Skip to main content
Fig. 3 | BMC Medical Informatics and Decision Making

Fig. 3

From: Use of diagnostic likelihood ratio of outcome to evaluate misclassification bias in the planning of database studies

Fig. 3

Expected RR as a function of DLR+ and the true disease risk. Expected risk ratio (RR) of the database (DB) study is shown as a function of positive diagnosis likelihood ratio (DLR+) and the true disease risk of the control (referent) group of the DB study. The true RR is set to 2.0. A Expected RR is plotted against DLR+. A hypothetical range of values for the true disease risk of the control group is shown (0.01–0.1). B Expected RR is plotted against the true disease risk of the control group. A range of values for the DLR+ is shown (20–1000). The right axis of each plot displays the scales in terms of % bias relative to the true RR. Both plots are based on Eq. 3. Figure X1 (Additional file 1) gives a more traditional display in which sensitivity and specificity are considered separately

Back to article page