G3SPSSAdvancedAbout 35 min to complete
Diagnostic accuracy and ROC curves
The 2x2 diagnostic table.
Slides · Dataset
About 35 minutes to complete
Enrol to open the slides and download the data.
What this module covers
the 2x2 diagnostic table; sensitivity, specificity, PPV, NPV and likelihood ratios, each with the clinical question it answers; why PPV depends on prevalence and sensitivity does not; the ROC curve and what a point on it represents; AUC and its interpretation; choosing a cut-off with the Youden index, and why the best cut-off depends on the cost of each error; the SPSS ROC path; comparing two tests; how to report diagnostic accuracy.
By the end of it
- The 2x2 diagnostic table
- Sensitivity, specificity, PPV, NPV and likelihood ratios, each with the clinical question it answers
- Why PPV depends on prevalence and sensitivity does not
- The ROC curve and what a point on it represents
- AUC and its interpretation
- Choosing a cut-off with the Youden index, and why the best cut-off depends on the cost of each error
- Comparing two tests
- How to report diagnostic accuracy
The dataset
TB symptom screening score against Xpert — Sigma teaching dataset
A teaching dataset of 600 cases, generated from a fixed seed so every figure in the module reproduces exactly.
- CodebookIncluded
- Dataset (CSV)Included
- Dataset (CSV)Included
- SPSS syntaxIncluded