In summary, we certainly do not have all the answers, but we can recommend data validation before the file is updated rather then at data analysis time. The best data-validation rules should use both covariates and previously recorded data. To be really accurate, the data repository needs to be in the patient-care loop rather than an afterthought. Allowing end-users from the start to browse through their data with an English-like query language will improve the accuracy of the data. Separate hierarchical source-oriented files have problems when joining data across files, but the advantages in implementation, storage costs, recovery time, and ease of end-user understanding far outweigh the disadvantages. Trying to get clinicians up-to-speed with modern-day multivariate statistical packages can easily backfire on you; it is better to leave the statistics to the statisticians and spend your energies in conveying the results with scattergrams and other graphs.
Computer-assisted detection of asymptomatic coronary artery disease in USAF flyers
Rechnerunterstuetzte Verfahren zur Erkennung der asymptomatischen koronaren Herzkrankheit von US Air Force Piloten
(9th) Computer in Cardiology ; 377-380
1982
4 Seiten, 4 Quellen
Conference paper
English
Online Contents | 1996
Online Contents | 2007
Screening for Asymptomatic Coronary Heart Disease
British Library Conference Proceedings | 1993
|Engineering Index Backfile | 1930
|American flyers over Pardubice
Online Contents | 1994