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

Transferability of medical decision support systems based on Bayesian classification.

R J Zagoria, J A Reggia

    Medical Decision Making : an International Journal of the Society for Medical Decision Making
    |January 1, 1983
    PubMed
    Summary

    Bayesian decision support systems using distant stroke patient data accurately predicted stroke causes locally. This demonstrates the surprising transferability and utility of clinical databases for building and testing such systems.

    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

    Controlling working memory with learned instructions.

    Neural networks : the official journal of the International Neural Network Society·2013
    Same author

    Post-lesion lateralisation shifts in a computational model of single-word reading.

    Laterality·2004
    Same author

    Radio frequency ablation of lung metastases from renal cell carcinoma.

    The Journal of urology·2001
    Same author

    Go forth and replicate.

    Scientific American·2001
    Same author

    Cortical spreading depression and the pathogenesis of brain disorders: a computational and neural network-based investigation.

    Neurological research·2001
    Same author

    The callosal dilemma: explaining diaschisis in the context of hemispheric rivalry via a neural network model.

    Neurological research·2001

    Area of Science:

    • Medical informatics
    • Clinical decision support
    • Epidemiology

    Background:

    • Predicting stroke etiology is complex and often relies on local data.
    • Existing Bayesian systems may be limited by site-specific probability data.

    Purpose of the Study:

    • To test if a Bayesian decision support system trained on geographically distant stroke data could accurately predict local stroke etiology.
    • To evaluate the performance of an "extrainstitutional" Bayesian system against local physician diagnoses and local probability-based Bayesian systems.

    Main Methods:

    • Retrospective analysis of 100 stroke cases.
    • Development and application of an "extrainstitutional" Bayesian classification system using a large, distant stroke patient database.
    • Performance assessment using error rates and a novel linear accuracy coefficient.

    Related Experiment Videos

    Main Results:

    • The "extrainstitutional" Bayesian system demonstrated surprising accuracy in predicting stroke etiology.
    • The system's performance surpassed both local physician classification and Bayesian classification based on local, subjective probabilities.
    • The linear accuracy coefficient provided a new metric for evaluating classification performance.

    Conclusions:

    • Bayesian classification systems can be highly transferable to new clinical sites for problems like stroke etiology prediction.
    • Clinical databases are valuable resources for developing, transferring, and validating Bayesian decision support systems.
    • This approach offers a promising method for improving the accuracy and consistency of stroke diagnosis globally.