PERFORMANCE TESTING FOR AUTOMATED PRESCREENING DEVICES IN CERVICAL CYTOLOGY

Abstract
Performance specifications for automated prescreening devices for cervical cancer will concern rates for missed positive samples and rates for false alarms. Testing whether a given device adheres to such specifications within given confidence limits should be feasible for the missed positive rates. The test for adherence to the specified false alarm rate presents a serious sampling problem.

This publication has 0 references indexed in Scilit: