Parameters of Screening Tests

Monte Carlo Simulation Parameters

Specify the ranges to simulate predictive values across multiple populations.

Number of Simulations

Sensitivity (%) (25, 75)

Specificity (%) (25, 75)

Prevalence (%) (25, 75)

Results based on Simulation

PPV Mean (SD)

--

NPV Mean (SD)

--

Interpretation of the Result

Run the simulation to generate an interpretation...

🔍 Why Combine Tests?

No single screening test is perfect. Healthcare professionals often combine tests to maximize the chances of correctly identifying a disease. Here are real-world examples:

Parallel Testing (Net Sensitivity): Multiple tests run at once. Example: An ER doctor ordering multiple blood cultures and imaging at the same time for suspected sepsis because missing the infection is fatal. Net sensitivity increases, specificity decreases.
Sequential Testing (Net Specificity): Tests run one after another. Example: Standard HIV screening where a highly sensitive ELISA test is done first, and if positive, a highly specific Western Blot is done to confirm. Net specificity increases, sensitivity decreases.

⚖️ Statistical Errors

Type I Error (Alpha)

False Positive

Telling a healthy patient they are sick. Causes anxiety.

Type II Error (Beta)

False Negative

Missing a disease in a sick patient. Extremely dangerous.

Network Parameters

Test A

Sensitivity95%

Specificity80%

Test B

Sensitivity80%

Specificity99%

Disease Prevalence (%) 10%

Combined Test Performace

Net Sensitivity

--

Net Specificity

--

Net PPV

--

Net NPV

--

Confusion Matrix for 1,000 Patients

Evaluate a New Screening Test

Enter the counts from your study comparing a new screening test against a Gold Standard test to automatically calculate its parameters.

Gold Standard (+)
(Disease Present)

Gold Standard (-)
(Disease Absent)

New Test (+)
(Screen Positive)

True Positive (a)

False Positive (b)

New Test (-)
(Screen Negative)

False Negative (c)

True Negative (d)

Sensitivity [a / (a+c)]

--

Specificity [d / (b+d)]

--

PPV [a / (a+b)]

--

NPV [d / (c+d)]

--

Accuracy [(a+d)/Total]

--

Prevalence [(a+c)/Total]

--

📈 Concept of ROC Curve Analysis

The Receiver Operating Characteristic (ROC) curve is a graphical plot that illustrates the diagnostic ability of a binary classifier system as its discrimination threshold is varied.

It is created by plotting the True Positive Rate (Sensitivity) on the Y-axis against the False Positive Rate (1 - Specificity) on the X-axis at various threshold settings.

AUC (Area Under the Curve): Measures the total two-dimensional area underneath the ROC curve. A model whose predictions are 100% wrong has an AUC of 0.0; one whose predictions are perfectly correct has an AUC of 1.0.
Perfect Test: Passes through the upper left corner (100% Sensitivity, 0% False Positives).
Useless Test: Follows the diagonal line (AUC = 0.5), meaning it is no better than randomly flipping a coin.

Interactive ROC Curve Visualization

Adjust the diagnostic sliding threshold to observe how moving the cutoff for a "positive" test result trades off Sensitivity versus Specificity. The left side of the distributions typically represents healthy individuals, while the right side represents diseased individuals.

Diagnostic Threshold Cutoff 50

Sensitivity (TPR)

--

False Positive Rate (FPR)

--

(100% - Specificity)

AUC-ROC (Area Under Curve)

--