Biostatistics Tools

Diagnostic Test Evaluation

Prof. Dr. Doğukan ÖZEN

Calculate diagnostic accuracy measures, confidence intervals, and post-test probabilities from a 2×2 contingency table.

Local calculation

2×2 Table Data

Compare the index test result with the reference standard.

	Reference positive Disease present	Reference negative Disease absent
Test positive	TPTrue positive	FPFalse positive
Test negative	FNFalse negative	TNTrue negative

Confidence level

Decimal places

Zero-cell correction: Adds 0.5 to all cells when required for likelihood ratios and the diagnostic odds ratio.

Main Results

Proportions are presented as percentages; likelihood ratios are presented as coefficients.

Total n = 200

SensitivitySe

—

SpecificitySp

—

Positive predictive valuePPV

—

Negative predictive valueNPV

—

Positive LRLR+

—

Negative LRLR−

—

Additional Performance Measures

Complementary indicators summarizing different aspects of test performance.

Measure	Result	Confidence interval / range	Definition

Post-test Probability

Updates the pre-test probability using LR+ and LR−.

Pre-test probability (%)

Post-test probability after a positive result —

Post-test probability of disease after a negative result —

Academic Reporting Text

An automatically generated draft that can be adapted for a Results section.

—

Interpretation note: PPV and NPV are affected by disease prevalence in the study population. Sensitivity, specificity, and likelihood ratios should also be interpreted in the context of participant selection, the accuracy of the reference standard, and the validation design.

Confidence intervals for proportions are calculated using the Wilson score method; confidence intervals for likelihood ratios and the diagnostic odds ratio are calculated using a logarithmic approach.