Lesson 5 · Other Designs and Approaches to Know · Step 6 of 7
Diagnostic accuracy studies
A test can miss disease or flag disease that isn't there, and the two errors have different costs. Sensitivity is the share of sick people who test positive. Specificity is the share of healthy people who test negative. Predictive values depend on prevalence, so at 1% prevalence a 98% specific test still produces mostly false positives. Accuracy studies need a sound reference standard, a single-gate sample with the realistic spectrum of patients, and blinding. Xpert MTB/RIF reached 90% sensitivity for culture-positive TB, though accuracy alone doesn't show that a test improves outcomes.
Also covered: sensitivity and specificity · predictive values and prevalence · ROC curves and thresholds · spectrum and verification bias
Alongside this step
- Slides and transcript — the whole video
- Read the chapter — Other Designs and Approaches to Know
Lesson 5 · 7 steps
Research Designs · lesson 5 of 5