Table of Contents
The Paradox Explained
The false positive paradox occurs when a test with high accuracy produces more false positives than true positives, simply because the condition being tested for is rare. Even a test that is 99% sensitive and 95% specific will produce mostly false positive results when prevalence is low (e.g., 1%).
This paradox has profound implications for mass screening programs, drug testing in workplaces, airport security screening, and criminal forensics. It demonstrates why understanding base rates is essential for interpreting test results correctly.
Worked Example
In a population of 10,000 with 1% prevalence (100 sick): Test sensitivity 99%: detects 99 of 100 sick (TP=99). Test specificity 95%: correctly clears 9,405 of 9,900 healthy, but falsely flags 495 (FP=495). Total positives: 99+495 = 594. Only 99/594 = 16.7% of positive results are truly sick!
| Test Positive | Test Negative | Total | |
|---|---|---|---|
| Sick | 99 (TP) | 1 (FN) | 100 |
| Healthy | 495 (FP) | 9,405 (TN) | 9,900 |
| Total | 594 | 9,406 | 10,000 |
Why It Happens
The healthy population vastly outnumbers the sick. Even a small false positive rate (5%) applied to 9,900 healthy people produces 495 false alarms, overwhelming the 99 true positives from the small sick group. The key insight: when prevalence is low, most positive results are false.
Solutions
- Sequential testing: Confirm positives with a second independent test
- Targeted screening: Test only high-risk populations (higher prevalence)
- Improve specificity: Even small specificity gains dramatically reduce false positives
- Bayesian updating: Use prior probability + test result to compute posterior probability
Frequently Asked Questions
Does this mean screening tests are useless?
No. Screening is still valuable because it catches true positives who need treatment. But positive results should be confirmed with follow-up testing rather than acted upon immediately.
How does prevalence affect PPV?
Higher prevalence dramatically increases PPV. At 10% prevalence with the same test, PPV rises to about 69%. At 50% prevalence, PPV exceeds 95%.
Is this relevant to COVID testing?
Yes. During low-prevalence periods, rapid antigen tests (lower specificity) can produce many false positives. PCR confirmation is recommended for positive rapid tests when community prevalence is low.