|
|
||||||||
Correspondence: Requests for reprints should be addressed to: Joseph M. Miller, M.D., M.P.H., Dept. of Ophthalmology and Vision Science, University of Arizona, Tucson, AZ.
Introduction: The design of a rational screening program is predicated upon reliable rates of referral with know sensitivity and specificity for a given referral threshold. Such decision making is facilitated through the use of Receiver Operator Characteristic (ROC) curves.
Methods: The generation of an ROC curve is demonstrated for a hypothetical vision screening instrument in a simulated population of children needing eye glasses.
Results: An ROC curve, once calculated for a screening technique when applied to a given population, allows reliable prediction of referral rates for a given sensitivity or specificity.
Conclusion: ROC curves provide a reliable method to facilitate decision making with regard to a screening program, and a method to compare the effectiveness of one screening program to another for the same population.
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |