Calculate false positive rate, PPV, NPV, and 11 diagnostic metrics from prevalence, sensitivity, and specificity with confusion matrix and PPV-prevalence curve.
False positive rates are critical in medical screening, drug testing, security screening, and any binary classification system. A test with 99% sensitivity and 95% specificity sounds excellent — but when the condition is rare (say 1% prevalence), most positive results are actually false positives. This calculator reveals that counterintuitive reality.
Enter the prevalence, sensitivity, and specificity of any diagnostic test to see the full picture: confusion matrix with raw counts, positive and negative predictive values, false discovery rate, likelihood ratios, and a curve showing how PPV changes with prevalence. The population-scaled confusion matrix makes abstract probabilities concrete by showing actual numbers of true positives, false positives, true negatives, and false negatives.
This tool is essential for understanding screening programs, evaluating diagnostic tests, and communicating risk to patients. The PPV-prevalence curve demonstrates why mass screening for rare conditions produces so many false alarms, and why targeted testing of high-risk groups is more effective.
Misunderstanding false positive rates leads to patient anxiety, unnecessary follow-up procedures, and poor clinical decisions. Studies show that even physicians often overestimate PPV for screening tests. This calculator makes the base rate fallacy tangible with concrete numbers.
The population-scaled confusion matrix and PPV-prevalence curve are powerful communication tools for explaining screening results to patients or stakeholders. Seeing that 5,940 positive results contain only 990 true cases is far more impactful than abstract probability statements.
PPV = TP / (TP + FP) = (Sensitivity × Prevalence) / (Sensitivity × Prevalence + (1 − Specificity) × (1 − Prevalence)) NPV = TN / (TN + FN) FPR = FP / (FP + TN) = 1 − Specificity FDR = FP / (TP + FP) = 1 − PPV LR+ = Sensitivity / (1 − Specificity) LR− = (1 − Sensitivity) / Specificity
Result: PPV = 16.67%, FDR = 83.33%, FPR = 5%, NPV = 99.99%
In a population of 100,000 with 1% prevalence: 1,000 are diseased, 99,000 healthy. The test correctly identifies 990 (TP) but misses 10 (FN). Among the healthy, 4,950 get false positives. So out of 5,940 positive results, only 990 (16.7%) actually have the disease. Most "positive" results are false alarms.
The base rate fallacy — ignoring how common a condition is when interpreting test results — is one of the most well-documented cognitive biases in medicine. In a famous study by Casscells, Schoenberger, and Graboy (1978), most Harvard Medical School students and staff incorrectly estimated the PPV of a test with 95% sensitivity and 5% FPR in a population with 1/1000 prevalence. The correct answer is about 2%, but most answered 95%. This calculator helps combat that error.
Screening tests are applied to asymptomatic populations (mammography, newborn screening, airport security). Diagnostic tests are applied to symptomatic or high-risk individuals. Because screening populations have lower prevalence, screening tests almost always have lower PPV than the same test used diagnostically. This is why positive screening results require confirmatory testing.
Research by Gerd Gigerenzer shows that people reason about probabilities far better when presented with natural frequencies (e.g., "out of 1,000 women, 10 have cancer, 9 test positive...") rather than conditional probabilities. This calculator's confusion matrix with raw counts implements that approach, making Bayesian reasoning accessible to anyone.
When prevalence is low, the false positives from the large healthy group vastly outnumber the true positives from the small diseased group. With 1% prevalence, there are 99 healthy people for every 1 sick person. Even a small FPR of 5% from 99,000 healthy people produces 4,950 false positives, dwarfing the 990 true positives.
FPR (False Positive Rate) = FP / (all healthy) answers "what fraction of healthy people get a false alarm?" FDR (False Discovery Rate) = FP / (all test positives) answers "what fraction of positive results are wrong?" FDR is what patients care about; FPR is a property of the test itself.
Three ways: (1) increase specificity (reduces false positives), (2) test higher-prevalence populations (targeted screening), or (3) use a confirmatory test on initial positives. Option 2 is why doctors order tests based on symptoms rather than screening everyone.
LR+ tells you how much more likely a positive result is in someone with the disease vs. without. LR+ > 10 is considered a strong diagnostic indicator. LR− tells how much less likely a negative result is in someone with the disease. LR− < 0.1 is strong for ruling out.
Absolutely. Any binary classifier has these metrics: spam filters (spam vs. not-spam), fraud detection (fraudulent vs. legitimate), quality control (defective vs. good), security screening (threat vs. non-threat). Replace "sensitivity" with "recall" and "PPV" with "precision" for ML terminology.
It shows that PPV increases steeply with prevalence. At very low prevalence, PPV is near zero regardless of test quality. This is the base rate fallacy — rare events produce mostly false positives even with good tests. The curve quantifies exactly where the transition happens.