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

Limits to diagnostic accuracy

B S Todd1, R Stamper

  • 1Programming Research Group, Oxford University Computing Laboratory, UK.

Medical Informatics = Medecine Et Informatique
|July 1, 1993
PubMed
Summary
This summary is machine-generated.

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

Role of central corneal thickness on baseline parameters and progression of visual fields in open angle glaucoma.

European journal of ophthalmology·2007
Same author

Effect of population ageing on emergency department speed and efficiency: a historical perspective from a district general hospital in the UK.

Emergency medicine journal : EMJ·2006
Same author

Mozart in AVF testing.

The British journal of ophthalmology·2006
Same author

A computer program for generating New Deal compliant SHO rosters.

Emergency medicine journal : EMJ·2004
Same author

A mathematical specification of the New Deal on junior doctors' hours.

Medical informatics and the Internet in medicine·2003
Same author

Bilateral chronic hypotony following trabeculectomy with mitomycin-C.

Journal of glaucoma·2001
Same journal

Backpropagation and adaptive resonance theory in predicting suicidal risk.

Medical informatics = Medecine et informatique·1999
Same journal

Enhancing security and improving interoperability in healthcare information systems.

Medical informatics = Medecine et informatique·1999
Same journal

A multi-agent architecture for teaching dermatology.

Medical informatics = Medecine et informatique·1999
Same journal

A network-based training environment: a medical image processing paradigm.

Medical informatics = Medecine et informatique·1999
Same journal

Hippocrates: an integrated platform for telemedicine applications.

Medical informatics = Medecine et informatique·1999
Same journal

MEDNET97. Proceedings of a conference on the internet in medicine. November 1997.

Medical informatics = Medecine et informatique·1998
See all related articles

Computer-aided diagnosis accuracy is limited by incomplete data, not symptom interactions. Near-optimal results (75-80%) are achievable with Bayes

Area of Science:

  • Medical Informatics
  • Artificial Intelligence in Medicine
  • Diagnostic Systems

Background:

  • Computer-aided diagnosis (CAD) systems aim to improve medical diagnostic accuracy.
  • Understanding the limitations of CAD is crucial for developing effective tools.
  • Gynaecological abdominal pain diagnosis serves as a specific case study for CAD limitations.

Purpose of the Study:

  • To explore the factors limiting the accuracy of computer-aided medical diagnosis.
  • To investigate the impact of symptom interactions and data completeness on diagnostic performance.
  • To determine optimal strategies for achieving high diagnostic accuracy in CAD systems.

Main Methods:

  • A simulation model was employed to generate realistic medical cases for training and testing.

Related Experiment Videos

  • The model allowed for the creation of arbitrarily large datasets to overcome sample size limitations.
  • Bayes' theorem was applied, with and without the assumption of conditional independence, to assess diagnostic accuracy.

    • Main Results:

    • Statistical dependencies between symptoms and signs offer minimal improvement when considered in diagnostic models.
    • Incomplete recording of all possible observations significantly reduces diagnostic accuracy.
    • Near-optimal diagnostic accuracy (75-80%) was achieved using Bayes' theorem with conditional independence, given a large training set (10^5 cases).

    Conclusions:

    • Data completeness is a more critical factor than complex symptom interactions for CAD accuracy.
    • Bayes' theorem, under the assumption of conditional independence, provides a robust and efficient method for CAD.
    • Large datasets are essential for achieving high diagnostic performance in computer-aided medical diagnosis.