Chapter 8 · Biostatistics for Clinicians

Diagnostic Test Statistics

Sensitivity, specificity, PPV, NPV, likelihood ratios, Fagan nomogram, and ROC/AUC. Exercise 4 (penKid for AKI) can be entered via the Sens/Spec mode in the Calculator.

AUC:0.5Coin flip0.6Poor0.70–0.75Fair0.75+Generally useful0.80+Good0.90+Excellent

Input method

Load a chapter preset

Enter your 2×2 table

Disease+ and Test+
Disease− but Test+
Disease+ but Test−
Disease− and Test−
Total n: 1000
Study prevalence: 35%
Disease+: 350
Disease−: 650

Your table

Disease +
Disease −
Test +
284
TP
182
FP
Test −
66
FN
468
TN

2×2 Table Reference

Disease +
Disease −
Test +
TPTrue Positive
FPFalse Positive
Test −
FNFalse Negative
TNTrue Negative
SensitivityTP / (TP+FN)Of all who have disease, how many did I catch?
SpecificityTN / (TN+FP)Of all who don't, how many did I correctly clear?
PPVTP / (TP+FP)Of my positives, how many are real?
NPVTN / (TN+FN)Of my negatives, how many are truly clear?
Sensitivity
81.1%
Specificity
72%
PPV
60.9%
NPV
87.6%
LR+
2.9
LR−
0.26

Copy-paste result summary

Sensitivity 81.1%, Specificity 72%. PPV 60.9%, NPV 87.6%. LR+ 2.9, LR− 0.26. (n = 1000; disease prevalence in study = 35%)

Before you trust a paper's PPV — ask this question

"What was the prevalence in their study population, and is it the same as mine?"

Sensitivity and specificity are stable properties of the test. PPV and NPV are not — they belong to the test in that population. Calculate them for your population using the Prevalence Explorer.