Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Generating a mortality model from a pediatric ICU (PICU) database utilizing knowledge discovery.

Curtis E Kennedy1, Noriaki Aoki

  • 1Department of Pediatrics, Baylor College of Medicine, Houston, TX, USA.

Proceedings. AMIA Symposium
|December 5, 2002
PubMed
Summary

Knowledge discovery algorithms created a pediatric intensive care unit (PICU) mortality model. This data-driven approach showed similar predictive performance to traditional methods.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Retreatment using temporary anchorage device for a patient with relapse after orthognathic surgery: a case report.

International journal of surgery case reports·2026
Same author

Prevalence of bone mineral disease in children with acute kidney disease on continuous kidney replacement therapy: a case-control study.

Pediatric nephrology (Berlin, Germany)·2026
Same author

Surgery First Bimaxillary Osteotomy for Marked Facial Asymmetry Developed After Aligner Treatment.

The Journal of craniofacial surgery·2026
Same author

Child Opportunity Index and Outcomes of Children Admitted to US PICUs for Critical Asthma.

Pediatric pulmonology·2025
Same author

Supernumerary intranasal floating tooth: clinical and literature review.

BMJ case reports·2025
Same author

Reduction of Temporomandibular Joint Dislocation With Open Approach Using Prearticular Incision Following Sagittal Split Ramus Osteotomy.

The Journal of craniofacial surgery·2025

Area of Science:

  • Medical Informatics
  • Machine Learning in Healthcare
  • Clinical Decision Support

Background:

  • Predictive outcome models are often limited by biases from pre-defined hypotheses.
  • Knowledge discovery algorithms offer a data-driven approach to model generation, reducing a priori bias.
  • Pediatric intensive care unit (PICU) data presents a complex environment for outcome prediction.

Purpose of the Study:

  • To develop a pediatric intensive care unit (PICU) mortality prediction model using knowledge discovery techniques.
  • To evaluate the performance of a data-driven model against a traditional algorithm.

Main Methods:

  • Utilized a PICU database of 5067 records with 192 variables.
  • Employed decision tree induction for model generation on a training set (75%).

Related Experiment Videos

  • Validated the model's performance on a separate validation set (25%) and compared it with the original PRISM algorithm.

    • Main Results:

    • The decision tree model identified 25 predictive variables.
    • The decision tree model achieved 33% sensitivity, 98% specificity, and 54% positive predictive value (PPV).
    • The PRISM algorithm achieved 30% sensitivity, 98% specificity, and 51% PPV; performance differences were not statistically significant.

    Conclusions:

    • Knowledge discovery algorithms can successfully generate mortality prediction models from PICU databases.
    • These data-driven techniques show comparable validity to established clinical prediction tools.
    • Validates the utility of knowledge discovery in the clinical medical domain for predictive modeling.