Exercise: Using Test Information II

This sample exercise and solution set supports the teaching pack on Using Test Information II, in which students are introduced to tests with continuous or categorical results, and calculate the sensitivity and specificity conditional on different cutoff points or positivity criterion. They are introduced to ROC curves and calculate the optimal positivity criterion. In addition to the sample exercise, materials include an instructor's note, videos, companion slides, a glossary, and an annotated bibliography.

Learning Objectives

1. Calculate the sensitivity and specificity conditional on different ‘cutoff points’ (i.e., positivity criterion) for a test with continuous results, and plot an ROC curve.
2. Describe the implications of operating at different parts of the ROC curve for any single test - explain the implications of ‘strict’ and ‘lenient’ cutoffs.
3. Calculate the optimal positivity criterion for a test given information on the prior probability, the test performance, and the relative consequences of true positives, true negatives, false negatives, and false positives.

The exercise provides just 2-3 simple examples of the types of short questions that can be useful for practicing some of the skills covered in this teaching pack. These should be modified, tailored, and expanded upon to meet the goals of specific classes.

This teaching pack was developed by Sue J. Goldie at the Center for Health Decision Science, Harvard T.H. Chan School of Public Health. The multimedia components were developed as part of a series of pilots in the CHDS Media Hub, led by Jake Waxman, where media-based pedagogy experiments contribute to new ways of thinking about short form content.