SNM Annual Meeting Abstracts
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

J Nucl Med. 2008; 49 (Supplement 1):271P
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Google Scholar
Right arrow Articles by Santana, C.
Right arrow Articles by Garcia, E.
PubMed
Right arrow Articles by Santana, C.
Right arrow Articles by Garcia, E.

General Clinical Specialties: Operations/Practice Based/Outcomes Research

Operations/Practice Based/Outcomes Research Posters

Medical literature as source of knowledge for decision support systems

Cesar Santana1, Li Baoli2, Liudmila Verdes1, Saura Sahay2, Eugene Agichtein1, Ashwin Ram2 and Ernest Garcia1

1 Radiology, Emory University, Atlanta, Georgia; ; 2 Georgia Institute of Technology, Atlanta, Georgia

1161

Objectives: Determine if original scientific articles can be used to generate domain-specific, evidence-based knowledge that could be incorporated into expert systems for the diagnosis and prognosis of CAD in patients imaged with ECG-gated myocardial perfusion SPECT.

Methods: The abstracts from 454 original scientific articles published in the Journal of Nuclear Cardiology from 1995 to 2004 were used for this research. Two experts classified, scored and selected the knowledge that can be applied to clinical practice ("tomorrow morning") for the diagnosis and prognosis of patients with CAD.

Results: More than 40% of the abstracts were selected by the experts as having scientific evidence that can be used to generate heuristic rules for a decision support system for diagnosis and prognostic stratification for patients with CAD (Expert 1: 221(49%), Expert 2: 202(44%)). The experts’ agreement by categories (experimental: 166, diagnostic: 48, prognostic: 34, viability: 9 and technique: 13) was 60% (270/454), showing a contingency coefficient of 0.74, p<0.0001 in cross tabulation analysis. Analysis of the expert scoring of the level of support showed that the most of the abstracts were graded useful for rule generation (fine, good and higher: expert 1: 190/221=86%, expert 2: 192/202=95%). The correlation in the grading score (1 to 5) in the 101 abstracts that both experts classified as diagnostic, prognostic or viability was r= 0.5 (95% CI for r= 0.33 to 0.66, p<0.0001).

Conclusions: The knowledge from the scientific abstracts published in the JNC can potentially be used to generate heuristic rules for incorporation into Knowledge-based Expert Systems for the diagnosis and prognosis of coronary artery disease.





Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Google Scholar
Right arrow Articles by Santana, C.
Right arrow Articles by Garcia, E.
PubMed
Right arrow Articles by Santana, C.
Right arrow Articles by Garcia, E.