
