This project will develop and study a model for the measurement of the diagnostic value of a Picture Archiving and Communication System (PACS). The model groups imaging examinations into diagnostic categories on the basis of clinical indications and for each category computes a diagnostic value index, that is based upon the accuracy and the timeliness of clinical access to the diagnostic image information. A mixture distribution technique will be studied as an alternative to the receiver operating characteristic (ROC) analysis for estimating diagnostic accuracy. The diagnostic value index will be related to the time required to take a clinical action. Data to support the model will be collected prospectively from three clinical sites: a medical intensive care unit, a surgical intensive care unit and an emergency department. Each site has different characteristics and will test the robustness of the diagnostic value model. At each site the PACS will be compared before and after an intervention designed to improve the effectiveness of the PACS.
Showing the most recent 10 out of 19 publications