Discover how test accuracy and disease prevalence interact to create surprisingly high false positive rates.
% of population with condition
True positive rate (test correctly identifies condition)
True negative rate (test correctly identifies absence)
Per 100,000 tested:
The false positive paradox occurs when a test with high accuracy produces mostly false results. This counterintuitive phenomenon happens when the condition being tested is rare in the population. Even with a test that correctly identifies the condition 99% of the time and correctly identifies its absence 95% of the time, if the condition affects only 1% of the population, more than half of positive test results will be false positives.
This paradox is crucial in medical testing, screening programs, and security applications. A positive test result doesn't necessarily mean the person has the condition—it depends on how common the condition is. This is captured by Bayes' theorem, which combines prior probability (how common the condition is) with test accuracy to compute the posterior probability (the likelihood you actually have the condition given a positive test).
Understanding this paradox helps explain why doctors often order confirmatory tests after positive screening results, and why mass screening programs require careful consideration of disease prevalence before implementation.
Suppose a rare disease affects 1% of the population. A test has 99% sensitivity (catches 99% of cases) and 95% specificity (correctly identifies 95% of non-cases). What is the probability that a person with a positive test actually has the disease?
Given:
• Prevalence = 1% = 0.01
• Sensitivity = 99% = 0.99
• Specificity = 95% = 0.95
Per 100,000 people:
• Infected: 100,000 × 0.01 = 1,000
• Not infected: 100,000 × 0.99 = 99,000
• True positives: 1,000 × 0.99 = 990
• False positives: 99,000 × (1 − 0.95) = 4,950
• Total positive tests: 990 + 4,950 = 5,940
• PPV = 990 / 5,940 = 16.7%
Result: Despite the test's high accuracy, only 16.7% of positive results indicate someone actually has the disease. This means 83.3% of positive tests are false positives! This illustrates why confirmatory testing is essential for rare conditions.
Why does a 99% accurate test still produce so many false positives?
Because for rare conditions, the 1% incorrect results from the large non-affected population outnumber the true positives from the small affected population. With 99,000 unaffected people, even 1% error (990 false positives) can exceed actual positives.
What's the difference between sensitivity and specificity?
Sensitivity measures how well a test identifies positive cases (true positive rate), while specificity measures how well it identifies negative cases (true negative rate). A test can be highly sensitive but have low specificity, or vice versa.
How does prevalence affect the results?
As prevalence increases, PPV increases and false positive rate decreases. For common conditions, positive tests are more reliable. For rare conditions, positive tests are less reliable and require confirmation.
What is Positive Predictive Value (PPV)?
PPV is the probability that a person with a positive test result actually has the condition. It's calculated using Bayes' theorem and depends on the test's sensitivity, specificity, and the disease's prevalence.
When would we expect high PPV?
PPV is high when either (1) prevalence is high, or (2) specificity is very high. Medical screening programs typically use tests with very high specificity to maintain reasonable PPV even for rare conditions.
Why do doctors order confirmatory tests?
A confirmatory test with different characteristics can reduce false positives. If both tests are positive, PPV increases substantially even for rare conditions, providing greater confidence in the diagnosis.
How is this relevant to COVID-19 testing?
During low-prevalence periods, rapid COVID tests (which have lower specificity) produce many false positives. Confirmatory PCR tests are recommended for positive rapid tests to rule out false positives.
Can PPV ever exceed sensitivity or specificity?
Yes. If a condition is common (high prevalence) and specificity is high, PPV can exceed sensitivity. PPV depends on the combination of all three factors, not just test accuracy alone.
Related Tools
Shared birthday probability.
Geometric probability.
Bayesian paradox.
Conditional gender probability.
Losing games winning strategy.
Money switching paradox.